arXiv Daily Digest - 2026-03-24

Large Reasoning Models (LRMs) achieve strong accuracy on challenging tasks by generating long Chain-of-Thought traces, but suffer from overthinking. Even after reaching the correct answer, they continue generating redundant reasoning steps. This behavior increases latency and compute cost and can also lead to answer drift. Existing mitigation methods either require training-heavy backbone modification or rely on hand-crafted heuristics that do not truly capture overthinking patterns. We propose ROM, the first method that formulates overthinking mitigation as a streaming prediction-and-control problem. ROM attaches a lightweight detection head to the late-layer hidden states of a frozen large language model backbone. It monitors tokens in real time and triggers an early transition to the final answer once overthinking is detected. We also introduce token-level supervision based on solution correctness boundaries and a data augmentation strategy that reduces distilled-data bias. Across seven benchmarks, ROM achieves the highest accuracy (93.51%), the shortest responses (1,159 tokens), and the best response efficiency. Compared with the vanilla baseline, it reduces response length by 47.2% and improves efficiency by 121%. These results show that streaming detection is a promising approach to real-time overthinking mitigation.

Detecting climate disinformation narratives typically relies on fixed taxonomies, which do not accommodate emerging narratives. Thus, we re-frame narrative detection as a retrieval task: given a narrative's core message as a query, rank texts from a corpus by alignment with that narrative. This formulation requires no predefined label set and can accommodate emerging narratives. We repurpose three climate disinformation datasets (CARDS, Climate Obstruction, climate change subset of PolyNarrative) for retrieval evaluation and propose SpecFi, a framework that generates hypothetical documents to bridge the gap between abstract narrative descriptions and their concrete textual instantiations. SpecFi uses community summaries from graph-based community detection as few-shot examples for generation, achieving a MAP of 0.505 on CARDS without access to narrative labels. We further introduce narrative variance, an embedding-based difficulty metric, and show via partial correlation analysis that standard retrieval degrades on high-variance narratives (BM25 loses 63.4% of MAP), while SpecFi-CS remains robust (32.7% loss). Our analysis also reveals that unsupervised community summaries converge on descriptions close to expert-crafted taxonomies, suggesting that graph-based methods can surface narrative structure from unlabeled text.

Retrieval over large codebases is a key component of modern LLM-based software engineering systems. Existing approaches predominantly rely on dense embedding models, while learned sparse retrieval (LSR) remains largely unexplored for code. However, applying sparse retrieval to code is challenging due to subword fragmentation, semantic gaps between natural-language queries and code, diversity of programming languages and sub-tasks, and the length of code documents, which can harm sparsity and latency. We introduce SPLADE-Code, the first large-scale family of learned sparse retrieval models specialized for code retrieval (600M-8B parameters). Despite a lightweight one-stage training pipeline, SPLADE-Code achieves state-of-the-art performance among retrievers under 1B parameters (75.4 on MTEB Code) and competitive results at larger scales (79.0 with 8B). We show that learned expansion tokens are critical to bridge lexical and semantic matching, and provide a latency analysis showing that LSR enables sub-millisecond retrieval on a 1M-passage collection with little effectiveness loss.

Upcoming stage-IV surveys such as Euclid and Rubin will deliver vast amounts of high-precision data, opening new opportunities to constrain cosmological models with unprecedented accuracy. A key step in this process is the reconstruction of the dark matter distribution from noisy weak lensing shear measurements. Current deep learning-based mass mapping methods achieve high reconstruction accuracy, but either require retraining a model for each new observed sky region (limiting practicality) or rely on slow MCMC sampling. Efficient exploitation of future survey data therefore calls for a new method that is accurate, flexible, and fast at inference. In addition, uncertainty quantification with coverage guarantees is essential for reliable cosmological parameter estimation. We introduce PnPMass, a plug-and-play approach for weak lensing mass mapping. The algorithm produces point estimates by alternating between a gradient descent step with a carefully chosen data fidelity term, and a denoising step implemented with a single deep learning model trained on simulated data corrupted by Gaussian white noise. We also propose a fast, sampling-free uncertainty quantification scheme based on moment networks, with calibrated error bars obtained through conformal prediction to ensure coverage guarantees. Finally, we benchmark PnPMass against both model-driven and data-driven mass mapping techniques. PnPMass achieves performance close to that of state-of-the-art deep-learning methods while offering fast inference (converging in just a few iterations) and requiring only a single training phase, independently of the noise covariance of the observations. It therefore combines flexibility, efficiency, and reconstruction accuracy, while delivering tighter error bars than existing approaches, making it well suited for upcoming weak lensing surveys.

The advent of Transformer and Mamba-based architectures has significantly advanced 3D medical image segmentation by enabling global contextual modeling, a capability traditionally limited in Convolutional Neural Networks (CNNs). However, state-of-the-art Transformer models often entail substantial computational complexity and parameter counts, which is particularly prohibitive for volumetric data and further exacerbated by the limited availability of annotated medical imaging datasets. To address these limitations, this work introduces SegMaFormer, a lightweight hybrid architecture that synergizes Mamba and Transformer modules within a hierarchical volumetric encoder for efficient long-range dependency modeling. The model strategically employs Mamba-based layers in early, high-resolution stages to reduce computational overhead while capturing essential spatial context, and reserves self-attention mechanisms for later, lower-resolution stages to refine feature representation. This design is augmented with generalized rotary position embeddings to enhance spatial awareness. Despite its compact structure, SegMaFormer achieves competitive performance on three public benchmarks (Synapse, BraTS, and ACDC), matching the Dice coefficient of significantly larger models. Empirically, our approach reduces parameters by up to 75x and substantially decreases FLOPs compared to current state-of-the-art models, establishing an efficient and high-performing solution for 3D medical image segmentation.

We propose a method for non-parametric conditional distribution estimation based on partitioning covariate-sorted observations into contiguous bins and using the within-bin empirical CDF as the predictive distribution. Bin boundaries are chosen to minimise the total leave-one-out Continuous Ranked Probability Score (LOO-CRPS), which admits a closed-form cost function with $O(n^2 \log n)$ precomputation and $O(n^2)$ storage; the globally optimal $K$-partition is recovered by a dynamic programme in $O(n^2 K)$ time. Minimisation of Within-sample LOO-CRPS turns out to be inappropriate for selecting $K$ as it results in in-sample optimism. So we instead select $K$ by evaluating test CRPS on an alternating held-out split, which yields a U-shaped criterion with a well-defined minimum. Having selected $K^*$ and fitted the full-data partition, we form two complementary predictive objects: the Venn prediction band and a conformal prediction set based on CRPS as the nonconformity score, which carries a finite-sample marginal coverage guarantee at any prescribed level $\varepsilon$. On real benchmarks against split-conformal competitors (Gaussian split conformal, CQR, and CQR-QRF), the method produces substantially narrower prediction intervals while maintaining near-nominal coverage.

StreamSampling$.$jl is a Julia library designed to provide general and efficient methods for sampling from data streams in a single pass, even when the total number of items is unknown. In this paper, we describe the capabilities of the library and its advantages over traditional sampling procedures, such as maintaining a small, constant memory footprint and avoiding the need to fully materialize the stream in memory. Furthermore, we provide empirical benchmarks comparing online sampling methods against standard approaches, demonstrating performance and memory improvements.

Gaussian Error Linear Unit (GELU) is a widely used smooth alternative to Rectifier Linear Unit (ReLU), yet many deployment, compression, and analysis toolchains are most naturally expressed for piecewise-linear (ReLU-type) networks. We study a hardness-parameterized formulation of GELU, f(x;λ)=xΦ(λ x), where Φ is the Gaussian CDF and λ \in [1, infty) controls gate sharpness, with the goal of turning smooth gated training into a controlled path toward ReLU-compatible models. Learning λ is non-trivial: naive updates yield unstable dynamics and effective gradient attenuation, so we introduce a constrained reparameterization and an optimizer-aware update scheme. Empirically, across a diverse set of model--dataset pairs spanning MLPs, CNNs, and Transformers, we observe structured layerwise hardness profiles and assess their robustness under different initializations. We further study a deterministic ReLU-ization strategy in which the learned gates are progressively hardened toward a principled target, enabling a post-training substitution of λ-GELU by ReLU with reduced disruption. Overall, λ-GELU provides a minimal and interpretable knob to profile and control gating hardness, bridging smooth training with ReLU-centric downstream pipelines.

Reinforcement Learning (RL) has demonstrated its ability to solve complex decision-making problems in a variety of domains, by optimizing reward signals obtained through interaction with an environment. However, many real-world scenarios involve multiple, potentially conflicting objectives that cannot be easily represented by a single scalar reward. Multi-Objective Reinforcement Learning (MORL) addresses this limitation by enabling agents to optimize several objectives simultaneously, explicitly reasoning about trade-offs between them. However, the ``black box" nature of the RL models makes the decision process behind chosen objective trade-offs unclear. Current Explainable Reinforcement Learning (XRL) methods are typically designed for single scalar rewards and do not account for explanations with respect to distinct objectives or user preferences. To address this gap, in this paper we propose TREX, a Trajectory based Explainability framework to explain Multi-objective Reinforcement Learning policies, based on trajectory attribution. TREX generates trajectories directly from the learned expert policy, across different user preferences and clusters them into semantically meaningful temporal segments. We quantify the influence of these behavioural segments on the Pareto trade-off by training complementary policies that exclude specific clusters, measuring the resulting relative deviation on the observed rewards and actions compared to the original expert policy. Experiments on multi-objective MuJoCo environments - HalfCheetah, Ant and Swimmer, demonstrate the framework's ability to isolate and quantify the specific behavioural patterns.

Autonomous vehicles face major perception and navigation challenges in adverse weather such as rain, fog, and snow, which degrade the performance of LiDAR, RADAR, and RGB camera sensors. While each sensor type offers unique strengths, such as RADAR robustness in poor visibility and LiDAR precision in clear conditions, they also suffer distinct limitations when exposed to environmental obstructions. This study proposes LRC-WeatherNet, a novel multi-sensor fusion framework that integrates LiDAR, RADAR, and camera data for real-time classification of weather conditions. By employing both early fusion using a unified Bird's Eye View representation and mid-level gated fusion of modality-specific feature maps, our approach adapts to the varying reliability of each sensor under changing weather. Evaluated on the extensive MSU-4S dataset covering nine weather types, LRC-WeatherNet achieves superior classification performance and computational efficiency, significantly outperforming unimodal baselines in adverse conditions. This work is the first to combine all three modalities for robust, real-time weather classification in autonomous driving. We release our trained models and source code in https://github.com/nouralhudaalbashir/LRC-WeatherNet.

Accurate power flow analysis is critical for modern distribution systems, yet classical solvers face scalability issues, and current machine learning models often struggle with generalization. We introduce BOOST-RPF, a novel method that reformulates voltage prediction from a global graph regression task into a sequential path-based learning problem. By decomposing radial networks into root-to-leaf paths, we leverage gradient-boosted decision trees (XGBoost) to model local voltage-drop regularities. We evaluate three architectural variants: Absolute Voltage, Parent Residual, and Physics-Informed Residual. This approach aligns the model architecture with the recursive physics of power flow, ensuring size-agnostic application and superior out-of-distribution robustness. Benchmarked against the Kerber Dorfnetz grid and the ENGAGE suite, BOOST-RPF achieves state-of-the-art results with its Parent Residual variant which consistently outperforms both analytical and neural baselines in standard accuracy and generalization tasks. While global Multi-Layer Perceptrons (MLPs) and Graph Neural Networks (GNNs) often suffer from performance degradation under topological shifts, BOOST-RPF maintains high precision across unseen feeders. Furthermore, the framework displays linear $O(N)$ computational scaling and significantly increased sample efficiency through per-edge supervision, offering a scalable and generalizable alternative for real-time distribution system operator (DSO) applications.

Large language models are becoming pervasive core components in many real-world applications. As a consequence, security alignment represents a critical requirement for their safe deployment. Although previous related works focused primarily on model architectures and alignment methodologies, these approaches alone cannot ensure the complete elimination of harmful generations. This concern is reinforced by the growing body of scientific literature showing that attacks, such as jailbreaking and prompt injection, can bypass existing security alignment mechanisms. As a consequence, additional security strategies are needed both to provide qualitative feedback on the robustness of the obtained security alignment at the training stage, and to create an ``ultimate'' defense layer to block unsafe outputs possibly produced by deployed models. To provide a contribution in this scenario, this paper introduces SecureBreak, a safety-oriented dataset designed to support the development of AI-driven solutions for detecting harmful LLM outputs caused by residual weaknesses in security alignment. The dataset is highly reliable due to careful manual annotation, where labels are assigned conservatively to ensure safety. It performs well in detecting unsafe content across multiple risk categories. Tests with pre-trained LLMs show improved results after fine-tuning on SecureBreak. Overall, the dataset is useful both for post-generation safety filtering and for guiding further model alignment and security improvements.

Reinforcement Learning (RL) is essential for evolving Large Language Models (LLMs) into autonomous agents capable of long-horizon planning, yet a practical recipe for scaling RL in complex, multi-turn environments remains elusive. This paper presents a systematic empirical study using TravelPlanner, a challenging testbed requiring tool orchestration to satisfy multifaceted constraints. We decompose the agentic RL design space along 5 axes: reward shaping, model scaling, data composition, algorithm selection, and environmental stability. Our controlled experiments yield 7 key takeaways, e.g., (1) reward and algorithm choices are scale-dependent as smaller models benefit from staged rewards and enhanced exploration, whereas larger models converge efficiently with simpler dense rewards, (2) ~ 1K training samples with a balanced difficulty mixture mark a sweet spot for both in-domain and out-of-domain performance, and (3) environmental stability is critical to prevent policy degradation. Based on our distilled recipe, our RL-trained models achieve state-of-the-art performance on TravelPlanner, significantly outperforming leading LLMs.

Fine-tuning large language models for domain-specific tasks such as medical text summarization demands substantial computational resources. Parameter-efficient fine-tuning (PEFT) methods offer promising alternatives by updating only a small fraction of parameters. This paper compares three adaptation approaches-Low-Rank Adaptation (LoRA), Prompt Tuning, and Full Fine-Tuning-across the Flan-T5 model family on the PubMed medical summarization dataset. Through experiments with multiple random seeds, we demonstrate that LoRA consistently outperforms full fine-tuning, achieving 43.52 +/- 0.18 ROUGE-1 on Flan-T5-Large with only 0.6% trainable parameters compared to 40.67 +/- 0.21 for full fine-tuning. Sensitivity analyses examine the impact of LoRA rank and prompt token count. Our findings suggest the low-rank constraint provides beneficial regularization, challenging assumptions about the necessity of full parameter updates. Code is available at https://github.com/eracoding/llm-medical-summarization

We introduce the Burmese Handwritten Digit Dataset (BHDD), a collection of 87,561 grayscale images of handwritten Burmese digits in ten classes. Each image is 28x28 pixels, following the MNIST format. The training set has 60,000 samples split evenly across classes; the test set has 27,561 samples with class frequencies as they arose during collection. Over 150 people of different ages and backgrounds contributed samples. We analyze the dataset's class distribution, pixel statistics, and morphological variation, and identify digit pairs that are easily confused due to the round shapes of the Myanmar script. Simple baselines (an MLP, a two-layer CNN, and an improved CNN with batch normalization and augmentation) reach 99.40%, 99.75%, and 99.83% test accuracy respectively. BHDD is available under CC BY-SA 4.0 at https://github.com/baseresearch/BHDD

We introduce Suiren-1.0, a family of molecular foundation models for the accurate modeling of diverse organic systems. Suiren-1.0 comprising three specialized variants (Suiren-Base, Suiren-Dimer, and Suiren-ConfAvg) is integrated within an algorithmic framework that bridges the gap between 3D conformational geometry and 2D statistical ensemble spaces. We first pre-train Suiren-Base (1.8B parameters) on a 70M-sample Density Functional Theory dataset using spatial self-supervision and SE(3)-equivariant architectures, achieving robust performance in quantum property prediction. Suiren-Dimer extends this capability through continued pre-training on 13.5M intermolecular interaction samples. To enable efficient downstream application, we propose Conformation Compression Distillation (CCD), a diffusion-based framework that distills complex 3D structural representations into 2D conformation-averaged representations. This yields the lightweight Suiren-ConfAvg, which generates high-fidelity representations from SMILES or molecular graphs. Our extensive evaluations demonstrate that Suiren-1.0 establishes state-of-the-art results across a range of tasks. All models and benchmarks are open-sourced.

Spoken Language Understanding (SLU) aims to extract the semantic information from the speech utterance of user queries. It is a core component in a task-oriented dialogue system. With the spectacular progress of deep neural network models and the evolution of pre-trained language models, SLU has obtained significant breakthroughs. However, only a few high-resource languages have taken advantage of this progress due to the absence of SLU resources. In this paper, we seek to mitigate this obstacle by introducing SLURP-TN. This dataset was created by recording 55 native speakers uttering sentences in Tunisian dialect, manually translated from six SLURP domains. The result is an SLU Tunisian dialect dataset that comprises 4165 sentences recorded into around 5 hours of acoustic material. We also develop a number of Automatic Speech Recognition and SLU models exploiting SLUTP-TN. The Dataset and baseline models are available at: https://huggingface.co/datasets/Elyadata/SLURP-TN.

Quantitative disease severity scoring in medical imaging is costly, time-consuming, and subject to inter-reader variability. At the same time, clinical archives contain far more longitudinal imaging data than expert-annotated severity scores. Existing self-supervised methods typically ignore this chronological structure. We introduce ChronoCon, a contrastive learning approach that replaces label-based ranking losses with rankings derived solely from the visitation order of a patient's longitudinal scans. Under the clinically plausible assumption of monotonic progression in irreversible diseases, the method learns disease-relevant representations without using any expert labels. This generalizes the idea of Rank-N-Contrast from label distances to temporal ordering. Evaluated on rheumatoid arthritis radiographs for severity assessment, the learned representations substantially improve label efficiency. In low-label settings, ChronoCon significantly outperforms a fully supervised baseline initialized from ImageNet weights. In a few-shot learning experiment, fine-tuning ChronoCon on expert scores from only five patients yields an intraclass correlation coefficient of 86% for severity score prediction. These results demonstrate the potential of chronological contrastive learning to exploit routinely available imaging metadata to reduce annotation requirements in the irreversible disease domain. Code is available at https://github.com/cirmuw/ChronoCon.

The pruning of 3D Gaussian splats is essential for reducing their complexity to enable efficient storage, transmission, and downstream processing. However, most of the existing pruning strategies depend on camera parameters, rendered images, or view-dependent measures. This dependency becomes a hindrance in emerging camera-agnostic exchange settings, where splats are shared directly as point-based representations (e.g., .ply). In this paper, we propose a camera-agnostic, one-shot, post-training pruning method for 3D Gaussian splats that relies solely on attribute-derived neighbourhood descriptors. As our primary contribution, we introduce a hybrid descriptor framework that captures structural and appearance consistency directly from the splat representation. Building on these descriptors, we formulate pruning as a statistical evidence estimation problem and introduce a Beta evidence model that quantifies per-splat reliability through a probabilistic confidence score. Experiments conducted on standardized test sequences defined by the ISO/IEC MPEG Common Test Conditions (CTC) demonstrate that our approach achieves substantial pruning while preserving reconstruction quality, establishing a practical and generalizable alternative to existing camera-dependent pruning strategies.

Continual Test-Time Adaptation (CTTA) aims to enable models to adapt online to unlabeled data streams under distribution shift without accessing source data. Existing CTTA methods face an efficiency-generalization trade-off: updating more parameters improves adaptation but severely reduces online inference efficiency. An ideal solution is to achieve comparable adaptation with minimal feature updates; we call this minimal subspace the golden subspace. We prove its existence in a single-step adaptation setting and show that it coincides with the row space of the pretrained classifier. To enable online maintenance of this subspace, we introduce the sample-wise Average Gradient Outer Product (AGOP) as an efficient proxy for estimating the classifier weights without retraining. Building on these insights, we propose Guided Online Low-rank Directional adaptation (GOLD), which uses a lightweight adapter to project features onto the golden subspace and learns a compact scaling vector while the subspace is dynamically updated via AGOP. Extensive experiments on classification and segmentation benchmarks, including autonomous-driving scenarios, demonstrate that GOLD attains superior efficiency, stability, and overall performance. Our code is available at https://github.com/AIGNLAI/GOLD.

In this work, we propose Oph-Guid-RAG, a multimodal visual RAG system for ophthalmology clinical question answering and decision support. We treat each guideline page as an independent evidence unit and directly retrieve page images, preserving tables, flowcharts, and layout information. We further design a controllable retrieval framework with routing and filtering, which selectively introduces external evidence and reduces noise. The system integrates query decomposition, query rewriting, retrieval, reranking, and multimodal reasoning, and provides traceable outputs with guideline page references. We evaluate our method on HealthBench using a doctor-based scoring protocol. On the hard subset, our approach improves the overall score from 0.2969 to 0.3861 (+0.0892, +30.0%) compared to GPT-5.2, and achieves higher accuracy, improving from 0.5956 to 0.6576 (+0.0620, +10.4%). Compared to GPT-5.4, our method achieves a larger accuracy gain of +0.1289 (+24.4%). These results show that our method is more effective on challenging cases that require precise, evidence-based reasoning. Ablation studies further show that reranking, routing, and retrieval design are critical for stable performance, especially under difficult settings. Overall, we show how combining visionbased retrieval with controllable reasoning can improve evidence grounding and robustness in clinical AI applications,while pointing out that further work is needed to be more complete.

The temporal difference (TD) error was first formalized in Sutton (1988), where it was first characterized as the difference between temporally successive predictions, and later, in that same work, formulated as the difference between a bootstrapped target and a prediction. Since then, these two interpretations of the TD error have been used interchangeably in the literature, with the latter eventually being adopted as the standard critic loss in deep reinforcement learning (RL) architectures. In this work, we show that these two interpretations of the TD error are not always equivalent. In particular, we show that increasingly-nonlinear deep RL architectures can cause these interpretations of the TD error to yield increasingly different numerical values. Then, building on this insight, we show how choosing one interpretation of the TD error over the other can affect the performance of deep RL algorithms that utilize the TD error to compute other quantities, such as with deep differential (i.e., average-reward) RL methods. All in all, our results show that the default interpretation of the TD error as the difference between a bootstrapped target and a prediction does not always hold in deep RL settings.

This paper develops a unified framework for measuring concentration in weighted systems embedded in networks of interactions. While traditional indices such as the Herfindahl-Hirschman Index capture dispersion in weights, they neglect the topology of relationships among the elements receiving those weights. To address this limitation, we introduce a family of topology-aware concentration indices that jointly account for weight distributions and network structure. At the core of the framework lies a baseline Network Concentration Index (NCI), defined as a normalized quadratic form that measures the fraction of potential weighted interconnection realized along observed network links. Building on this foundation, we construct a flexible class of extensions that modify either the interaction structure or the normalization benchmark, including weighted, density-adjusted, null-model, degree-constrained, transformed-data, and multi-layer variants. This family of indices preserves key properties such as normalization, invariance, and interpretability, while allowing concentration to be evaluated across different dimensions of dependence, including intensity, higher-order interactions, and extreme events. Theoretical results characterize the indices and establish their relationship with classical concentration and network measures. Empirical and simulation evidence demonstrate that systems with identical weight distributions may exhibit markedly different levels of structural concentration depending on network topology, highlighting the additional information captured by the proposed framework. The approach is broadly applicable to economic, financial, and complex systems in which weighted elements interact through networks.

Synthetic hyperspectral image (HSI) generation is essential for large-scale simulation, algorithm development, and mission design, yet traditional radiative transfer models remain computationally expensive and often limited to spectrum-level outputs. In this work, we propose a latent representation-based framework for hyperspectral emulation that learns a latent generative representation of hyperspectral data. The proposed approach supports both spectrum-level and spatial-spectral emulation and can be trained either in a direct one-step formulation or in a two-step strategy that couples variational autoencoder (VAE) pretraining with parameter-to-latent interpolation. Experiments on PROSAIL-simulated vegetation data and Sentinel-3 OLCI imagery demonstrate that the method outperforms classical regression-based emulators in reconstruction accuracy, spectral fidelity, and robustness to real-world spatial variability. We further show that emulated HSIs preserve performance in downstream biophysical parameter retrieval, highlighting the practical relevance of emulated data for remote sensing applications.

In this article, we study the impact of self-heating effects (SHEs) and middle of line (MOL) and back-end of line (BEOL) induced parasitics on multi-tier CFET design, where multiple nanosheet devices are vertically stacked. We analyze and compare the 4-tier CFET design with the conventional 2-tier CFET, using TCAD models calibrated to experimental measurements. Additionally, TCAD simulations are used to model and analyze SHE-induced heat distribution and temperature profiles and to extract the detailed parasitic RC network from 3D models of CMOS inverters designed with full MOL and BEOL interconnects. At the device level, the maximum temperature rise (TMAX) caused by SHE in nFET and pFET devices of the 2-tier CFET architecture is 62 K and 74 K, respectively. Due to the increased distance from the substrate heat sink, the upper-tier nFET and pFET devices in the 4-tier design show higher TMAX of 83.5 K and 98.5 K and more heat trapping in the stacked layers. Furthermore, in the 4-tier CFET-based CMOS inverters, the BEOL-induced parasitic RCs are, respectively, 10 and 6.5 times higher in the top-tier than in the 2-tier CFET-based inverters. In the bottom tier, the corresponding parasitic RC elements are 6.26 and 2 times higher, respectively, than in the 2-tier inverters. Finally, compared to the 4-tier design without parasitics, the propagation delay of the top and bottom tier inverters increases by 10% and 8.2%, respectively, due to the interconnect parasitic RCs. For the conventional 2-tier inverter, the corresponding degradation of delay with parasitic RCs is 37.25%.

We present a systematic method for exactly enforcing Dirichlet, Neumann, and Robin type conditions on general quadrilateral domains with arbitrary curved boundaries. Our method is built upon exact mappings between general quadrilateral domains and the standard domain, and employs a combination of TFC (theory of functional connections) constrained expressions and transfinite interpolations. When Neumann or Robin boundaries are present, especially when two Neumann (or Robin) boundaries meet at a vertex, it is critical to enforce exactly the induced compatibility constraints at the intersection, in order to enforce exactly the imposed conditions on the joining boundaries. We analyze in detail and present constructions for handling the imposed boundary conditions and the induced compatibility constraints for two types of situations: (i) when Neumann (or Robin) boundary only intersects with Dirichlet boundaries, and (ii) when two Neumann (or Robin) boundaries intersect with each other. We describe a four-step procedure to systematically formulate the general form of functions that exactly satisfy the imposed Dirichlet, Neumann, or Robin conditions on general quadrilateral domains. The method developed herein has been implemented together with the extreme learning machine (ELM) technique we have developed recently for scientific machine learning. Ample numerical experiments are presented with several linear/nonlinear stationary/dynamic problems on a variety of two-dimensional domains with complex boundary geometries. Simulation results demonstrate that the proposed method has enforced the Dirichlet, Neumann, and Robin conditions on curved domain boundaries exactly, with the numerical boundary-condition errors at the machine accuracy.

Deploying deep neural networks (DNNs) on power-sensitive edge devices presents a formidable challenge. While Dynamic Voltage and Frequency Scaling (DVFS) is widely employed for energy optimization, traditional model-level scaling is often too coarse to capture intra-inference variations, whereas fine-grained operator-level scaling suffers from prohibitive performance degradation due to significant hardware switching latency. This paper presents SparseDVFS, a fine-grained, sparse-aware DVFS framework designed for energy-efficient edge inference. Our key insight is that operator sparsity is a primary metric for hardware frequency modulation. By distinguishing between compute-bound dense operators and memory-bound sparse operators, the system can apply specialized frequency triplets to maximize energy efficiency. To overcome switching overheads and component interference, SparseDVFS incorporates three key innovations: (1) an offline modeler that established a deterministic mapping between operator sparsity and optimal frequency triplets (CPU/GPU/EMC) via white-box timeline analysis; (2) a runtime graph partitioner that utilizes a greedy merging heuristic to aggregate operators into super-blocks, balancing scaling granularity and DVFS switching latency through a latency amortization constraint; and (3) a unified co-governor that employs a frequency unified scaling engine (FUSE) and a look-ahead instruction queue to eliminate antagonistic effects between independent controllers and hide hardware transition latencies. Extensive evaluations show that SparseDVFS achieves an average 78.17% energy efficiency gain over state-of-the-art solutions while maintaining a superior 14% cost-gain ratio.

Unsupervised Domain Adaptation (UDA) is essential for deploying medical segmentation models across diverse clinical environments. Existing methods are fundamentally limited, suffering from semantically unaware feature alignment that results in poor distributional fidelity and from pseudo-label validation that disregards global anatomical constraints, thus failing to prevent the formation of globally implausible structures. To address these issues, we propose SHAPE (Structure-aware Hierarchical Unsupervised Domain Adaptation with Plausibility Evaluation), a framework that reframes adaptation towards global anatomical plausibility. Built on a DINOv3 foundation, its Hierarchical Feature Modulation (HFM) module first generates features with both high fidelity and class-awareness. This shifts the core challenge to robustly validating pseudo-labels. To augment conventional pixel-level validation, we introduce Hypergraph Plausibility Estimation (HPE), which leverages hypergraphs to assess the global anatomical plausibility that standard graphs cannot capture. This is complemented by Structural Anomaly Pruning (SAP) to purge remaining artifacts via cross-view stability. SHAPE significantly outperforms prior methods on cardiac and abdominal cross-modality benchmarks, achieving state-of-the-art average Dice scores of 90.08% (MRI->CT) and 78.51% (CT->MRI) on cardiac data, and 87.48% (MRI->CT) and 86.89% (CT->MRI) on abdominal data. The code is available at https://github.com/BioMedIA-repo/SHAPE.

We present Ara-BEST-RQ, a family of self-supervised learning (SSL) models specifically designed for multi-dialectal Arabic speech processing. Leveraging 5,640 hours of crawled Creative Commons speech and combining it with publicly available datasets, we pre-train conformer-based BEST-RQ models up to 600M parameters. Our models are evaluated on dialect identification (DID) and automatic speech recognition (ASR) tasks, achieving state-of-the-art performance on the former while using fewer parameters than competing models. We demonstrate that family-targeted pre-training on Arabic dialects significantly improves downstream performance compared to multilingual or monolingual models trained on non-Arabic data. All models, code, and pre-processed datasets will be publicly released to support reproducibility and further research in Arabic speech technologies.

New technologies, such as blockchain, are designed to address various system weaknesses, particularly those related to security. Blockchain can enhance numerous aspects of traditional banking systems by transforming them into digital, immutable, secure, and anonymous ledger. This paper proposes a new banking application ALBank, which is based on blockchain and smart contract technologies. Its functionality relies on invoking functions within smart contracts deployed on the Ethereum blockchain. This approach enables decentralization and enhances both security and trust. In this context, the paper first presents a critical analysis of existing research on blockchain and traditional banking systems, with a focus on their respective challenges. It then examines the Know Your Customer (KYC) process and its various models. Finally, it introduces the design and development of ALBank, a decentralized banking application built on the Ethereum blockchain using smart contracts. The results show that the integration of blockchain and smart contracts effectively addresses key issues in traditional banking systems, including centralization, inefficiency, and security vulnerabilities by storing critical data on a decentralized, immutable ledger, managing processes autonomously, and making transactions transparent to all users.

Low Rank Adaptation (LoRA) is the de facto fine-tuning strategy to generate personalized images from pre-trained diffusion models. Choosing a good rank is extremely critical, since it trades off performance and memory consumption, but today the decision is often left to the community's consensus, regardless of the personalized subject's complexity. The reason is evident: the cost of selecting a good rank for each LoRA component is combinatorial, so we opt for practical shortcuts such as fixing the same rank for all components. In this paper, we take a first step to overcome this challenge. Inspired by variational methods that learn an adaptive width of neural networks, we let the ranks of each layer freely adapt during fine-tuning on a subject. We achieve it by imposing an ordering of importance on the rank's positions, effectively encouraging the creation of higher ranks when strictly needed. Qualitatively and quantitatively, our approach, LoRA$^2$, achieves a competitive trade-off between DINO, CLIP-I, and CLIP-T across 29 subjects while requiring much less memory and lower rank than high rank LoRA versions. Code: https://github.com/donaldssh/NotAllLayersAreCreatedEqual.

Weather forecasting supports critical socioeconomic activities and complements environmental protection, yet operational Numerical Weather Prediction (NWP) systems remain computationally intensive, thus being inefficient for certain applications. Meanwhile, recent advances in deep data-driven models have demonstrated promising results in nowcasting tasks. This paper presents SmaAT-QMix-UNet, an enhanced variant of SmaAT-UNet that introduces two key innovations: a vector quantization (VQ) bottleneck at the encoder-decoder bridge, and mixed kernel depth-wise convolutions (MixConv) replacing selected encoder and decoder blocks. These enhancements both reduce the model's size and improve its nowcasting performance. We train and evaluate SmaAT-QMix-UNet on a Dutch radar precipitation dataset (2016-2019), predicting precipitation 30 minutes ahead. Three configurations are benchmarked: using only VQ, only MixConv, and the full SmaAT-QMix-UNet. Grad-CAM saliency maps highlight the regions influencing each nowcast, while a UMAP embedding of the codewords illustrates how the VQ layer clusters encoder outputs. The source code for SmaAT-QMix-UNet is publicly available on GitHub \footnote{\href{https://github.com/nstavr04/MasterThesisSnellius}{https://github.com/nstavr04/MasterThesisSnellius}}.

Reinforcement Learning with Verifiable Rewards (RLVR) has emerged as a powerful paradigm for enhancing the reasoning capabilities of Large Language Models (LLMs). However, vanilla RLVR suffers from inefficient exploration, particularly when confronting "hard samples" that yield nearzero success rates. In such scenarios, the reliance on sparse outcome rewards typically results in zero-advantage estimates, effectively starving the model of supervision signals despite the high informational value of these instances. To address this, we propose P^2O, a novel framework that synergizes Prompt Optimization with Policy Optimization. P^2O identifies hard samples during training iterations and leverages the GeneticPareto (GEPA) prompt optimization algorithm to evolve prompt templates that guide the model toward discovering successful trajectories. Crucially, unlike traditional prompt engineering methods that rely on input augmentation, P^2O distills the reasoning gains induced by these optimized prompts directly into the model parameters. This mechanism provides denser positive supervision signals for hard samples and accelerates convergence. Extensive experiments demonstrate that P^2O not only achieves superior performance on in-distribution datasets but also exhibits strong generalization, yielding substantial improvements on out-of-distribution benchmarks (+4.7% avg.).

Speech deepfake source verification systems aims to determine whether two synthetic speech utterances originate from the same source generator, often assuming that the resulting source embeddings are independent of speaker traits. However, this assumption remains unverified. In this paper, we first investigate the impact of speaker factors on source verification. We propose a speaker-disentangled metric learning (SDML) framework incorporating two novel loss functions. The first leverages Chebyshev polynomial to mitigate gradient instability during disentanglement optimization. The second projects source and speaker embeddings into hyperbolic space, leveraging Riemannian metric distances to reduce speaker information and learn more discriminative source features. Experimental results on MLAAD benchmark, evaluated under four newly proposed protocols designed for source-speaker disentanglement scenarios, demonstrate the effectiveness of SDML framework. The code, evaluation protocols and demo website are available at https://github.com/xxuan-acoustics/RiemannSD-Net.

Group Relative Policy Optimization (GRPO) methods for video generation like FlowGRPO remain far less reliable than their counterparts for language models and images. This gap arises because video generation has a complex solution space, and the ODE-to-SDE conversion used for exploration can inject excess noise, lowering rollout quality and making reward estimates less reliable, which destabilizes post-training alignment. To address this problem, we view the pre-trained model as defining a valid video data manifold and formulate the core problem as constraining exploration within the vicinity of this manifold, ensuring that rollout quality is preserved and reward estimates remain reliable. We propose SAGE-GRPO (Stable Alignment via Exploration), which applies constraints at both micro and macro levels. At the micro level, we derive a precise manifold-aware SDE with a logarithmic curvature correction and introduce a gradient norm equalizer to stabilize sampling and updates across timesteps. At the macro level, we use a dual trust region with a periodic moving anchor and stepwise constraints so that the trust region tracks checkpoints that are closer to the manifold and limits long-horizon drift. We evaluate SAGE-GRPO on HunyuanVideo1.5 using the original VideoAlign as the reward model and observe consistent gains over previous methods in VQ, MQ, TA, and visual metrics (CLIPScore, PickScore), demonstrating superior performance in both reward maximization and overall video quality. The code and visual gallery are available at https://dungeonmassster.github.io/SAGE-GRPO-Page/.

While the rapid development of facial recognition algorithms has enabled numerous beneficial applications, their widespread deployment has raised significant concerns about the risks of mass surveillance and threats to individual privacy. In this paper, we introduce \textit{Adversarial Camouflage} as a novel solution for protecting users' privacy. This approach is designed to be efficient and simple to reproduce for users in the physical world. The algorithm starts by defining a low-dimensional pattern space parameterized by color, shape, and angle. Optimized patterns, once found, are projected onto semantically valid facial regions for evaluation. Our method maximizes recognition error across multiple architectures, ensuring high cross-model transferability even against black-box systems. It significantly degrades the performance of all tested state-of-the-art face recognition models during simulations and demonstrates promising results in real-world human experiments, while revealing differences in model robustness and evidence of attack transferability across architectures.

The emergence of generative AI is bringing about a significant transformation in knowledge management. Generative AI has the potential to address the limitations of conventional knowledge management systems, and it is increasingly being deployed in real-world settings with promising results. Related research is also expanding rapidly. However, much of this work focuses on research and practice related to the management of explicit knowledge. While fragmentary efforts have been made regarding the management of tacit knowledge using generative AI, the modeling and systematization that handle both tacit and explicit knowledge in an integrated manner remain insufficient. In this paper, we propose the "GenAI SECI" model as an updated version of the knowledge creation process (SECI) model, redesigned to leverage the capabilities of generative AI. A defining feature of the "GenAI SECI" model is the introduction of "Digital Fragmented Knowledge", a new concept that integrates explicit and tacit knowledge within cyberspace. Furthermore, a concrete system architecture for the proposed model is presented, along with a comparison with prior research models that share a similar problem awareness and objectives.

Video generation has recently emerged as a central task in the field of generative AI. However, the substantial computational cost inherent in video synthesis makes model distillation a critical technique for efficient deployment. Despite its significance, there is a scarcity of methods specifically designed for video diffusion models. Prevailing approaches often directly adapt image distillation techniques, which frequently lead to artifacts such as oversaturation, temporal inconsistency, and mode collapse. To address these challenges, we propose a novel distillation framework tailored specifically for video diffusion models. Its core innovations include: (1) an adaptive regression loss that dynamically adjusts spatial supervision weights to prevent artifacts arising from excessive distribution shifts; (2) a temporal regularization loss to counteract temporal collapse, promoting smooth and physically plausible sampling trajectories; and (3) an inference-time frame interpolation strategy that reduces sampling overhead while preserving perceptual quality. Extensive experiments and ablation studies on the VBench and VBench2 benchmarks demonstrate that our method achieves stable few-step video synthesis, significantly enhancing perceptual fidelity and motion realism. It consistently outperforms existing distillation baselines across multiple metrics.

Scaling laws for Large Language Models govern macroscopic resource allocation, yet translating them into precise Mixture-of-Experts (MoE) architectural configurations remains an open problem due to the combinatorially vast design space. Existing MoE scaling studies are constrained by experimental budgets to either augment scaling formulas with extra MoE variables, risking unreliable fits, or fix all non-MoE factors, ignoring global interactions. We propose a reusable framework for holistic MoE architectural optimization that bridges this gap. We first show that FLOPs per token alone is an inadequate fairness metric for MoE models because differing computational densities across layer types can inflate parameters without proportional compute cost, and establish a joint constraint triad of FLOPs per token, active parameters, and total parameters. We then reduce the 16-dimensional architectural search space to two sequential low-dimensional phases through algebraic constraints and a rank-preserving property of the hidden dimension. Validated across hundreds of MoE models spanning six orders of magnitude in compute, our framework yields robust scaling laws that map any compute budget to a complete, optimal MoE architecture. A key finding is that the near-optimal configuration band widens with scale, giving practitioners quantitative flexibility to balance scaling law recommendations against infrastructure constraints.

Do large language models reason morally, or do they merely sound like they do? We investigate whether LLM responses to moral dilemmas exhibit genuine developmental progression through Kohlberg's stages of moral development, or whether alignment training instead produces reasoning-like outputs that superficially resemble mature moral judgment without the underlying developmental trajectory. Using an LLM-as-judge scoring pipeline validated across three judge models, we classify more than 600 responses from 13 LLMs spanning a range of architectures, parameter scales, and training regimes across six classical moral dilemmas, and conduct ten complementary analyses to characterize the nature and internal coherence of the resulting patterns. Our results reveal a striking inversion: responses overwhelmingly correspond to post-conventional reasoning (Stages 5-6) regardless of model size, architecture, or prompting strategy, the effective inverse of human developmental norms, where Stage 4 dominates. Most strikingly, a subset of models exhibit moral decoupling: systematic inconsistency between stated moral justification and action choice, a form of logical incoherence that persists across scale and prompting strategy and represents a direct reasoning consistency failure independent of rhetorical sophistication. Model scale carries a statistically significant but practically small effect; training type has no significant independent main effect; and models exhibit near-robotic cross-dilemma consistency producing logically indistinguishable responses across semantically distinct moral problems. We posit that these patterns constitute evidence for moral ventriloquism: the acquisition, through alignment training, of the rhetorical conventions of mature moral reasoning without the underlying developmental trajectory those conventions are meant to represent.

This paper proposes a novel alternative to existing sim-to-real methods for training control policies with simulated experiences. Unlike prior methods that typically rely on domain randomization over a fixed finite set of parameters, the proposed approach injects state-dependent perturbations into the input joint torque during forward simulation. These perturbations are designed to simulate a broader spectrum of reality gaps than standard parameter randomization without requiring additional training. By using neural networks as flexible perturbation generators, the proposed method can represent complex, state-dependent uncertainties, such as nonlinear actuator dynamics and contact compliance, that parametric randomization cannot capture. Experimental results demonstrate that the proposed approach enables humanoid locomotion policies to achieve superior robustness against complex, unseen reality gaps in both simulation and real-world deployment.

A single two-input gate suffices for all of Boolean logic in digital hardware. No comparable primitive has been known for continuous mathematics: computing elementary functions such as sin, cos, sqrt, and log has always required multiple distinct operations. Here I show that a single binary operator, eml(x,y)=exp(x)-ln(y), together with the constant 1, generates the standard repertoire of a scientific calculator. This includes constants such as $e$, $π$, and $i$; arithmetic operations including $+$, $-$, $\times$, $/$, and exponentiation as well as the usual transcendental and algebraic functions. For example, $e^x=\operatorname{eml}(x,1)$, $\ln x=\operatorname{eml}(1,\operatorname{eml}(\operatorname{eml}(1,x),1))$, and likewise for all other operations. That such an operator exists was not anticipated; I found it by systematic exhaustive search and established constructively that it suffices for the concrete scientific-calculator basis. In EML (Exp-Minus-Log) form, every such expression becomes a binary tree of identical nodes, yielding a grammar as simple as $S \to 1 \mid \operatorname{eml}(S,S)$. This uniform structure also enables gradient-based symbolic regression: using EML trees as trainable circuits with standard optimizers (Adam), I demonstrate the feasibility of exact recovery of closed-form elementary functions from numerical data at shallow tree depths up to 4. The same architecture can fit arbitrary data, but when the generating law is elementary, it may recover the exact formula.

Verifying whether two implementations of the same large model are equivalent across frameworks is difficult in practice. Even when they realize the same computation, their graphs may differ substantially in operator decomposition, tensor layout, and the use of fused or opaque kernels, making manual rewrite rules hard to build and maintain. We present Emerge, a framework for checking Implementation Equivalence over computation graphs of large-model implementations. Instead of writing rules manually, Emerge represents the two implementations in an e-graph, infers candidate relations from execution values, and synthesizes rewrite rules on demand when existing rules are insufficient. Each synthesized rule is validated using the strongest applicable method, including SMT- based checking for symbolically tractable cases and constraint-aware randomized testing for opaque kernels, and then propagated through e-graph rebuilding to establish larger equivalences. Our current implementation targets inference computation graphs captured from HuggingFace Transformers and vLLM. Our evaluation shows that Emerge establishes equivalence for correct implementation pairs at practical cost, while also providing useful by-products for debugging: it detects 10 of 13 known implementation bugs and uncovers 8 previously unknown implementation issues that were later confirmed by developers. In addition, Emerge synthesizes block-level rules that compare favorably with manually authored ones.

Consumer-grade EEG is entering everyday devices, from earbuds to headbands, raising the question of whether language models can be adapted to individual neural responses. We test this by asking whether frozen LLM representations encode person-specific EEG signals, directions in activation space that predict one person's brain activity but not another's. Using word-level EEG from 30 participants reading naturalistic sentences (ZuCo corpus), we train a separate linear probe for each person, mapping hidden states from a frozen Qwen 2.5 7B to that individual's EEG power. Person-specific probes outperform a single population probe on every EEG feature tested; for high-gamma power, the person-specific probe achieves rho = 0.183, a ninefold improvement over the population probe (rho = 0.020, p < 10^-4). A negative control, fixation count, shows no person-specific advantage (p = 0.360); fixation count reflects word length and frequency rather than individual cognition. The individual directions are temporally stable (split-half cosine = 0.824), non-transferable across people (self rho = 0.369 vs. other rho = 0.143, p < 10^-19), and distinct from the shared population signal: person-specific probes retain predictive power after the population component is removed. The person-specific signal concentrates in the model's deep layers, rising consistently with depth and peaking at Layer 24 of 28. The results are consistent across architectures (LLaMA 3.1 8B) and survive word-level confound controls. Frozen language models contain stable, person-specific neural directions in their deep layers, providing a geometric foundation for EEG-driven personalization.

AI explanation methods often assume a static user model, producing non-adaptive explanations regardless of expert goals, reasoning strategies, or decision contexts. Knowledge graph-based explanations, despite their capacity for grounded, path-based reasoning, inherit this limitation. In complex domains such as scientific discovery, this assumption fails to capture the diversity of cognitive strategies and epistemic stances among experts, preventing explanations that foster deeper understanding and informed decision-making. However, the scarcity of human experts limits the use of direct human feedback to produce adaptive explanations. We present a reinforcement learning approach for scientific explanation generation that incorporates agentic personas, structured representations of expert reasoning strategies, that guide the explanation agent towards specific epistemic preferences. In an evaluation of knowledge graph-based explanations for drug discovery, we tested two personas that capture distinct epistemic stances derived from expert feedback. Results show that persona-driven explanations match state-of-the-art predictive performance while persona preferences closely align with those of their corresponding experts. Adaptive explanations were consistently preferred over non-adaptive baselines (n = 22), and persona-based training reduces feedback requirements by two orders of magnitude. These findings demonstrate how agentic personas enable scalable adaptive explainability for AI systems in complex and high-stakes domains.

Learning causal relations from observational data is a fundamental problem with wide-ranging applications across many fields. Constraint-based methods infer the underlying causal structure by performing conditional independence tests. However, existing algorithms such as the prominent PC algorithm need to perform a large number of independence tests, which in the worst case is exponential in the maximum degree of the causal graph. Despite extensive research, it remains unclear if there exist algorithms with better complexity without additional assumptions. Here, we establish an algorithm that achieves a better complexity of $p^{\mathcal{O}(s)}$ tests, where $p$ is the number of nodes in the graph and $s$ denotes the maximum undirected clique size of the underlying essential graph. Complementing this result, we prove that any constraint-based algorithm must perform at least $2^{Ω(s)}$ conditional independence tests, establishing that our proposed algorithm achieves exponent-optimality up to a logarithmic factor in terms of the number of conditional independence tests needed. Finally, we validate our theoretical findings through simulations, on semi-synthetic gene-expression data, and real-world data, demonstrating the efficiency of our algorithm compared to existing methods in terms of number of conditional independence tests needed.

Human annotation cost and time remain significant bottlenecks in Natural Language Processing (NLP), with test data annotation being particularly expensive due to the stringent requirement for low-error and high-quality labels necessary for reliable model evaluation. Traditional approaches require annotating entire test sets, leading to substantial resource requirements. Active Testing is a framework that selects the most informative test samples for annotation. Given a labeling budget, it aims to choose the subset that best estimates model performance while minimizing cost and human effort. In this work, we formalize Active Testing in NLP and we conduct an extensive benchmarking of existing approaches across 18 datasets and 4 embedding strategies spanning 4 different NLP tasks. The experiments show annotation reductions of up to 95%, with performance estimation accuracy difference from the full test set within 1%. Our analysis reveals variations in method effectiveness across different data characteristics and task types, with no single approach emerging as universally superior. Lastly, to address the limitation of requiring a predefined annotation budget in existing sample selection strategies, we introduce an adaptive stopping criterion that automatically determines the optimal number of samples.

A fundamental but largely unaddressed obstacle in Symbolic regression (SR) is structural redundancy: every expression DAG with admits many distinct node-numbering schemes that all encode the same expression, each occupying a separate point in the search space and consuming fitness evaluations without adding diversity. We present IsalSR (Instruction Set and Language for Symbolic Regression), a representation framework that encodes expression DAGs as strings over a compact two-tier alphabet and computes a pruned canonical string -- a complete labeled-DAG isomorphism invariant -- that collapses all the equivalent representations into a single canonical form.

Photoplethysmography (PPG) is one of the most widely captured biosignals for clinical prediction tasks, yet PPG-based algorithms are typically trained on small-scale datasets of uncertain quality, which hinders meaningful algorithm comparisons. We present a comprehensive benchmark for PPG-based clinical prediction using the \dbname~dataset, establishing baselines across the full spectrum of clinically relevant applications: multi-class heart rhythm classification, and regression of physiological parameters including respiratory rate (RR), heart rate (HR), and blood pressure (BP). Most notably, we provide the first comprehensive assessment of PPG for general arrhythmia detection beyond atrial fibrillation (AF) and atrial flutter (AFLT), with performance stratified by BP, HR, and demographic subgroups. Using established deep learning architectures, we achieved strong performance for AF detection (AUROC = 0.96) and accurate physiological parameter estimation (RR MAE: 2.97 bpm; HR MAE: 1.13 bpm; SBP/DBP MAE: 16.13/8.70 mmHg). Cross-dataset validation demonstrates excellent generalizability for AF detection (AUROC = 0.97), while clinical subgroup analysis reveals marked performance differences across subgroups by BP, HR, and demographic strata. These variations appear to reflect population-specific waveform differences rather than systematic bias in model behavior. This framework establishes the first integrated benchmark for multi-task PPG-based clinical prediction, demonstrating that PPG signals can effectively support multiple simultaneous monitoring tasks and providing essential baselines for future algorithm development.

Most existing Time Series Foundation Models (TSFMs) use channel independent modeling and focus on capturing and generalizing temporal dependencies, while neglecting the correlations among channels or overlooking the different aspects of correlations. However, these correlations play a vital role in Multivariate time series forecasting. To address this, we propose a CoRrelation-aware Adapter (CoRA), a lightweight plug-and-play method that requires only fine-tuning with TSFMs and is able to capture different types of correlations, so as to improve forecast performance. Specifically, to reduce complexity, we innovatively decompose the correlation matrix into low-rank Time-Varying and Time-Invariant components. For the Time-Varying component, we further design learnable polynomials to learn dynamic correlations by capturing trends or periodic patterns. To learn positive and negative correlations that appear only among some channels, we introduce a novel dual contrastive learning method that identifies correlations through projection layers, regulated by a Heterogeneous-Partial contrastive loss during training, without introducing additional complexity in the inference stage. Extensive experiments on 10 real-world datasets demonstrate that CoRA can improve TSFMs in multivariate forecasting performance.

Evaluating badminton performance often requires expert coaching, which is rarely accessible for amateur players. We present adminSense, a smartwatch-based system for fine-grained badminton performance analysis using wearable sensing. Through interviews with experienced badminton players, we identified four system design requirements with three implementation insights that guide the development of BadminSense. We then collected a badminton strokes dataset on 12 experienced badminton amateurs and annotated it with fine-grained labels, including stroke type, expert-assessed stroke rating, and shuttle impact location. Built on this dataset, BadminSense segments and classifies strokes, predicts stroke quality, and estimates shuttle impact location using vibration signal from an off-the-shelf smartwatch. Our evaluations show that

Steel surface defect detection is essential for ensuring product quality and reliability in modern manufacturing. Current methods often rely on basic image classification models trained on label-only datasets, which limits their interpretability and generalization. To address these challenges, we introduce SteelDefectX, a vision-language dataset containing 7,778 images across 25 defect categories, annotated with coarse-to-fine textual descriptions. At the coarse-grained level, the dataset provides class-level information, including defect categories, representative visual attributes, and associated industrial causes. At the fine-grained level, it captures sample-specific attributes, such as shape, size, depth, position, and contrast, enabling models to learn richer and more detailed defect representations. We further establish a benchmark comprising four tasks, vision-only classification, vision-language classification, few/zero-shot recognition, and zero-shot transfer, to evaluate model performance and generalization. Experiments with several baseline models demonstrate that coarse-to-fine textual annotations significantly improve interpretability, generalization, and transferability. We hope that SteelDefectX will serve as a valuable resource for advancing research on explainable, generalizable steel surface defect detection. The data will be publicly available on https://github.com/Zhaosxian/SteelDefectX.

Interrogatives in news discourse have been examined in linguistics and conversation analysis, but mostly in broadcast interviews and relatively small, often English-language corpora, while large-scale computational studies of news rarely distinguish interrogatives from declaratives or differentiate their functions. This paper brings these strands together through a mixed-methods study of the "Politics of Questions" in contemporary French-language digital news. Using over one million articles published between January 2023 and June 2024, we automatically detect interrogative stances, approximate their functional types, and locate textual answers when present, linking these quantitative measures to a qualitatively annotated subcorpus grounded in semantic and pragmatic theories of questions. Interrogatives are sparse but systematically patterned: they mainly introduce or organize issues, with most remaining cases being information-seeking or echo-like, while explicitly leading or tag questions are rare. Although their density and mix vary across outlets and topics, our heuristic suggests that questions are overwhelmingly taken up within the same article and usually linked to a subsequent answer-like span, most often in the journalist's narrative voice and less often through quoted speech. Interrogative contexts are densely populated with named individuals, organizations, and places, whereas publics and broad social groups are mentioned much less frequently, suggesting that interrogative discourse tends to foreground already prominent actors and places and thus exhibits strong personalization. We show how interrogative stance, textual uptake, and voice can be operationalized at corpus scale, and argue that combining computational methods with pragmatic and sociological perspectives can help account for how questioning practices structure contemporary news discourse.

We introduce ControlAugment (Ctrl-A), an automated data augmentation algorithm for image-vision tasks, which incorporates principles from control theory for online adjustment of augmentation strength distributions during model training. Ctrl-A eliminates the need for initialization of individual augmentation strengths. Instead, augmentation strength distributions are dynamically, and individually, adapted during training based on a control-loop architecture and what we define as relative operation response curves. Using an operation-dependent update procedure provides Ctrl-A with the potential to suppress augmentation styles that negatively impact model performance, alleviating the need for manually engineering augmentation policies for new image-vision tasks. Experiments on the CIFAR-10, CIFAR-100, and SVHN-core benchmark datasets using the common WideResNet-28-10 architecture demonstrate that Ctrl-A is highly competitive with existing state-of-the-art data augmentation strategies.

Attention mechanisms excel at learning sequential patterns by discriminating data based on relevance and importance. This provides state-of-the-art performance in advanced generative artificial intelligence models. This paper applies this concept of an attention mechanism for multi-agent safe control. We specifically consider the design of a neural network to control autonomous vehicles in a highway merging scenario. The environment is modeled as a Decentralized Partially Observable Markov Decision Process (Dec-POMDP). Within a QMIX framework, we include partial attention for each autonomous vehicle, thus allowing each ego vehicle to focus on the most relevant neighboring vehicles. Moreover, we propose a comprehensive reward signal that considers the global objectives of the environment (e.g., safety and vehicle flow) and the individual interests of each agent. Simulations are conducted in the Simulation of Urban Mobility (SUMO). The results show better performance compared to other driving algorithms in terms of safety, driving speed, and reward.

We propose a method for reasoning about trust in multi-agent systems, specifying a language for describing communication protocols and making trust assumptions and derivations. This is given an interpretation in a modal logic for describing the beliefs and communications of agents in a network. We define how information in the network can be shared via forwarding, and how trust between agents can be generalized to trust across networks. We give specifications for the modal logic which can be readily adapted into a lambda calculus of proofs. We show that by nesting modalities, we can describe chains of communication between agents, and establish suitable notions of trust for such chains. We see how this can be applied to trust models in public key infrastructures, as well as other interaction protocols in distributed systems.

Convolutional neural networks (CNNs) require a large number of multiply-accumulate (MAC) operations. To meet real-time constraints, they often need to be executed on specialized accelerators composed of an on-chip memory and a processing unit. However, the on-chip memory is often insufficient to store all the data required to compute a CNN layer. Thus, the computation must be performed in several offloading steps. We formalise such sequences of steps and apply our formalism to a state of the art decomposition of convolutions. In order to find optimal strategies in terms of duration, we encode the problem with a set of constraints. A Python-based simulator allows to analyse in-depth computed strategies.

The performance of machine learning models is determined by the quality of their learned features. They should be invariant under irrelevant data variation but sensitive to task-relevant details. To visualize whether this is the case, we propose a method to analyze feature extractors by sampling from their fibers -- equivalence classes defined by their invariances -- given an arbitrary representative. Unlike existing work where a dedicated generative model is trained for each feature detector, our algorithm is training-free and exploits a pretrained diffusion or flow-matching model as a prior. The fiber loss -- which penalizes mismatch in features -- guides the denoising process toward the desired equivalence class, via non-linear diffusion trajectory matching. This replaces days of training for invariance learning with a single guided generation procedure at comparable fidelity. Experiments on popular datasets (ImageNet, CheXpert) and model types (ResNet, DINO, BiomedClip) demonstrate that our framework can reveal invariances ranging from very desirable to concerning behaviour. For instance, we show how Qwen-2B places patients with situs inversus (heart on the right side) in the same fiber as typical anatomy.

This manuscript presents a comprehensive analysis of predictive modeling optimization in managed Wi-Fi networks through the integration of clustering algorithms and model evaluation techniques. The study addresses the challenges of deploying forecasting algorithms in large-scale environments managed by a central controller constrained by memory and computational resources. Feature-based clustering, supported by Principal Component Analysis (PCA) and advanced feature engineering, is employed to group time series data based on shared characteristics, enabling the development of cluster-specific predictive models. Comparative evaluations between global models (GMs) and cluster-specific models demonstrate that cluster-specific models consistently achieve superior accuracy in terms of Mean Absolute Error (MAE) values in high-activity clusters. The trade-offs between model complexity (and accuracy) and resource utilization are analyzed, highlighting the scalability of tailored modeling approaches. The findings advocate for adaptive network management strategies that optimize resource allocation through selective model deployment, enhance predictive accuracy, and ensure scalable operations in large-scale, centrally managed Wi-Fi environments.

Energy efficiency has become a growing concern in software development, leading to the need for tools designed to measure energy consumption. While several energy measurement tools are available as open-source projects, their characteristics and adoption remain underexplored. This work presents an empirical study based on a Mining Software Repositories (MSR) approach to identify, classify, and analyze software energy monitoring tools publicly available on GitHub. We qualitatively analyzed an initial dataset of 585 repositories to identify key design aspects, including measurement granularity and underlying design principles. After this analysis, we retained 24 repositories as relevant energy measuring software tools. The qualitative analysis we conduct reveals a clear evolution from early CPU-centric and machine-level monitoring utilities toward more diverse tools that support multi-level granularity (process, container, and AI workload levels) and integrate emission estimation capabilities. This study provides the first structured overview of open-source energy and emission measurement tools from an MSR perspective, which may be beneficial for software architects when designing energy-aware software.

Accurate and reliable safety metrics are paramount for functional safety verification of ASICs in automotive systems. Traditional FMEDA (Failure Modes, Effects, and Diagnostic Analysis) metrics, such as SPFM (Single Point Fault Metric) and LFM (Latent Fault Metric), depend on the precision of failure mode distribution (FMD) and diagnostic coverage (DC) estimations. This reliance can often leads to significant, unquantified uncertainties and a dependency on expert judgment, compromising the quality of the safety analysis. This paper proposes a novel approach that introduces error propagation theory into the calculation of FMEDA safety metrics. By quantifying the maximum deviation and providing confidence intervals for SPFM and LFM, our method offers a direct measure of analysis quality. Furthermore, we introduce an Error Importance Identifier (EII) to pinpoint the primary sources of uncertainty, guiding targeted improvements. This approach significantly enhances the transparency and trustworthiness of FMEDA, enabling more robust ASIC safety verification for ISO 26262 compliance, addressing a longstanding open question in the functional safety community.

Precipitation nowcasting is critical for disaster mitigation and aviation safety. However, radar-only models frequently suffer from a lack of large-scale atmospheric context, leading to performance degradation at longer lead times. While integrating meteorological variables predicted by weather foundation models offers a potential remedy, existing architectures fail to reconcile the profound representational heterogeneities between radar imagery and meteorological data. To bridge this gap, we propose PW-FouCast, a novel frequency-domain fusion framework that leverages Pangu-Weather forecasts as spectral priors within a Fourier-based backbone. Our architecture introduces three key innovations: (i) Pangu-Weather-guided Frequency Modulation to align spectral magnitudes and phases with meteorological priors; (ii) Frequency Memory to correct phase discrepancies and preserve temporal evolution; and (iii) Inverted Frequency Attention to reconstruct high-frequency details typically lost in spectral filtering. Extensive experiments on the SEVIR and MeteoNet benchmarks demonstrate that PW-FouCast achieves state-of-the-art performance, effectively extending the reliable forecast horizon while maintaining structural fidelity. Our code is available at https://github.com/Onemissed/PW-FouCast.

Deformable image registration plays a fundamental role in medical image analysis by enabling spatial alignment of anatomical structures across subjects. While recent deep learning-based approaches have significantly improved computational efficiency, many existing methods remain limited in capturing long-range anatomical correspondence and maintaining deformation consistency. In this work, we present a cycle inverse-consistent transformer-based framework for deformable brain MRI registration. The model integrates a Swin-UNet architecture with bidirectional consistency constraints, enabling the joint estimation of forward and backward deformation fields. This design allows the framework to capture both local anatomical details and global spatial relationships while improving deformation stability. We conduct a comprehensive evaluation of the proposed framework on a large multi-center dataset consisting of 2851 T1-weighted brain MRI scans aggregated from 13 public datasets. Experimental results demonstrate that the proposed framework achieves strong and balanced performance across multiple quantitative evaluation metrics while maintaining stable and physically plausible deformation fields. Detailed quantitative comparisons with baseline methods, including ANTs, ICNet, and VoxelMorph, are provided in the appendix. Experimental results demonstrate that CICTM achieves consistently strong performance across multiple evaluation criteria while maintaining stable and physically plausible deformation fields. These properties make the proposed framework suitable for large-scale neuroimaging datasets where both accuracy and deformation stability are critical.

Recently, Interleaved-modal Chain-of-Thought (ICoT) reasoning has achieved remarkable success by leveraging both multimodal inputs and outputs, attracting increasing attention. While achieving promising performance, current ICoT methods still suffer from two major limitations: (1) Static Visual Thought Positioning, which statically inserts visual information at fixed steps, resulting in inefficient and inflexible reasoning; and (2) Broken Visual Thought Representation, which involves discontinuous and semantically incoherent visual tokens. To address these limitations, we introduce Interleaved-modal Chain-of-Thought reasoning with Dynamic and Precise Visual Thoughts (DaP-ICoT), which incorporates two key components: (1) Dynamic Visual Thought Integration adaptively introduces visual inputs based on reasoning needs, reducing redundancy and improving efficiency. (2) Precise Visual Thought Guidance ensures visual semantically coherent and contextually aligned representations. Experiments across multiple benchmarks and models demonstrate that DaP-ICoT achieves state-of-the-art performance. In addition, DaP-ICoT significantly reduces the number of inserted images, leading to a 72.6% decrease in token consumption, enabling more efficient ICoT reasoning.

Bayesian inference is a powerful tool for parameter estimation and uncertainty quantification in dynamical systems. However, for nonlinear oscillator networks such as Kuramoto models, widely used to study synchronization phenomena in physics, biology, and engineering, inference is often computationally prohibitive due to high-dimensional state spaces and intractable likelihood functions. We present an amortized Bayesian inference approach that learns a neural approximation of the posterior from simulated phase dynamics, enabling fast, scalable inference without repeated sampling or optimization. Applied to synthetic Kuramoto networks, the method shows promising results in approximating posterior distributions and capturing uncertainty, with computational savings compared to traditional Bayesian techniques. These findings suggest that amortized inference is a practical and flexible framework for uncertainty-aware analysis of oscillator networks.

Quantum machine learning models generally lack principled design guidelines, often requiring full resource-intensive training across numerous choices of encodings, quantum circuit designs and initialization strategies to find effective configuration. To address this challenge, we develope the Quantum Bias-Expressivity Toolbox ($\texttt{QBET}$), a framework for evaluating quantum, classical, and hybrid transformer architectures. In this toolbox, we introduce lean metrics for Simplicity Bias ($\texttt{SB}$) and Expressivity ($\texttt{EXP}$), for comparing across various models, and extend the analysis of $\texttt{SB}$ to generative and multiclass-classification tasks. We show that $\texttt{QBET}$ enables efficient pre-screening of promising model variants obviating the need to execute complete training pipelines. In evaluations on transformer-based classification and generative tasks we employ a total of $18$ qubits for embeddings ($6$ qubits each for query, key, and value). We identify scenarios in which quantum self-attention variants surpass their classical counterparts by ranking the respective models according to the $\texttt{SB}$ metric and comparing their relative performance.

Large language models are the first systems to achieve high cognitive performance without clearly undergoing representation genesis: the transition from a non-representing physical system to one whose states guide behavior in a content-sensitive way. Prior cognitive systems had already made this transition before we could examine it, and philosophy of mind treated genesis as a background condition rather than an explanatory target. LLMs provide a case that does not clearly involve this transition, making the genesis question newly urgent: if genesis did not occur, which cognitive capacities are affected, and why? We currently lack the conceptual resources to answer this. The reason, this paper argues, is structural. Major frameworks in philosophy of mind, including the Language of Thought hypothesis, teleosemantics, predictive processing, enactivism, and genetic phenomenology, share a common feature when applied to the genesis question: at some explanatory step, each deploys concepts whose explanatory purchase depends on the system already being organized as a representer. This pattern, which we call the Representation Presupposition structure, generates systematic explanatory deferral. Attempts to explain the first acquisition of content-manipulable representation within the existing categorical vocabulary import resources from the representational side of the transition itself. We call this the Representation Regress. The paper offers a conceptual diagnosis rather than a new theory, establishing the structure of the problem and deriving two minimum adequacy conditions for any account that avoids this pattern. LLMs make the absence of such a theory consequential rather than merely theoretical.

Building virtual cells with generative models to simulate cellular behavior in silico is emerging as a promising paradigm for accelerating drug discovery. However, prior image-based generative approaches can produce implausible cell images that violate basic physical and biological constraints. To address this, we propose to post-train virtual cell models with reinforcement learning (RL), leveraging biologically meaningful evaluators as reward functions. We design seven rewards spanning three categories-biological function, structural validity, and morphological correctness-and optimize the state-of-the-art CellFlux model to yield CellFluxRL. CellFluxRL consistently improves over CellFlux across all rewards, with further performance boosts from test-time scaling. Overall, our results present a virtual cell modeling framework that enforces physically-based constraints through RL, advancing beyond "visually realistic" generations towards "biologically meaningful" ones.

Different types of reasoning impose different structural demands on representational systems, yet no systematic account of these demands exists across psychology, AI, and philosophy of mind. I propose a framework identifying four structural properties of representational systems: operability, consistency, structural preservation, and compositionality. These properties are demanded to different degrees by different forms of reasoning, from induction through analogy and causal inference to deduction and formal logic. Each property excludes a distinct class of reasoning failure. The analysis reveals a principal structural boundary: reasoning types below it can operate on associative, probabilistic representations, while those above it require all four properties to be fully satisfied. Scaling statistical learning without structural reorganization is insufficient to cross this boundary, because the structural guarantees required by deductive reasoning cannot be approximated through probabilistic means. Converging evidence from AI evaluation, developmental psychology, and cognitive neuroscience supports the framework at different levels of directness. Three testable predictions are derived, including compounding degradation, selective vulnerability to targeted structural disruption, and irreducibility under scaling. The framework is a necessary-condition account, agnostic about representational format, that aims to reorganize existing debates rather than close them.

The proliferation of Generative Artificial Intelligence has transformed benign cognitive offloading into a systemic risk of cognitive agency surrender. Driven by the commercial dogma of "zero-friction" design, highly fluent AI interfaces actively exploit human cognitive miserliness, prematurely satisfying the need for cognitive closure and inducing severe automation bias. To empirically quantify this epistemic erosion, we deployed a zero-shot semantic classification pipeline ($τ=0.7$) on 1,223 high-confidence AI-HCI papers from 2023 to early 2026. Our analysis reveals an escalating "agentic takeover": a brief 2025 surge in research defending human epistemic sovereignty (19.1%) was abruptly suppressed in early 2026 (13.1%) by an explosive shift toward optimizing autonomous machine agents (19.6%), while frictionless usability maintained a structural hegemony (67.3%). To dismantle this trap, we theorize "Scaffolded Cognitive Friction," repurposing Multi-Agent Systems (MAS) as explicit cognitive forcing functions (e.g., computational Devil's Advocates) to inject germane epistemic tension and disrupt heuristic execution. Furthermore, we outline a multimodal computational phenotyping agenda -- integrating gaze transition entropy, task-evoked pupillometry, fNIRS, and Hierarchical Drift Diffusion Modeling (HDDM) -- to mathematically decouple decision outcomes from cognitive effort. Ultimately, intentionally designed friction is not merely a psychological intervention, but a foundational technical prerequisite for enforcing global AI governance and preserving societal cognitive resilience.

Scientific idea generation is a cornerstone of autonomous knowledge discovery, yet the iterative evolution required to transform initial concepts into high-quality research proposals remains a formidable challenge for Large Language Models (LLMs). Existing Reinforcement Learning (RL) paradigms often rely on rubric-based scalar rewards that provide global quality scores but lack actionable granularity. Conversely, language-based refinement methods are typically confined to inference-time prompting, targeting models that are not explicitly optimized to internalize such critiques. To bridge this gap, we propose \textbf{EvoIdeator}, a framework that facilitates the evolution of scientific ideas by aligning the RL training objective with \textbf{checklist-grounded feedback}. EvoIdeator leverages a structured judge model to generate two synergistic signals: (1) \emph{lexicographic rewards} for multi-dimensional optimization, and (2) \emph{fine-grained language feedback} that offers span-level critiques regarding grounding, feasibility, and methodological rigor. By integrating these signals into the RL loop, we condition the policy to systematically utilize precise feedback during both optimization and inference. Extensive experiments demonstrate that EvoIdeator, built on Qwen3-4B, significantly outperforms much larger frontier models across key scientific metrics. Crucially, the learned policy exhibits strong generalization to diverse external feedback sources without further fine-tuning, offering a scalable and rigorous path toward self-refining autonomous ideation.

Fine-tuning large language models (LLMs) with backpropagation achieves high performance but incurs substantial memory overhead, limiting scalability on resource-constrained hardware. Zeroth-order (ZO) optimization provides a memory-efficient alternative by relying solely on forward passes, yet it typically suffers from slow or unstable convergence due to high-variance gradient estimates. Sparse ZO updates partially address this issue by perturbing only a subset of parameters, but their effectiveness hinges on selecting informative parameters, which is challenging in ZO optimization because each query yields only scalar feedback. We propose \textbf{Adaptive Curvature-Guided Sparse Zeroth-Order Optimization (CurvZO)}, which tracks curvature signals online from scalar ZO feedback and leverages these signals to construct a parameter-wise sampling distribution for selecting coordinates at each update, reducing the variance of the sparse ZO gradient estimator. Moreover, CurvZO dynamically adapts the perturbation budget to the evolving curvature signal distribution, yielding sparse ZO updates that remain both focused and sufficiently exploratory. Extensive experiments on OPT and Llama across diverse NLP tasks show that CurvZO consistently improves fine-tuning performance and reduces training time over ZO baselines. It improves accuracy by up to 4.4 points and achieves up to a $2\times$ speedup, while preserving memory efficiency.

Transformers have achieved remarkable progress in time series forecasting, yet their reliance on deterministic dot-product attention limits their capacity to model uncertainty and nonlinear dependencies across multivariate temporal dimensions. To address this limitation, we propose FISFormer, a Fuzzy Inference System-driven Transformer that replaces conventional attention with a FIS Interaction mechanism. In this framework, each query-key pair undergoes a fuzzy inference process for every feature dimension, where learnable membership functions and rule-based reasoning estimate token-wise relational strengths. These FIS-derived interaction weights capture uncertainty and provide interpretable, continuous mappings between tokens. A softmax operation is applied along the token axis to normalize these weights, which are then combined with the corresponding value features through element-wise multiplication to yield the final context-enhanced token representations. This design fuses the interpretability and uncertainty modeling of fuzzy logic with the representational power of Transformers. Extensive experiments on multiple benchmark datasets demonstrate that FISFormer achieves superior forecasting accuracy, noise robustness, and interpretability compared to state-of-the-art Transformer variants, establishing fuzzy inference as an effective alternative to conventional attention mechanisms.

We present Triple Zero Path Planning (TZPP), a collaborative framework for heterogeneous multi-robot systems that requires zero training, zero prior knowledge, and zero simulation. TZPP employs a coordinator--explorer architecture: a humanoid robot handles task coordination, while a quadruped robot explores and identifies feasible paths using guidance from a multimodal large language model. We implement TZPP on Unitree G1 and Go2 robots and evaluate it across diverse indoor and outdoor environments, including obstacle-rich and landmark-sparse settings. Experiments show that TZPP achieves robust, human-comparable efficiency and strong adaptability to unseen scenarios. By eliminating reliance on training and simulation, TZPP offers a practical path toward real-world deployment of heterogeneous robot cooperation. Our code and video are provided at: https://github.com/triple-zeropp/Triple-zero-robot-agent

Understanding why real-world events occur is important for both natural language processing and practical decision-making, yet direct-cause inference remains underexplored in evidence-rich settings. To address this gap, we organized SemEval-2026 Task 12: Abductive Event Reasoning (AER).\footnote{The task data is available at https://github.com/sooo66/semeval2026-task12-dataset.git} The task asks systems to identify the most plausible direct cause of a target event from supporting evidence. We formulate AER as an evidence-grounded multiple-choice benchmark that captures key challenges of real-world causal reasoning, including distributed evidence, indirect background factors, and semantically related but non-causal distractors. The shared task attracted 122 participants and received 518 submissions. This paper presents the task formulation, dataset construction pipeline, evaluation setup, and system results. AER provides a focused benchmark for abductive reasoning over real-world events and highlights challenges for future work on causal reasoning and multi-document understanding.

As real-world tasks grow increasingly complex, long-context reasoning has become a core capability for Large Language Models (LLMs). However, few studies explore which data types are effective for long-context reasoning and why. We find that structured table data with periodic structures shows strong potential for long-context reasoning. Motivated by this observation, we mathematically analyze tabular dependency structures using mutual information, revealing periodic non-vanishing dependencies in table data. Furthermore, we systematically analyze the capabilities of structured table data, conduct relevant scaling experiments, and validate its underlying mechanisms for enhancing long-context reasoning, yielding several meaningful insights. Leveraging these insights, we propose a simple yet scalable pipeline(TableLong) for synthesizing high-quality, diverse, and verifiable structured table data to boost long-context reasoning via RL. Extensive experimental results demonstrate that table data significantly enhances the long-context reasoning capability of LLMs across multiple long-context benchmarks (+8.24\% on average), and even improves performance on out-of-domain benchmarks (+8.06\% on average). We hope that our insights provide practical guidance for effective post-training data to enhance long-context reasoning in LLMs.

Distribution-to-distribution generative models support scientific imaging tasks ranging from modeling cellular perturbation responses to translating medical images across conditions. Trustworthy generation requires both reliability (generalization across labs, devices, and experimental conditions) and accountability (detecting out-of-distribution cases where predictions may be unreliable). Uncertainty quantification (UQ) based approaches serve as promising candidates for these tasks, yet UQ for distribution-to-distribution generative models remains underexplored. We present a unified UQ framework, Bayesian Stochastic Flow Matching (BSFM), that disentangles aleatoric and epistemic uncertainty. The Stochastic Flow Matching (SFM) component augments deterministic flows with a diffusion term to improve model generalization to unseen scenarios. For UQ, we develop a scalable Bayesian approach -- MCD-Antithetic -- that combines Monte Carlo Dropout with sample-efficient antithetic sampling to produce effective anomaly scores for out-of-distribution detection. Experiments on cellular imaging (BBBC021, JUMP) and brain fMRI (Theory of Mind) across diverse scenarios show that SFM improves reliability while MCD-Antithetic enhances accountability.

Efficient selection among multiple generative models is increasingly important in modern generative AI, where sampling from suboptimal models is costly. This problem can be formulated as a multi-armed bandit task. Under diversity-aware evaluation metrics, a non-degenerate mixture of generators can outperform any individual model, distinguishing this setting from classical best-arm identification. Prior approaches therefore incorporate an Upper Confidence Bound (UCB) exploration bonus into the mixture objective. However, across multiple datasets and evaluation metrics, we observe that the UCB term consistently slows convergence and often reduces sample efficiency. In contrast, a simple \emph{Mixture-Greedy} strategy without explicit UCB-type optimism converges faster and achieves even better performance, particularly for widely used metrics such as FID and Vendi where tight confidence bounds are difficult to construct. We provide theoretical insight explaining this behavior: under transparent structural conditions, diversity-aware objectives induce implicit exploration by favoring interior mixtures, leading to linear sampling of all arms and sublinear regret guarantees for entropy-based, kernel-based, and FID-type objectives. These results suggest that in diversity-aware multi-armed bandits for generative model selection, exploration can arise intrinsically from the objective geometry, questioning the necessity of explicit confidence bonuses.

The transition to Electric Vehicles (EVs) demands intelligent, congestion-aware infrastructure planning to balance user convenience, economic viability, and traffic efficiency. We present a joint optimisation framework for EV Charging Station (CS) placement and pricing, explicitly capturing strategic driver behaviour through coupled non-atomic congestion games over road networks and charging facilities. From a Public Authority (PA) perspective, the model minimises social cost, travel times, queuing delays and charging expenses, while ensuring infrastructure profitability. To solve the resulting Mixed-Integer Nonlinear Programme, we propose a scalable two-level approximation method, Joint Placement and Pricing Optimisation under Driver Equilibrium (JPPO-DE), combining driver behaviour decomposition with integer relaxation. Experiments on the benchmark Sioux Falls Transportation Network (TN) demonstrate that our method consistently outperforms single-parameter baselines, effectively adapting to varying budgets, EV penetration levels, and station capacities. It achieves performance improvements of at least 16% over state-of-the-art approaches. A generalisation procedure further extends scalability to larger networks. By accurately modelling traffic equilibria and enabling adaptive, efficient infrastructure design, our framework advances key intelligent transportation system goals for sustainable urban mobility.

Long-tail class incremental learning (LT CIL) remains highly challenging because the scarcity of samples in tail classes not only hampers their learning but also exacerbates catastrophic forgetting under continuously evolving and imbalanced data distributions. To tackle these issues, we exploit the informativeness and scalability of language knowledge. Specifically, we analyze the LT CIL data distribution to guide large language models (LLMs) in generating a stratified language tree that hierarchically organizes semantic information from coarse to fine grained granularity. Building upon this structure, we introduce stratified adaptive language guidance, which leverages learnable weights to merge multi-scale semantic representations, thereby enabling dynamic supervisory adjustment for tail classes and alleviating the impact of data imbalance. Furthermore, we introduce stratified alignment language guidance, which exploits the structural stability of the language tree to constrain optimization and reinforce semantic visual alignment, thereby alleviating catastrophic forgetting. Extensive experiments on multiple benchmarks demonstrate that our method achieves state of the art performance.

Model merging has emerged as a practical approach to combine capabilities of specialized large language models (LLMs) without additional training. In the Long-to-Short (L2S) scenario, merging a base model with a long-chain-of-thought reasoning model aims to preserve reasoning accuracy while reducing output length. Existing methods rely on Task Arithmetic and its variants, which implicitly assume that model outputs vary linearly with the merging coefficient -- an assumption we show is systematically violated in L2S settings. We provide the first theoretical justification for layer-adaptive merging: we prove that merging error is bounded by a term proportional to the per-layer Hessian norm (Proposition~1), and establish that the Fisher Information Matrix (FIM) is a principled, computable proxy for this bound via the Fisher-Hessian equivalence at local optima. Building on this theory, we propose \textbf{FIM-Merging}, which computes diagonal FIM using only random token inputs (no domain-specific calibration data required) and uses it to assign per-layer merging coefficients. On the 7B L2S benchmark, FIM-TIES achieves state-of-the-art performance on five out of six evaluation benchmarks, including a \textbf{+6.2} point gain on MATH500 over ACM-TIES (90.2 vs.\ 84.0), while requiring no calibration data. On the 1.5B benchmark, FIM-TIES achieves an average accuracy of \textbf{47.3}, surpassing the previous best ACM-TIES (43.3) by \textbf{+3.9} points, while reducing average response length by \textbf{91.9\%} relative to the long-CoT model. Our framework also provides a unified theoretical explanation for why existing layer-adaptive methods such as ACM empirically outperform uniform merging.

Large Vision-Language Models (LVLMs) excel in visual understanding and reasoning, but the excessive visual tokens lead to high inference costs. Although recent token reduction methods mitigate this issue, they mainly target single-turn Visual Question Answering (VQA), leaving the more practical multi-turn VQA (MT-VQA) scenario largely unexplored. MT-VQA introduces additional challenges, as subsequent questions are unknown beforehand and may refer to arbitrary image regions, making existing reduction strategies ineffective. Specifically, current approaches fall into two categories: prompt-dependent methods, which bias toward the initial text prompt and discard information useful for subsequent turns; prompt-agnostic ones, which, though technically applicable to multi-turn settings, rely on heuristic reduction metrics such as attention scores, leading to suboptimal performance. In this paper, we propose a learning-based prompt-agnostic method, termed MetaCompress, overcoming the limitations of heuristic designs. We begin by formulating token reduction as a learnable compression mapping, unifying existing formats such as pruning and merging into a single learning objective. Upon this formulation, we introduce a data-efficient training paradigm capable of learning optimal compression mappings with limited computational costs. Extensive experiments on MT-VQA benchmarks and across multiple LVLM architectures demonstrate that MetaCompress achieves superior efficiency-accuracy trade-offs while maintaining strong generalization across dialogue turns. Our code is available at https://github.com/MArSha1147/MetaCompress.

High-fidelity vehicle drag evaluation is constrained less by solver runtime than by workflow friction: geometry cleanup, meshing retries, queue contention, and reproducibility failures across teams. We present a contract-centric blueprint for self-evolving coding agents that discover executable surrogate pipelines for predicting drag coefficient $C_d$ under industrial constraints. The method formulates surrogate discovery as constrained optimization over programs, not static model instances, and combines Famou-Agent-style evaluator feedback with population-based island evolution, structured mutations (data, model, loss, and split policies), and multi-objective selection balancing ranking quality, stability, and cost. A hard evaluation contract enforces leakage prevention, deterministic replay, multi-seed robustness, and resource budgets before any candidate is admitted. Across eight anonymized evolutionary operators, the best system reaches a Combined Score of 0.9335 with sign-accuracy 0.9180, while trajectory and ablation analyses show that adaptive sampling and island migration are primary drivers of convergence quality. The deployment model is explicitly ``screen-and-escalate'': surrogates provide high-throughput ranking for design exploration, but low-confidence or out-of-distribution cases are automatically escalated to high-fidelity CFD. The resulting contribution is an auditable, reusable workflow for accelerating aerodynamic design iteration while preserving decision-grade reliability, governance traceability, and safety boundaries.

Multimodal Large Language Models (MLLMs) extend text-only LLMs with visual reasoning, but also introduce new safety failure modes under visually grounded instructions. We study comic-template jailbreaks that embed harmful goals inside simple three-panel visual narratives and prompt the model to role-play and "complete the comic." Building on JailbreakBench and JailbreakV, we introduce ComicJailbreak, a comic-based jailbreak benchmark with 1,167 attack instances spanning 10 harm categories and 5 task setups. Across 15 state-of-the-art MLLMs (six commercial and nine open-source), comic-based attacks achieve success rates comparable to strong rule-based jailbreaks and substantially outperform plain-text and random-image baselines, with ensemble success rates exceeding 90% on several commercial models. Then, with the existing defense methodologies, we show that these methods are effective against the harmful comics, they will induce a high refusal rate when prompted with benign prompts. Finally, using automatic judging and targeted human evaluation, we show that current safety evaluators can be unreliable on sensitive but non-harmful content. Our findings highlight the need for safety alignment robust to narrative-driven multimodal jailbreaks.

While Multi-Agent Debate (MAD) research has advanced, its efficacy in coordinating complex stakeholder interests such as travel planning remains largely unexplored. To bridge this gap, we propose MIND (Multi-agent Inference for Negotiation Dialogue), a framework designed to simulate realistic consensus-building among travelers with heterogeneous preferences. Grounded in the Theory of Mind (ToM), MIND introduces a Strategic Appraisal phase that infers opponent willingness (w) from linguistic nuances with 90.2% accuracy. Experimental results demonstrate that MIND outperforms traditional MAD frameworks, achieving a 20.5% improvement in High-w Hit and a 30.7% increase in Debate Hit-Rate, effectively prioritizing high-stakes constraints. Furthermore, qualitative evaluations via LLM-as-a-Judge confirm that MIND surpasses baselines in Rationality (68.8%) and Fluency (72.4%), securing an overall win rate of 68.3%. These findings validate that MIND effectively models human negotiation dynamics to derive persuasive consensus.

Multimodal large language models (MLLMs) have shown strong potential for medical Visual Question Answering (VQA), yet they remain prone to hallucinations, defined as generating responses that contradict the input image, posing serious risks in clinical settings. Current hallucination detection methods, such as Semantic Entropy (SE) and Vision-Amplified Semantic Entropy (VASE), require 10 to 20 stochastic generations per sample together with an external natural language inference model for semantic clustering, making them computationally expensive and difficult to deploy in practice. We observe that hallucinated responses exhibit a distinctive signature directly in the model's own log-probabilities: inconsistent token-level confidence and weak sensitivity to visual evidence. Based on this observation, we propose Confidence-Evidence Bayesian Gain (CEBaG), a deterministic hallucination detection method that requires no stochastic sampling, no external models, and no task-specific hyperparameters. CEBaG combines two complementary signals: token-level predictive variance, which captures inconsistent confidence across response tokens, and evidence magnitude, which measures how much the image shifts per-token predictions relative to text-only inference. Evaluated across four medical MLLMs and three VQA benchmarks (16 experimental settings), CEBaG achieves the highest AUC in 13 of 16 settings and improves over VASE by 8 AUC points on average, while being fully deterministic and self-contained. The code will be made available upon acceptance.

As AI agents transition from human-supervised copilots to autonomous platform infrastructure, the ability to analyze their reasoning behavior across populations of investigations becomes a pressing infrastructure requirement. Existing operational tooling addresses adjacent needs effectively: state checkpoint systems enable fault tolerance; observability platforms provide execution traces for debugging; telemetry standards ensure interoperability. What current systems do not natively provide as a first-class, schema-level primitive is structured reasoning provenance -- normalized, queryable records of why the agent chose each action, what it concluded from each observation, how each conclusion shaped its strategy, and which evidence supports its final verdict. This paper introduces the Agent Execution Record (AER), a structured reasoning provenance primitive that captures intent, observation, and inference as first-class queryable fields on every step, alongside versioned plans with revision rationale, evidence chains, structured verdicts with confidence scores, and delegation authority chains. We formalize the distinction between computational state persistence and reasoning provenance, argue that the latter cannot in general be faithfully reconstructed from the former, and show how AERs enable population-level behavioral analytics: reasoning pattern mining, confidence calibration, cross-agent comparison, and counterfactual regression testing via mock replay. We present a domain-agnostic model with extensible domain profiles, a reference implementation and SDK, and outline an evaluation methodology informed by preliminary deployment on a production platformized root cause analysis agent.

Real-world infrastructure planning increasingly involves strategic interactions among autonomous agents competing over congestible, limited resources. Applications such as Electric Vehicle (EV) charging, emergency response, and intelligent transportation require coordinated resource placement and pricing decisions, while anticipating the adaptive behaviour of decentralised, self-interested agents. We propose a novel multi-agent framework for joint placement and pricing under such interactions, formalised as a bi-level optimisation model. The upper level represents a central planner, while the lower level captures agent responses via coupled non-atomic congestion games. Motivated by the EV charging domain, we study a setting where a central planner provisions chargers and road capacity under budget and profitability constraints. The agent population includes both EV drivers and non-charging drivers (NCDs), who respond to congestion, delays, and costs. To solve the resulting NP-hard problem, we introduce ABO-MPN, a double-layer approximation framework that decouples agent types, applies integer adjustment and rounding, and targets high-impact placement and pricing decisions. Experiments on benchmark networks show that our model reduces social cost by up to 40% compared to placement- or pricing-only baselines, and generalises to other MAS-relevant domains.

As large language models (LLMs) and vision-language-action models (VLAs) become widely deployed, the tokens consumed by AI inference are evolving into a new type of commodity. This paper systematically analyzes the commodity attributes of tokens, arguing for their transition from intelligent service outputs to compute infrastructure raw materials, and draws comparisons with established commodities such as electricity, carbon emission allowances, and bandwidth. Building on the historical experience of electricity futures markets and the theory of commodity financialization, we propose a complete design for standardized token futures contracts, including the definition of a Standard Inference Token (SIT), contract specifications, settlement mechanisms, margin systems, and market-maker regimes. By constructing a mean-reverting jump-diffusion stochastic process model and conducting Monte Carlo simulations, we evaluate the hedging efficiency of the proposed futures contracts for application-layer enterprises. Simulation results show that, under an application-layer demand explosion scenario, token futures can reduce enterprise compute cost volatility by 62%-78%. We also explore the feasibility of GPU compute futures and discuss the regulatory framework for token futures markets, providing a theoretical foundation and practical roadmap for the financialization of compute resources.

Multimodal AI systems have achieved remarkable performance across a broad range of real-world tasks, yet the mechanisms underlying visual-language reasoning remain surprisingly poorly understood. We report three findings that challenge prevailing assumptions about how these systems process and integrate visual information. First, Frontier models readily generate detailed image descriptions and elaborate reasoning traces, including pathology-biased clinical findings, for images never provided; we term this phenomenon mirage reasoning. Second, without any image input, models also attain strikingly high scores across general and medical multimodal benchmarks, bringing into question their utility and design. In the most extreme case, our model achieved the top rank on a standard chest X-ray question-answering benchmark without access to any images. Third, when models were explicitly instructed to guess answers without image access, rather than being implicitly prompted to assume images were present, performance declined markedly. Explicit guessing appears to engage a more conservative response regime, in contrast to the mirage regime in which models behave as though images have been provided. These findings expose fundamental vulnerabilities in how visual-language models reason and are evaluated, pointing to an urgent need for private benchmarks that eliminate textual cues enabling non-visual inference, particularly in medical contexts where miscalibrated AI carries the greatest consequence. We introduce B-Clean as a principled solution for fair, vision-grounded evaluation of multimodal AI systems.

Hate speech online targets individuals or groups based on identity attributes and spreads rapidly, posing serious social risks. Memes, which combine images and text, have emerged as a nuanced vehicle for disseminating hate speech, often relying on cultural knowledge for interpretation. However, existing multimodal hate speech datasets suffer from coarse-grained labeling and a lack of integration with surrounding discourse, leading to imprecise and incomplete assessments. To bridge this gap, we propose an agentic annotation framework that coordinates seven specialized agents to generate hierarchical labels and rationales. Based on this framework, we construct M^3 (Multi-platform, Multi-lingual, and Multimodal Meme), a dataset of 2,455 memes collected from X, 4chan, and Weibo, featuring fine-grained hate labels and human-verified rationales. Benchmarking state-of-the-art Multimodal Large Language Models reveals that these models struggle to effectively utilize surrounding post context, which often fails to improve or even degrades detection performance. Our finding highlights the challenges these models face in reasoning over memes embedded in real-world discourse and underscores the need for a context-aware multimodal architecture. Our dataset and code are available at https://github.com/mira-ai-lab/M3.

The robustness of deep neural networks (DNNs) can be certified through their Lipschitz continuity, which has made the construction of Lipschitz-continuous DNNs an active research field. However, DNNs for audio processing have not been a major focus due to their poor compatibility with existing results. In this paper, we consider the amplitude modifier (AM), a popular architecture for handling audio signals, and propose its Lipschitz-continuous variants, which we refer to as LipsAM. We prove a sufficient condition for an AM to be Lipschitz continuous and propose two architectures as examples of LipsAM. The proposed architectures were applied to a Plug-and-Play algorithm for speech dereverberation, and their improved stability is demonstrated through numerical experiments.

Time-dependent reliability analysis of nonlinear dynamical systems under stochastic excitations is a critical yet computationally demanding task. Conventional approaches, such as Monte Carlo simulation, necessitate repeated evaluations of computationally expensive numerical solvers, leading to significant computational bottlenecks. To address this challenge, we propose \textit{CoNBONet}, a neuroscience-inspired surrogate model that enables fast, energy-efficient, and uncertainty-aware reliability analysis, providing a scalable alternative to techniques such as Monte Carlo simulations. CoNBONet, short for \textbf{Co}nformalized \textbf{N}euroscience-inspired \textbf{B}ayesian \textbf{O}perator \textbf{Net}work, leverages the expressive power of deep operator networks while integrating neuroscience-inspired neuron models to achieve fast, low-power inference. Unlike traditional surrogates such as Gaussian processes, polynomial chaos expansions, or support vector regression, that may face scalability challenges for high-dimensional, time-dependent reliability problems, CoNBONet offers \textit{fast and energy-efficient inference} enabled by a neuroscience-inspired network architecture, \textit{calibrated uncertainty quantification with theoretical guarantees} via split conformal prediction, and \textit{strong generalization capability} through an operator-learning paradigm that maps input functions to system response trajectories. Validation of the proposed CoNBONet for various nonlinear dynamical systems demonstrates that CoNBONet preserves predictive fidelity, and achieves reliable coverage of failure probabilities, making it a powerful tool for robust and scalable reliability analysis in engineering design.

Standard Transformers have a fixed computational depth, fundamentally limiting their ability to generalize to tasks requiring variable-depth reasoning, such as multi-hop graph traversal or nested logic. We propose a depth-recurrent Transformer that decouples computational depth from parameter count by iteratively applying a shared-weight Transformer block in latent space -- enabling the model to trade recurrence steps for deeper reasoning at inference time. Our architecture incorporates three mechanisms to make deep recurrence (20+ steps) stable: (1) a silent thinking objective that supervises only the final output, forcing genuine multi-step reasoning rather than intermediate heuristic shortcuts; (2) LayerScale initialization to protect fragile reasoning states from untrained layer noise; and (3) an identity-biased recurrence that creates a gradient highway across many steps. We evaluate on three compositional reasoning domains with decreasing inductive biases: graph reachability (strict adjacency masking), nested boolean logic (relative positioning), and unstructured relational text (where sequence position provides no structural hints). Across all tasks, we observe a clear \emph{computational frontier} -- a boundary where performance transitions from chance to near-perfect as thinking steps scale with task complexity. Moreover, these tasks reveal qualitatively different generalization behaviors: precise but brittle (graph), approximate but robust (logic), and autonomous latent routing without structural hints (text). This progression illuminates how the interplay between a task-invariant recurrent reasoning core and task-specific perceptual interfaces shapes out-of-distribution (OOD) generalization, offering a mechanistic perspective on vertical chain-of-thought that complements the prevailing horizontal token-generation paradigm.

Energy efficiency remains a critical challenge in deploying physics-informed operator learning models for computational mechanics and scientific computing, particularly in power-constrained settings such as edge and embedded devices, where repeated operator evaluations in dense networks incur substantial computational and energy costs. To address this challenge, we introduce the Separable Physics-informed Neuroscience-inspired Operator Network (SPINONet), a neuroscience-inspired framework that reduces redundant computation across repeated evaluations while remaining compatible with physics-informed training. SPINONet incorporates regression-friendly neuroscience-inspired spiking neurons through an architecture-aware design that enables sparse, event-driven computation, improving energy efficiency while preserving the continuous, coordinate-differentiable pathways required for computing spatio-temporal derivatives. We evaluate SPINONet on a range of partial differential equations representative of computational mechanics problems, with spatial, temporal, and parametric dependencies in both time-dependent and steady-state settings, and demonstrate predictive performance comparable to conventional physics-informed operator learning approaches despite the induced sparse communication. In addition, limited data supervision in a hybrid setup is shown to improve performance in challenging regimes where purely physics-informed training may converge to spurious solutions. Finally, we provide an analytical discussion linking architectural components and design choices of SPINONet to reductions in computational load and energy consumption.

Generating interpretable natural language captions from weather time series data remains a significant challenge at the intersection of meteorological science and natural language processing. While recent advances in Large Language Models (LLMs) have demonstrated remarkable capabilities in time series forecasting and analysis, existing approaches either produce numerical predictions without human-accessible explanations or generate generic descriptions lacking domain-specific depth. We introduce WeatherTGD, a training-free multi-agent framework that reinterprets collaborative caption refinement through the lens of Text Gradient Descent (TGD). Our system deploys three specialized LLM agents including a Statistical Analyst, a Physics Interpreter, and a Meteorology Expert that generate domain-specific textual gradients from weather time series observations. These gradients are aggregated through a novel Consensus-Aware Gradient Fusion mechanism that extracts common signals while preserving unique domain perspectives. The fused gradients then guide an iterative refinement process analogous to gradient descent, where each LLM-generated feedback signal updates the caption toward an optimal solution. Experiments on real-world meteorological datasets demonstrate that WeatherTGD achieves significant improvements in both LLM-based evaluation and human expert evaluation, substantially outperforming existing multi-agent baselines while maintaining computational efficiency through parallel agent execution.

Physical reservoir computing exploits the intrinsic dynamics of physical systems for information processing, while keeping the internal dynamics fixed and training only linear readouts; yet the role of input encoding remains poorly understood. We show that optimal input encoding is a geometric problem governed by the system's fluctuation-response structure. By measuring steady-state fluctuations and linear response, we derive an analytical criterion for the input direction that maximizes task-specific linear memory under a fixed power constraint, termed Response-based Optimal Memory Encoding (ROME). Backpropagation-based encoder optimization is shown to be equivalent to ROME, revealing a trade-off between task-dependent feature mixing and intrinsic noise. We apply ROME to various reservoir platforms, including spin-wave waveguides and spiking neural networks, demonstrating effective encoder design across physical and neuromorphic reservoirs, even in non-differentiable systems.

The rapid progress of large language models (LLMs) has led to remarkable performance gains across a wide range of tasks. However, when handling long documents that exceed the model's context window limit, the entire context cannot be processed in a single pass, making chunk-wise processing necessary. This requires multiple turns to read different chunks and update memory. However, supervision is typically provided only by the final outcome, which makes it difficult to evaluate the quality of memory updates at each turn in the multi-turn training setting. This introduces a temporal credit assignment challenge. Existing approaches, such as LLM-as-a-judge or process reward models, incur substantial computational overhead and suffer from estimation noise. To better address the credit assignment problem in multi-turn memory training, we propose Teacher-Aligned Reward Reshaping for Multi-Turn Reinforcement Learning (TAMTRL). TAMTRL leverages relevant documents as teacher signals by aligning them with each turn of model input and assigns rewards through normalized probabilities in a self-supervised manner. This provides fine-grained learning signals for each memory update and improves long-context processing. Experiments with multiple models of varying scales across seven long-context benchmarks show that TAMTRL consistently outperforms strong baselines, demonstrating its effectiveness. Our code is available at https://anonymous.4open.science/r/TAMTRL-F1F8.

Image deraining plays a pivotal role in low-level computer vision, serving as a prerequisite for robust outdoor surveillance and autonomous driving systems. While deep learning paradigms have achieved remarkable success in firmly aligned settings, they often suffer from severe performance degradation when generalized to unseen Out-of-Distribution (OOD) scenarios. This failure stems primarily from the significant domain discrepancy between synthetic training datasets and the complex physical dynamics of real-world rain. To address these challenges, this paper proposes a pioneering cross-scenario deraining adaptation framework. Diverging from conventional approaches, our method obviates the requirements for paired rainy observations in the target domain, leveraging exclusively rain-free background images. We design a Superpixel Generation (Sup-Gen) module to extract stable structural priors from the source domain using Simple Linear Iterative Clustering. Subsequently, a Resolution-adaptive Fusion strategy is introduced to align these source structures with target backgrounds through texture similarity, ensuring the synthesis of diverse and realistic pseudo-data. Finally, we implement a pseudo-label re-Synthesize mechanism that employs multi-stage noise generation to simulate realistic rain streaks. This framework functions as a versatile plug-and-play module capable of seamless integration into arbitrary deraining architectures. Extensive experiments on state-of-the-art models demonstrate that our approach yields remarkable PSNR gains of up to 32% to 59% in OOD domains while significantly accelerating training convergence.

The growing demand for multi-DNN workloads with unpredictable task arrival times has highlighted the need for interruptible scheduling on edge accelerators. However, existing preemptive frameworks typically assume known task arrival times and rely on CPU-based offline scheduling, which incurs heavy runtime overhead and struggles to handle unpredictable task arrivals. Even worse, prior studies have shown that multi-DNN scheduling requires solving an NP-hard subgraph isomorphism problem on large directed acyclic graphs within limited time, which is extremely challenging. To tackle this, we propose IMMSched, a parallel subgraph isomorphism method that combines Multi-Particle Optimization with the Ullmann algorithm based on a probabilistic continuous-relaxation scheme, eliminating the serial data dependencies of previous works. Finally, a quantized scheduling scheme and a global controller in the hardware architecture further combine multi-particle results for consensus-guided exploration. Evaluations demonstrate that IMMSched achieves orders-of-magnitude reductions in scheduling latency and energy consumption, enabling real-time execution of unpredictable DNN tasks on edge accelerators.

Memorization is a fundamental component of intelligence for both humans and LLMs. However, while LLM performance scales rapidly, our understanding of memorization lags. Due to limited access to the pre-training data of LLMs, most previous studies focus on a single model series, leading to isolated observations among series, making it unclear which findings are general or specific. In this study, we collect multiple model series (Pythia, OpenLLaMa, StarCoder, OLMo1/2/3) and analyze their shared or unique memorization behavior at both the statistical and internal levels, connecting individual observations while showing new findings. At the statistical level, we reveal that the memorization rate scales log-linearly with model size, and memorized sequences can be further compressed. Further analysis demonstrated a shared frequency and domain distribution pattern for memorized sequences. However, different models also show individual features under the above observations. At the internal level, we find that LLMs can remove certain injected perturbations, while memorized sequences are more sensitive. By decoding middle layers and attention head ablation, we revealed the general decoding process and shared important heads for memorization. However, the distribution of those important heads differs between families, showing a unique family-level feature. Through bridging various experiments and revealing new findings, this study paves the way for a universal and fundamental understanding of memorization in LLM.

Protecting patient privacy remains a fundamental barrier to scaling machine learning across healthcare institutions, where centralizing sensitive data is often infeasible due to ethical, legal, and regulatory constraints. Federated learning offers a promising alternative by enabling privacy-preserving, multi-institutional training without sharing raw patient data; however, real-world deployments face severe challenges from data heterogeneity, site-specific biases, and class imbalance, which degrade predictive reliability and render existing uncertainty quantification methods ineffective. Here, we present TrustFed, a federated uncertainty quantification framework that provides distribution-free, finite-sample coverage guarantees under heterogeneous and imbalanced healthcare data, without requiring centralized access. TrustFed introduces a representation-aware client assignment mechanism that leverages internal model representations to enable effective calibration across institutions, along with a soft-nearest threshold aggregation strategy that mitigates assignment uncertainty while producing compact and reliable prediction sets. Using over 430,000 medical images across six clinically distinct imaging modalities, we conduct one of the most comprehensive evaluations of uncertainty-aware federated learning in medical imaging, demonstrating robust coverage guarantees across datasets with diverse class cardinalities and imbalance regimes. By validating TrustFed at this scale and breadth, our study advances uncertainty-aware federated learning from proof-of-concept toward clinically meaningful, modality-agnostic deployment, positioning statistically guaranteed uncertainty as a core requirement for next-generation healthcare AI systems.

Retrieval-Augmented Generation (RAG) significantly mitigates the hallucinations and domain knowledge deficiency in large language models by incorporating external knowledge bases. However, the multi-module architecture of RAG introduces complex system-level security vulnerabilities. Guided by the RAG workflow, this paper analyzes the underlying vulnerability mechanisms and systematically categorizes core threat vectors such as data poisoning, adversarial attacks, and membership inference attacks. Based on this threat assessment, we construct a taxonomy of RAG defense technologies from a dual perspective encompassing both input and output stages. The input-side analysis reviews data protection mechanisms including dynamic access control, homomorphic encryption retrieval, and adversarial pre-filtering. The output-side examination summarizes advanced leakage prevention techniques such as federated learning isolation, differential privacy perturbation, and lightweight data sanitization. To establish a unified benchmark for future experimental design, we consolidate authoritative test datasets, security standards, and evaluation frameworks. To the best of our knowledge, this paper presents the first end-to-end survey dedicated to the security of RAG systems. Distinct from existing literature that isolates specific vulnerabilities, we systematically map the entire pipeline-providing a unified analysis of threat models, defense mechanisms, and evaluation benchmarks. By enabling deep insights into potential risks, this work seeks to foster the development of highly robust and trustworthy next-generation RAG systems.

Predicting the next mobile app a user will launch is essential for proactive mobile services. Yet accurate prediction remains challenging in real-world settings, where user intent can shift rapidly within short sessions and user-specific historical profiles are often sparse or unavailable, especially under cold-start conditions. Existing approaches mainly model app usage as sequential behavior or local session transitions, limiting their ability to capture higher-order structural dependencies and evolving session intent. To address this issue, we propose MISApp, a profile-free framework for next app prediction based on multi-hop session graph learning. MISApp constructs multi-hop session graphs to capture transition dependencies at different structural ranges, learns session representations through lightweight graph propagation, incorporates temporal and spatial context to characterize session conditions, and captures intent evolution from recent interactions. Experiments on two real-world app usage datasets show that MISApp consistently outperforms competitive baselines under both standard and cold-start settings, while maintaining a favorable balance between predictive accuracy and practical efficiency. Further analyses show that the learned hop-level attention weights align well with structural relevance, offering interpretable evidence for the effectiveness of the proposed multi-hop modeling strategy.

Federated learning (FL) has emerged as a promising paradigm for privacy-preserving multi-camera video understanding. However, applying FL to cross-view scenarios faces three major challenges: (i) heterogeneous viewpoints and backgrounds lead to highly non-IID client distributions and overfitting to view-specific patterns, (ii) local distribution biases cause misaligned representations that hinder consistent cross-view semantics, and (iii) large video architectures incur prohibitive communication overhead. To address these issues, we propose FedCVU, a federated framework with three components: VS-Norm, which preserves normalization parameters to handle view-specific statistics; CV-Align, a lightweight contrastive regularization module to improve cross-view representation alignment; and SLA, a selective layer aggregation strategy that reduces communication without sacrificing accuracy. Extensive experiments on action understanding and person re-identification tasks under a cross-view protocol demonstrate that FedCVU consistently boosts unseen-view accuracy while maintaining strong seen-view performance, outperforming state-of-the-art FL baselines and showing robustness to domain heterogeneity and communication constraints.

Prompt injection is listed as the number-one vulnerability class in the OWASP Top 10 for LLM Applications that can subvert LLM guardrails, disclose sensitive data, and trigger unauthorized tool use. Developers are rapidly adopting AI-assisted development tools built on the Model Context Protocol (MCP). However, their convenience comes with security risks, especially prompt-injection attacks delivered via tool-poisoning vectors. While prior research has studied prompt injection in LLMs, the security posture of real-world MCP clients remains underexplored. We present the first empirical analysis of prompt injection with the tool-poisoning vulnerability across seven widely used MCP clients: Claude Desktop, Claude Code, Cursor, Cline, Continue, Gemini CLI, and Langflow. We identify their detection and mitigation mechanisms, as well as the coverage of security features, including static validation, parameter visibility, injection detection, user warnings, execution sandboxing, and audit logging. Our evaluation reveals significant disparities. While some clients, such as Claude Desktop, implement strong guardrails, others, such as Cursor, exhibit high susceptibility to cross-tool poisoning, hidden parameter exploitation, and unauthorized tool invocation. We further provide actionable guidance for MCP implementers and the software engineering community seeking to build secure AI-assisted development workflows.

The Model Context Protocol (MCP) has emerged as a standard for connecting Large Language Models (LLMs) to external tools and data. However, MCP servers often expose privileged capabilities, such as file system access, network requests, and command execution that can be exploited if not properly secured. We present mcp-sec-audit, an extensible security assessment toolkit designed specifically for MCP servers. It implements static pattern matching for Python-based MCP servers and dynamic sandboxed fuzzing and monitoring via Docker and eBPF. The tool detects risky capabilities through configurable rule-based analysis and provides mitigation recommendations.

Most research in urban informatics and tourism focuses on mitigating overtourism in dense global cities. However, for regions experiencing demographic decline and structural stagnation, the primary risk is "under-vibrancy", a condition where low visitor density suppresses economic activity and diminishes satisfaction. This paper introduces the Distributed Human Data Engine (DHDE), a socio-technical framework previously validated in biological crisis management, and adapts it for regional economic flow optimization. Using high-granularity data from Japan's least-visited prefecture (Fukui), we utilize an AI-driven decision support system (DSS) to analyze two datasets: a raw Fukui spending database (90,350 records) and a regional standardized sentiment database (97,719 responses). The system achieves in-sample explanatory power of 81% (R^2 = 0.810) and out-of-sample predictive performance of 68% (R^2 = 0.683). We quantify an annual opportunity gap of 865,917 unrealized visits, equivalent to approximately 11.96 billion yen (USD 76.2 million) in lost revenue. We propose a dual-nudge governance architecture leveraging the DHDE to redistribute cross-prefectural flows and reduce economic leakage.

Public benchmarks increasingly govern how large language models (LLMs) are ranked, selected, and deployed. We frame this benchmark-centered regime as Silicon Bureaucracy and AI Test-Oriented Education, and argue that it rests on a fragile assumption: that benchmark scores directly reflect genuine generalization. In practice, however, such scores may conflate exam-oriented competence with principled capability, especially when contamination and semantic leakage are difficult to exclude from modern training pipelines. We therefore propose an audit framework for analyzing contamination sensitivity and score confidence in LLM benchmarks. Using a router-worker setup, we compare a clean-control condition with noisy conditions in which benchmark problems are systematically deleted, rewritten, and perturbed before being passed downstream. For a genuinely clean benchmark, noisy conditions should not consistently outperform the clean-control baseline. Yet across multiple models, we find widespread but heterogeneous above-baseline gains under noisy conditions, indicating that benchmark-related cues may be reassembled and can reactivate contamination-related memory. These results suggest that similar benchmark scores may carry substantially different levels of confidence. Rather than rejecting benchmarks altogether, we argue that benchmark-based evaluation should be supplemented with explicit audits of contamination sensitivity and score confidence.

Deploying AI agents in enterprise environments requires balancing capability with data sovereignty and cost constraints. While small language models offer privacy-preserving alternatives to frontier models, their specialization is hindered by fragmented development pipelines that separate tool integration, data generation, and training. We introduce EnterpriseLab, a full-stack platform that unifies these stages into a closed-loop framework. EnterpriseLab provides (1) a modular environment exposing enterprise applications via Model Context Protocol, enabling seamless integration of proprietary and open-source tools; (2) automated trajectory synthesis that programmatically generates training data from environment schemas; and (3) integrated training pipelines with continuous evaluation. We validate the platform through EnterpriseArena, an instantiation with 15 applications and 140+ tools across IT, HR, sales, and engineering domains. Our results demonstrate that 8B-parameter models trained within EnterpriseLab match GPT-4o's performance on complex enterprise workflows while reducing inference costs by 8-10x, and remain robust across diverse enterprise benchmarks, including EnterpriseBench (+10%) and CRMArena (+10%). EnterpriseLab provides enterprises a practical path to deploying capable, privacy-preserving agents without compromising operational capability.

On-policy reinforcement learning with generative policies is promising but remains underexplored. A central challenge is that proximal policy optimization (PPO) is traditionally formulated in terms of action-space probability ratios, whereas diffusion- and flow-based policies are more naturally represented as trajectory-level generative processes. In this work, we propose GSB-PPO, a path-space formulation of generative PPO inspired by the Generalized Schrödinger Bridge (GSB). Our framework lifts PPO-style proximal updates from terminal actions to full generation trajectories, yielding a unified view of on-policy optimization for generative policies. Within this framework, we develop two concrete objectives: a clipping-based objective, GSB-PPO-Clip, and a penalty-based objective, GSB-PPO-Penalty. Experimental results show that while both objectives are compatible with on-policy training, the penalty formulation consistently delivers better stability and performance than the clipping counterpart. Overall, our results highlight path-space proximal regularization as an effective principle for training generative policies with PPO.

The rapid progress of text-to-image models has made AI-generated images increasingly realistic, posing significant challenges for accurate detection of generated content. While training-based detectors often suffer from limited generalization to unseen images, training-free approaches offer better robustness, yet struggle to capture subtle discrepancies between real and synthetic images. In this work, we propose a training-free AI-generated image detection method that measures representation sensitivity to structured frequency perturbations, enabling detection of minute manipulations. The proposed method is computationally lightweight, as perturbation generation requires only a single Fourier transform for an input image. As a result, it achieves one to two orders of magnitude faster inference than most training-free detectors.Extensive experiments on challenging benchmarks demonstrate the efficacy of our method over state-of-the-art (SoTA). In particular, on OpenFake benchmark, our method improves AUC by nearly $10\%$ compared to SoTA, while maintaining substantially lower computational cost.

In recent years, numerous data-intensive broadcasting applications have emerged at the wireless edge, calling for a flexible tradeoff between distortion, transmission rate, and processing complexity. While deep learning-based joint source-channel coding (DeepJSCC) has been identified as a potential solution to data-intensive communications, most of these schemes are confined to worst-case solutions, lack adaptive complexity, and are inefficient in broadcast settings. To overcome these limitations, this paper introduces nonlinear transform rateless source-channel coding (NTRSCC), a variable-length JSCC framework for broadcast channels based on rateless codes. In particular, we integrate learned source transformations with physical-layer LT codes, develop unequal protection schemes that exploit decoder side information, and devise approximations to enable end-to-end optimization of rateless parameters. Our framework enables heterogeneous receivers to adaptively adjust their received number of rateless symbols and decoding iterations in belief propagation, thereby achieving a controllable tradeoff between distortion, rate, and decoding complexity. Simulation results demonstrate that the proposed method enhances image broadcast quality under stringent communication and processing budgets over heterogeneous edge devices.

Recommender agents built on Large Language Models offer a promising paradigm for recommendation. However, existing recommender agents typically suffer from a disconnect between intermediate reasoning and final ranking feedback, and are unable to capture fine-grained preferences. To address this, we present AgenticRec, a ranking-oriented agentic recommendation framework that optimizes the entire decision-making trajectory (including intermediate reasoning, tool invocation, and final ranking list generation) under sparse implicit feedback. Our approach makes three key contributions. First, we design a suite of recommendation-specific tools integrated into a ReAct loop to support evidence-grounded reasoning. Second, we propose theoretically unbiased List-Wise Group Relative Policy Optimization (list-wise GRPO) to maximize ranking utility, ensuring accurate credit assignment for complex tool-use trajectories. Third, we introduce Progressive Preference Refinement (PPR) to resolve fine-grained preference ambiguities. By mining hard negatives from ranking violations and applying bidirectional preference alignment, PPR minimizes the convex upper bound of pairwise ranking errors. Experiments on benchmarks confirm that AgenticRec significantly outperforms baselines, validating the necessity of unifying reasoning, tool use, and ranking optimization.

Time series anomaly detection plays a critical role in many dynamic systems. Despite its importance, previous approaches have primarily relied on unimodal numerical data, overlooking the importance of complementary information from other modalities. In this paper, we propose a novel multimodal time series anomaly detection model (MindTS) that focuses on addressing two key challenges: (1) how to achieve semantically consistent alignment across heterogeneous multimodal data, and (2) how to filter out redundant modality information to enhance cross-modal interaction effectively. To address the first challenge, we propose Fine-grained Time-text Semantic Alignment. It integrates exogenous and endogenous text information through cross-view text fusion and a multimodal alignment mechanism, achieving semantically consistent alignment between time and text modalities. For the second challenge, we introduce Content Condenser Reconstruction, which filters redundant information within the aligned text modality and performs cross-modal reconstruction to enable interaction. Extensive experiments on six real-world multimodal datasets demonstrate that the proposed MindTS achieves competitive or superior results compared to existing methods. The code is available at: https://github.com/decisionintelligence/MindTS.

Existing machine learning frameworks for compliance monitoring -- Markov Logic Networks, Probabilistic Soft Logic, supervised models -- share a fundamental paradigm: they treat observed data as ground truth and attempt to approximate rules from it. This assumption breaks down in rule-governed domains such as taxation or regulatory compliance, where authoritative rules are known a priori and the true challenge is to infer the latent state of rule activation, compliance, and parametric drift from partial and noisy observations. We propose Rule-State Inference (RSI), a Bayesian framework that inverts this paradigm by encoding regulatory rules as structured priors and casting compliance monitoring as posterior inference over a latent rule-state space S = {(a_i, c_i, delta_i)}, where a_i captures rule activation, c_i models the compliance rate, and delta_i quantifies parametric drift. We prove three theoretical guarantees: (T1) RSI absorbs regulatory changes in O(1) time via a prior ratio correction, independently of dataset size; (T2) the posterior is Bernstein-von Mises consistent, converging to the true rule state as observations accumulate; (T3) mean-field variational inference monotonically maximizes the Evidence Lower BOund (ELBO). We instantiate RSI on the Togolese fiscal system and introduce RSI-Togo-Fiscal-Synthetic v1.0, a benchmark of 2,000 synthetic enterprises grounded in real OTR regulatory rules (2022-2025). Without any labeled training data, RSI achieves F1=0.519 and AUC=0.599, while absorbing regulatory changes in under 1ms versus 683-1082ms for full model retraining -- at least a 600x speedup.

Recent advances in generative models, such as diffusion and flow matching, have shown strong performance in audio tasks. However, speech enhancement (SE) models are typically trained on limited datasets and evaluated under narrow conditions, limiting real-world applicability. To address this, we propose DiT-Flow, a flow matching-based SE framework built on the latent Diffusion Transformer (DiT) backbone and trained for robustness across diverse distortions, including noise, reverberation, and compression. DiT-Flow operates on compact variational auto-encoders (VAEs)-derived latent features. We validated our approach on StillSonicSet, a synthetic yet acoustically realistic dataset composed of LibriSpeech, FSD50K, FMA, and 90 Matterport3D scenes. Experiments show that DiT-Flow consistently outperforms state-of-the-art generative SE models, demonstrating the effectiveness of flow matching in multi-condition speech enhancement. Despite ongoing efforts to expand synthetic data realism, a persistent bottleneck in SE is the inevitable mismatch between training and deployment conditions. By integrating LoRA with the MoE framework, we achieve both parameter-efficient and high-performance training for DiT-Flow robust to multiple distortions with using 4.9% percentage of the total parameters to obtain a better performance on five unseen distortions.

While retrieval-augmented generation (RAG) significantly improves the factual reliability of LLMs, it does not eliminate hallucinations, so robust uncertainty quantification (UQ) remains essential. In this paper, we reveal that standard entropy-based UQ methods often fail in RAG settings due to a mechanistic paradox. An internal "tug-of-war" inherent to context utilization appears: while induction heads promote grounded responses by copying the correct answer, they collaterally trigger the previously established "entropy neurons". This interaction inflates predictive entropy, causing the model to signal false uncertainty on accurate outputs. To address this, we propose INTRYGUE (Induction-Aware Entropy Gating for Uncertainty Estimation), a mechanistically grounded method that gates predictive entropy based on the activation patterns of induction heads. Evaluated across four RAG benchmarks and six open-source LLMs (4B to 13B parameters), INTRYGUE consistently matches or outperforms a wide range of UQ baselines. Our findings demonstrate that hallucination detection in RAG benefits from combining predictive uncertainty with interpretable, internal signals of context utilization.

Current language model training commonly applies multi-task Supervised Fine-Tuning (SFT) using a homogeneous compute budget across all sub-datasets. This approach is fundamentally sub-optimal: heterogeneous learning dynamics cause faster-learning tasks to overfit early while slower ones remain under-fitted. To address this, we introduce mSFT, an iterative, overfitting-aware search algorithm for multi-task data mixtures. mSFT trains the model on an active mixture, identifies and excludes the earliest overfitting sub-dataset, and reverts to that specific optimal checkpoint before continuing. Extensive evaluations demonstrate that mSFT consistently outperforms 4 baselines across 10 benchmarks and 6 base models. Further analysis confirms mSFT maintains robust gains across diverse dataset sizes, task granularities, and is insensitive to its single new hyperparameter (compute budget). Notably, at low compute budget, mSFT can improve performance while lowering training FLOPs. Ultimately, mSFT establishes a practical overfitting-aware algorithm for multi-task SFT that maximizes the potential of models across diverse data mixtures.

Asynchronous messaging is a cornerstone of modern distributed systems, enabling decoupled communication for scalable and resilient applications. Today's message queue (MQ) ecosystem spans a wide range of designs, from high-throughput streaming platforms to lightweight protocols tailored for edge and IoT environments. Despite this diversity, choosing an appropriate MQ system remains difficult. Existing evaluations largely focus on throughput and latency on fixed hardware, while overlooking CPU and memory footprint and the effects of resource constraints, factors that are critical for edge and IoT deployments. In this paper, we present a systematic performance study of eight prominent message brokers: Mosquitto, EMQX, HiveMQ, RabbitMQ, ActiveMQ Artemis, NATS Server, Redis (Pub/Sub), and Zenoh Router. We introduce mq-bench, a unified benchmarking framework to evaluate these systems under identical conditions, scaling up to 10,000 concurrent client pairs across three VM configurations representative of edge hardware. This study reveals several interesting and sometimes counter-intuitive insights. Lightweight native brokers achieve sub-millisecond latency, while feature-rich enterprise platforms incur 2-3X higher overhead. Under high connection loads, multi-threaded brokers like NATS and Zenoh scale efficiently, whereas the widely-deployed Mosquitto saturates earlier due to its single-threaded architecture. We also find that Java-based brokers consume significantly more memory than native implementations, which has important implications for memory-constrained edge deployments. Based on these findings, we provide practical deployment guidelines that map workload requirements and resource constraints to appropriate broker choices for telemetry, streaming analytics, and IoT use cases.

Graphs provide a natural description of the complex relationships among objects, and play a pivotal role in communications, transportation, social computing, the life sciences, etc. Currently, there is strong agreement that Graph Foundation Models (GFMs) are essential for advancing graph learning, yet considerable disagreement persists on how to build a powerful, general-purpose GFM analogous to Large Language Models (LLMs). Graph Neural Networks (GNNs) exhibit limitations in memory retention and principled interpretability when confronted with multi-domain pretraining and adaptation. The challenge of graph serialization hinders the direct application of LLMs, as the words struggle to capture the structural complexity and diversity inherent in graphs. In contrast, Riemannian geometry offers an elegant mathematical framework for modeling structures, while remaining compatible with graph semantic learning, even with LLMs. In this paper, we argue that, for graphs, Riemannian geometry speaks louder than words, and lay out the foundational principles for GFM. Reimagining with Riemannian geometry, we introduce a blue sky idea-Riemannian Foundation Model (RFM)-that opens a new pathway for capturing complex structural patterns and uncovering cross-domain generalities. RFM emphasizes intrinsic graph geometry and embodies endogenous capacities for structural inference and generation, moving beyond mere representation-space switching. Accordingly, we outline a progressive agenda that begins with universal structural understanding through intrinsic geometry, and then rebuilds LLM with a Riemannian engine for general-purpose graph modeling and beyond. Thus, RFM enables a paradigm shift from designing graph models to solving graph-structured applications with RFM agents, unlocking the next-generation graph intelligence.

Modern clinical practice increasingly depends on reasoning over heterogeneous, evolving, and incomplete patient data. Although recent advances in multimodal foundation models have improved performance on various clinical tasks, most existing models remain static, opaque, and poorly aligned with real-world clinical workflows. We present Cerebra, an interactive multi-agent AI team that coordinates specialized agents for EHR, clinical notes, and medical imaging analysis. These outputs are synthesized into a clinician-facing dashboard that combines visual analytics with a conversational interface, enabling clinicians to interrogate predictions and contextualize risk at the point of care. Cerebra supports privacy-preserving deployment by operating on structured representations and remains robust when modalities are incomplete. We evaluated Cerebra using a massive multi-institutional dataset spanning 3 million patients from four independent healthcare systems. Cerebra consistently outperformed both state-of-the-art single-modality models and large multimodal language model baselines. In dementia risk prediction, it achieved AUROCs up to 0.80, compared with 0.74 for the strongest single-modality model and 0.68 for language model baselines. For dementia diagnosis, it achieved an AUROC of 0.86, and for survival prediction, a C-index of 0.81. In a reader study with experienced physicians, Cerebra significantly improved expert performance, increasing accuracy by 17.5 percentage points in prospective dementia risk estimation. These results demonstrate Cerebra's potential for interpretable, robust decision support in clinical care.

The rapid expansion of the Internet of Things (IoT) and its integration with backbone networks have heightened the risk of security breaches. Traditional centralized approaches to anomaly detection, which require transferring large volumes of data to central servers, suffer from privacy, scalability, and latency limitations. This paper proposes a lightweight autoencoder-based anomaly detection framework designed for deployment on resource-constrained edge devices, enabling real-time detection while minimizing data transfer and preserving privacy. Federated learning is employed to train models collaboratively across distributed devices, where local training occurs on edge nodes and only model weights are aggregated at a central server. A real-world IoT testbed using Raspberry Pi sensor nodes was developed to collect normal and attack traffic data. The proposed federated anomaly detection system, implemented and evaluated on the testbed, demonstrates its effectiveness in accurately identifying network attacks. The communication overhead was reduced significantly while achieving comparable performance to the centralized method.

In this paper, we employ multiple UAVs to accelerate data transmissions from ground users (GUs) to a remote base station (BS) via the UAVs' relay communications. The UAVs' intermittent information exchanges typically result in delays in acquiring the complete system state and hinder their effective collaboration. To maximize the overall throughput, we first propose a delay-tolerant multi-agent deep reinforcement learning (MADRL) algorithm that integrates a delay-penalized reward to encourage information sharing among UAVs, while jointly optimizing the UAVs' trajectory planning, network formation, and transmission control strategies. Additionally, considering information loss due to unreliable channel conditions, we further propose a spatio-temporal attention based prediction approach to recover the lost information and enhance each UAV's awareness of the network state. These two designs are envisioned to enhance the network capacity in UAV-assisted wireless networks with limited communications. The simulation results reveal that our new approach achieves over 50\% reduction in information delay and 75% throughput gain compared to the conventional MADRL. Interestingly, it is shown that improving the UAVs' information sharing will not sacrifice the network capacity. Instead, it significantly improves the learning performance and throughput simultaneously. It is also effective in reducing the need for UAVs' information exchange and thus fostering practical deployment of MADRL in UAV-assisted wireless networks.

In many learning systems, such as activity recognition systems, as new data collection methods continue to emerge in various dynamic environmental applications, the attributes of instances accumulate incrementally, with data being stored in gradually expanding feature spaces. How to design theoretically guaranteed algorithms to effectively cluster this special type of data stream, commonly referred to as activity recognition, remains unexplored. Compared to traditional scenarios, we will face at least two fundamental questions in this feature incremental scenario. (i) How to design preliminary and effective algorithms to address the feature incremental clustering problem? (ii) How to analyze the generalization bounds for the proposed algorithms and under what conditions do these algorithms provide a strong generalization guarantee? To address these problems, by tailoring the most common clustering algorithm, i.e., $k$-means, as an example, we propose four types of Feature Incremental Clustering (FIC) algorithms corresponding to different situations of data access: Feature Tailoring (FT), Data Reconstruction (DR), Data Adaptation (DA), and Model Reuse (MR), abbreviated as FIC-FT, FIC-DR, FIC-DA, and FIC-MR. Subsequently, we offer a detailed analysis of the generalization error bounds for these four algorithms and highlight the critical factors influencing these bounds, such as the amounts of training data, the complexity of the hypothesis space, the quality of pre-trained models, and the discrepancy of the reconstruction feature distribution. The numerical experiments show the effectiveness of the proposed algorithms, particularly in their application to activity recognition clustering tasks.

Multimodal large language models (MLLMs) achieve strong performance by jointly processing inputs from multiple modalities, such as vision, audio, and language. However, building such models or extending them to new modalities often requires large paired datasets and substantial computational resources. Since many pretrained MLLMs (e.g., vision-language or audio-language) are publicly available, we ask whether we can merge them into a single MLLM that can handle multiple modalities? Merging MLLMs with different input modalities remains challenging, partly because of differences in the learned representations and interference between their parameter spaces. To address these challenges, we propose Singular Subspace Alignment and Merging (SSAM), a training-free model merging framework that unifies independently trained specialist MLLMs into a single model capable of handling any combination of input modalities. SSAM maintains modality-specific parameter updates separately and identifies a shared low-rank subspace for language-related parameter updates, aligns them within this subspace, and merges them to preserve complementary knowledge while minimizing parameter interference. Without using any multimodal training data, SSAM achieves state-of-the-art performance across four datasets, surpassing prior training-free merging methods and even jointly trained multimodal models. These results demonstrate that aligning models in parameter space provides a scalable and resource-efficient alternative to conventional joint multimodal training.

Despite the widespread adoption of MLLMs in embodied agents, their capabilities remain largely confined to reactive planning from immediate observations, consistently failing in spatial reasoning across extensive spatiotemporal scales. Cognitive science reveals that Biological Intelligence (BI) thrives on "mental navigation": the strategic construction of spatial representations from experience and the subsequent mental simulation of paths prior to action. To bridge the gap between AI and BI, we introduce Video2Mental, a pioneering benchmark for evaluating the mental navigation capabilities of MLLMs. The task requires constructing hierarchical cognitive maps from long egocentric videos and generating landmark-based path plans step by step, with planning accuracy verified through simulator-based physical interaction. Our benchmarking results reveal that mental navigation capability does not naturally emerge from standard pre-training. Frontier MLLMs struggle profoundly with zero-shot structured spatial representation, and their planning accuracy decays precipitously over extended horizons. To overcome this, we propose \textbf{NavMind}, a reasoning model that internalizes mental navigation using explicit, fine-grained cognitive maps as learnable intermediate representations. Through a difficulty-stratified progressive supervised fine-tuning paradigm, NavMind effectively bridges the gap between raw perception and structured planning. Experiments demonstrate that NavMind achieves superior mental navigation capabilities, significantly outperforming frontier commercial and spatial MLLMs.

Long-context LLM inference is bottlenecked not by compute but by the O(n) memory bandwidth cost of scanning the KV cache at every decode step -- a wall that no amount of arithmetic scaling can break. Recent photonic accelerators have demonstrated impressive throughput for dense attention computation; however, these approaches inherit the same O(n) memory scaling as electronic attention when applied to long contexts. We observe that the real leverage point is the coarse block-selection step: a memory-bound similarity search that determines which KV blocks to fetch. We identify, for the first time, that this task is structurally matched to the photonic broadcast-and-weight paradigm -- the query fans out to all candidates via passive splitting, signatures are quasi-static (matching electro-optic MRR programming), and only rank order matters (relaxing precision to 4-6 bits). Crucially, the photonic advantage grows with context length: as N increases, the electronic scan cost rises linearly while the photonic evaluation remains O(1). We instantiate this insight in PRISM (Photonic Ranking via Inner-product Similarity with Microring weights), a thin-film lithium niobate (TFLN) similarity engine. Hardware-impaired needle-in-a-haystack evaluation on Qwen2.5-7B confirms 100% accuracy from 4K through 64K tokens at k=32, with 16x traffic reduction at 64K context. PRISM achieves a four-order-of-magnitude energy advantage over GPU baselines at practical context lengths (n >= 4K).

Multi-agent collaboration has emerged as a powerful paradigm for enhancing the reasoning capabilities of large language models, yet it suffers from interaction-level ambiguity that blurs generation, critique, and revision, making credit assignment across agents difficult. Moreover, policy optimization in this setting is vulnerable to heavy-tailed and noisy rewards, which can bias advantage estimation and trigger unstable or even divergent training. To address both issues, we propose a robust multi-agent reinforcement learning framework for collaborative reasoning, consisting of two components: Dual-Agent Answer-Critique-Rewrite (DACR) and an Adaptive Robust Estimator (ARE). DACR decomposes reasoning into a structured three-stage pipeline: answer, critique, and rewrite, while enabling explicit attribution of each agent's marginal contribution to its partner's performance. ARE provides robust estimation of batch experience means during multi-agent policy optimization. Across mathematical reasoning and embodied intelligence benchmarks, even under noisy rewards, our method consistently outperforms the baseline in both homogeneous and heterogeneous settings. These results indicate stronger robustness to reward noise and more stable training dynamics, effectively preventing optimization failures caused by noisy reward signals.

DATASHI is a new parallel English-Tashlhiyt corpus that fills a critical gap in computational resources for Amazigh languages. It contains 5,000 sentence pairs, including a 1,500-sentence subset with expert-standardized and non-standard user-generated versions, enabling systematic study of orthographic diversity and normalization. This dual design supports text-based NLP tasks - such as tokenization, translation, and normalization - and also serves as a foundation for read-speech data collection and multimodal alignment. Comprehensive evaluations with state-of-the-art Large Language Models (GPT-5, Claude-Sonnet-4.5, Gemini-2.5-Pro, Mistral, Qwen3-Max) show clear improvements from zero-shot to few-shot prompting, with Gemini-2.5-Pro achieving the lowest word and character-level error rates and exhibiting robust cross-lingual generalization. A fine-grained analysis of edit operations - deletions, substitutions, and insertions - across phonological classes (geminates, emphatics, uvulars, and pharyngeals) further highlights model-specific sensitivities to marked Tashlhiyt features and provides new diagnostic insights for low-resource Amazigh orthography normalization.

We address a numerical framework for the stability and bifurcation analysis of nonlinear partial differential equations (PDEs) in which the solution is sought in the function space spanned by physics-informed random projection neural networks (PI-RPNNs), and discretized via a collocation approach. These are single-hidden-layer networks with randomly sampled and fixed a priori hidden-layer weights; only the linear output layer weights are optimized, reducing training to a single least-squares solve. This linear output structure enables the direct and explicit formulation of the eigenvalue problem governing the linear stability of stationary solutions. This takes a generalized eigenvalue form, which naturally separates the physical domain interior dynamics from the algebraic constraints imposed by boundary conditions, at no additional training cost and without requiring additional PDE solves. However, the random projection collocation matrix is inherently numerically rank-deficient, rendering naive eigenvalue computation unreliable and contaminating the true eigenvalue spectrum with spurious near-zero modes. To overcome this limitation, we introduce a matrix-free shift-invert Krylov-Arnoldi method that operates directly in weight space, avoiding explicit inversion of the numerically rank-deficient collocation matrix and enabling the reliable computation of several leading eigenpairs of the physical Jacobian - the discretized Frechet derivative of the PDE operator with respect to the solution field, whose eigenvalue spectrum determines linear stability. We further prove that the PI-RPNN-based generalized eigenvalue problem is almost surely regular, guaranteeing solvability with standard eigensolvers, and that the singular values of the random projection collocation matrix decay exponentially for analytic activation functions.

Large language models can rewrite text to embed hidden payloads while preserving surface-level meaning, a capability that opens covert channels between cooperating AI systems and poses challenges for alignment monitoring. We study the information-theoretic cost of such embedding. Our main result is that any steganographic scheme that preserves the semantic load of a covertext~$M_1$ while encoding a payload~$P$ into a stegotext~$M_2$ must satisfy $K(M_2) \geq K(M_1) + K(P) - O(\log n)$, where $K$ denotes Kolmogorov complexity and $n$ is the combined message length. A corollary is that any non-trivial payload forces a strict complexity increase in the stegotext, regardless of how cleverly the encoder distributes the signal. Because Kolmogorov complexity is uncomputable, we ask whether practical proxies can detect this predicted increase. Drawing on the classical correspondence between lossless compression and Kolmogorov complexity, we argue that language-model perplexity occupies an analogous role in the probabilistic regime and propose the Binoculars perplexity-ratio score as one such proxy. Preliminary experiments with a color-based LLM steganographic scheme support the theoretical prediction: a paired $t$-test over 300 samples yields $t = 5.11$, $p < 10^{-6}$.

We present CataractSAM-2, a domain-adapted extension of Meta's Segment Anything Model 2, designed for real-time semantic segmentation of cataract ophthalmic surgery videos with high accuracy. Positioned at the intersection of computer vision and medical robotics, CataractSAM-2 enables precise intraoperative perception crucial for robotic-assisted and computer-guided surgical systems. Furthermore, to alleviate the burden of manual labeling, we introduce an interactive annotation framework that combines sparse prompts with video-based mask propagation. This tool significantly reduces annotation time and facilitates the scalable creation of high-quality ground-truth masks, accelerating dataset development for ocular anterior segment surgeries. We also demonstrate the model's strong zero-shot generalization to glaucoma trabeculectomy procedures, confirming its cross-procedural utility and potential for broader surgical applications. The trained model and annotation toolkit are released as open-source resources, establishing CataractSAM-2 as a foundation for expanding anterior ophthalmic surgical datasets and advancing real-time AI-driven solutions in medical robotics, as well as surgical video understanding.

Synthetic aperture radar automatic target recognition (SAR ATR) is of considerable importance in marine navigation and disaster monitoring. However, the coherent speckle noise inherent in SAR imagery often obscures salient target features, leading to degraded recognition accuracy and limited model generalization. To address this issue, this paper proposes a target-aware frequency-spatial enhancement framework with noise-resilient knowledge guidance (FSCE) for SAR target recognition. The proposed framework incorporates a frequency-spatial shallow feature adaptive enhancement (DSAF) module, which processes shallow features through spatial multi-scale convolution and frequency-domain wavelet convolution. In addition, a teacher-student learning paradigm combined with an online knowledge distillation method (KD) is employed to guide the student network to focus more effectively on target regions, thereby enhancing its robustness to high-noise backgrounds. Through the collaborative optimization of attention transfer and noise-resilient representation learning, the proposed approach significantly improves the stability of target recognition under noisy conditions. Based on the FSCE framework, two network architectures with different performance emphases are developed: lightweight DSAFNet-M and high-precision DSAFNet-L. Extensive experiments are conducted on the MSTAR, FUSARShip and OpenSARShip datasets. The results show that DSAFNet-L achieves competitive or superior performance compared with various methods on three datasets; DSAFNet-M significantly reduces the model complexity while maintaining comparable accuracy. These results indicate that the proposed FSCE framework exhibits strong cross-model generalization.

Many recent long-context and agentic systems address context-length limitations by adding hierarchical memory: they extract atomic units from raw data, build multi-level representatives by grouping and compression, and traverse this structure to retrieve content under a token budget. Despite recurring implementations, there is no shared formalism for comparing design choices. We propose a unifying theory in terms of three operators. Extraction ($α$) maps raw data to atomic information units; coarsening ($C = (π, ρ)$) partitions units and assigns a representative to each group; and traversal ($τ$) selects which units to include in context given a query and budget. We identify a self-sufficiency spectrum for the representative function $ρ$ and show how it constrains viable retrieval strategies (a coarsening-traversal coupling). Finally, we instantiate the decomposition on eleven existing systems spanning document hierarchies, conversational memory, and agent execution traces, showcasing its generality.

Collaborative multi-agent large language models (LLMs) can solve complex reasoning tasks by decomposing roles and aggregating diverse hypotheses. Yet, reinforcement learning (RL) for such systems is often undermined by credit assignment: a shared global reward obscures individual contributions, inflating update variance and encouraging free-riding. We introduce Counterfactual Credit Policy Optimization (CCPO), a framework that assigns agent-specific learning signals by estimating each agent's marginal contribution through counterfactual trajectories. CCPO builds dynamic counterfactual baselines that simulate outcomes with an agent's contribution removed, yielding role-sensitive advantages for policy optimization. To further improve stability under heterogeneous tasks and data distributions, we propose a global-history-aware normalization scheme that calibrates advantages using global rollout statistics. We evaluate CCPO on two collaboration topologies: a sequential Think--Reason dyad and multi-agent voting. Across mathematical and logical reasoning benchmarks, CCPO mitigates free-riding and outperforms strong multi-agent RL baselines, yielding finer-grained and more effective credit assignment for collaborative LLM training. Our code is available at https://github.com/bhai114/ccpo.

Recursive self-improvement--where a model iteratively trains on its own outputs--promises sustained capability growth but faces a fundamental obstacle: recursive drift. As models train on self-generated data across multiple iterations, errors in intermediate reasoning compound, leading to mode collapse and performance degradation. We propose Neuro-Symbolic Recursive Self-Alignment (NSRSA), which stabilizes iterative self-training by embedding a symbolic verification subsystem that gates training data quality at the reasoning step level. Unlike outcome-only filtering (which admits "lucky guesses" with flawed reasoning), NSRSA verifies each arithmetic operation via sympy, checks logical flow consistency across reasoning steps, and enforces domain constraints. We evaluate NSRSA on GSM8K using Qwen3-4B-Thinking across 5 self-training iterations under five conditions: no verification, outcome verification, majority voting, full NSRSA symbolic verification, and NSRSA with DPO. Our filtering analysis shows that NSRSA rejects approximately 34% of correct-answer solutions that pass outcome verification, eliminating "lucky guesses" with flawed reasoning from the training set. We further demonstrate that constructing DPO preference pairs from NSRSA verification teaches the model to distinguish sound from flawed reasoning (reward accuracy 46% to 63%). NSRSA provides an extensible framework that demonstrates how external symbolic verification can make recursive self-improvement measurable and reliable within domains where automated verification is available.

Developers rely on online tutorials to learn web application security, but tutorial quality varies. We reviewed 132 free security tutorials to examine topic coverage, authorship, and technical depth. Our analysis shows that most tutorials come from vendors and emphasize high-level explanations over concrete implementation guidance. Few tutorials provide complete runnable code examples or direct links to authoritative security resources such as the Open Web Application Security Project (OWASP), Common Weakness Enumeration (CWE), or Common Vulnerabilities and Exposures (CVE). We found that two visible signals help identify more useful tutorials: the presence of runnable code and direct links to official resources. These signals can help developers distinguish broad awareness material from tutorials that better support secure implementation.

Large language models are rapidly changing how learners acquire and demonstrate cybersecurity skills. However, when human--AI collaboration is allowed, educators still lack validated competition designs and evaluation practices that remain fair and evidence-based. This paper presents a cross-regional study of LLM-centered Capture-the-Flag competitions built on the Cyber Security Awareness Week competition system. To understand how autonomy levels and participants' knowledge backgrounds influence problem-solving performance and learning-related behaviors, we formalize three autonomy levels: human-in-the-loop, autonomous agent frameworks, and hybrid. To enable verification, we require traceable submissions including conversation logs, agent trajectories, and agent code. We analyze multi-region competition data covering an in-class track, a standard track, and a year-long expert track, each targeting participants with different knowledge backgrounds. Using data from the 2025 competition, we compare solve performance across autonomy levels and challenge categories, and observe that autonomous agent frameworks and hybrid achieve higher completion rates on challenges requiring iterative testing and tool interactions. In the in-class track, we classify participants' agent designs and find a preference for lightweight, tool-augmented prompting and reflection-based retries over complex multi-agent architectures. Our results offer actionable guidance for designing LLM-assisted cybersecurity competitions as learning technologies, including autonomy-specific scoring criteria, evidence requirements that support solution verification, and track structures that improve accessibility while preserving reliable evaluation and engagement.

The inclusion of internship courses in Software Engineering (SE) programs is essential for closing knowledge gaps and improving graduates' readiness for the software industry. Our study focuses on year-long internships at RMIT University (Melbourne, Australia), which offers in-depth industry engagement. We analysed how the course evolved over the last 10 years to incorporate students' needs and summarised the lessons learned that can be helpful for other educators supporting internship courses. Our qualitative analysis of internship data based on 91 reports during 2023-2024 identified three challenge themes the students faced, and which courses were found by students to be particularly beneficial during their internships. On this basis, we proposed recommendations for educators and companies to help interns overcome challenges and maximise their learning experience.

World models learn to simulate environment dynamics from experience, enabling sample-efficient reinforcement learning. But what do these models actually represent internally? We apply interpretability techniques--including linear and nonlinear probing, causal interventions, and attention analysis--to two architecturally distinct world models: IRIS (discrete token transformer) and DIAMOND (continuous diffusion UNet), trained on Atari Breakout and Pong. Using linear probes, we find that both models develop linearly decodable representations of game state variables (object positions, scores), with MLP probes yielding only marginally higher R^2, confirming that these representations are approximately linear. Causal interventions--shifting hidden states along probe-derived directions--produce correlated changes in model predictions, providing evidence that representations are functionally used rather than merely correlated. Analysis of IRIS attention heads reveals spatial specialization: specific heads attend preferentially to tokens overlapping with game objects. Multi-baseline token ablation experiments consistently identify object-containing tokens as disproportionately important. Our findings provide interpretability evidence that learned world models develop structured, approximately linear internal representations of environment state across two games and two architectures.

Unmanned aerial vehicles (UAVs) have been widely used in urban missions, and proper planning of UAV paths can improve mission efficiency while reducing the risk of potential third-party impact. Existing work has considered all efficiency and safety objectives for a single decision-maker (DM) and regarded this as a multiobjective optimization problem (MOP). However, there is usually not a single DM but two DMs, i.e., an efficiency DM and a safety DM, and the DMs are only concerned with their respective objectives. The final decision is made based on the solutions of both DMs. In this paper, for the first time, biparty multiobjective UAV path planning (BPMO-UAVPP) problems involving both efficiency and safety departments are modeled. The existing multiobjective immune algorithm with nondominated neighbor-based selection (NNIA), the hybrid evolutionary framework for the multiobjective immune algorithm (HEIA), and the adaptive immune-inspired multiobjective algorithm (AIMA) are modified for solving the BPMO-UAVPP problem, and then biparty multiobjective optimization algorithms, including the BPNNIA, BPHEIA, and BPAIMA, are proposed and comprehensively compared with traditional multiobjective evolutionary algorithms and typical multiparty multiobjective evolutionary algorithms (i.e., OptMPNDS and OptMPNDS2). The experimental results show that BPAIMA performs better than ordinary multiobjective evolutionary algorithms such as NSGA-II and multiparty multiobjective evolutionary algorithms such as OptMPNDS, OptMPNDS2, BPNNIA and BPHEIA.

This paper studies generalization error bounds for Transformer models. Based on the offset Rademacher complexity, we derive sharper generalization bounds for different Transformer architectures, including single-layer single-head, single-layer multi-head, and multi-layer Transformers. We first express the excess risk of Transformers in terms of the offset Rademacher complexity. By exploiting its connection with the empirical covering numbers of the corresponding hypothesis spaces, we obtain excess risk bounds that achieve optimal convergence rates up to constant factors. We then derive refined excess risk bounds by upper bounding the covering numbers of Transformer hypothesis spaces using matrix ranks and matrix norms, leading to precise, architecture-dependent generalization bounds. Finally, we relax the boundedness assumption on feature mappings and extend our theoretical results to settings with unbounded (sub-Gaussian) features and heavy-tailed distributions.

We investigate the generalization capabilities of In-Context Operator Networks (ICONs), a new class of operator networks that build on the principles of in-context learning, for higher-order partial differential equations. We extend previous work by expanding the type and scope of differential equations handled by the foundation model. We demonstrate that while processing complex inputs requires some new computational methods, the underlying machine learning techniques are largely consistent with simpler cases. Our implementation shows that although point-wise accuracy degrades for higher-order problems like the heat equation, the model retains qualitative accuracy in capturing solution dynamics and overall behavior. This demonstrates the model's ability to extrapolate fundamental solution characteristics to problems outside its training regime.

Database Management Systems (DBMSs) are fundamental infrastructure for modern data-driven applications, where thorough testing with high-quality SQL test cases is essential for ensuring system reliability. Traditional approaches such as fuzzing can be effective for specific DBMSs, but adapting them to different proprietary dialects requires substantial manual effort. Large Language Models (LLMs) present promising opportunities for automated SQL test generation, but face critical challenges in industrial environments. First, lightweight models are widely used in organizations due to security and privacy constraints, but they struggle to generate syntactically valid queries for proprietary SQL dialects. Second, LLM-generated queries are often semantically similar and exercise only shallow execution paths, thereby quickly reaching a coverage plateau. To address these challenges, we propose MIST, an LLM-based test case generatIon framework for DBMS through Monte Carlo Tree search. MIST consists of two stages: Feature-Guided Error-Driven Test Case Synthetization, which constructs a hierarchical feature tree and uses error feedback to guide LLM generation, aiming to produce syntactically valid and semantically diverse queries for different DBMS dialects, and Monte Carlo Tree Search-Based Test Case Mutation, which jointly optimizes seed query selection and mutation rule application guided by coverage feedback, aiming at boosting code coverage by exploring deeper execution paths. Experiments on three widely-used DBMSs with four lightweight LLMs show that MIST achieves average improvements of 43.3% in line coverage, 32.3% in function coverage, and 46.4% in branch coverage compared to the baseline approach with the highest line coverage of 69.3% in the Optimizer module.

Psychiatric symptom identification on social media aims to infer fine-grained mental health symptoms from user-generated posts, allowing a detailed understanding of users' mental states. However, the construction of large-scale symptom-level datasets remains challenging due to the resource-intensive nature of expert labeling and the lack of standardized annotation guidelines, which in turn limits the generalizability of models to identify diverse symptom expressions from user-generated text. To address these issues, we propose SynSym, a synthetic data generation framework for constructing generalizable datasets for symptom identification. Leveraging large language models (LLMs), SynSym constructs high-quality training samples by (1) expanding each symptom into sub-concepts to enhance the diversity of generated expressions, (2) producing synthetic expressions that reflect psychiatric symptoms in diverse linguistic styles, and (3) composing realistic multi-symptom expressions, informed by clinical co-occurrence patterns. We validate SynSym on three benchmark datasets covering different styles of depressive symptom expression. Experimental results demonstrate that models trained solely on the synthetic data generated by SynSym perform comparably to those trained on real data, and benefit further from additional fine-tuning with real data. These findings underscore the potential of synthetic data as an alternative resource to real-world annotations in psychiatric symptom modeling, and SynSym serves as a practical framework for generating clinically relevant and realistic symptom expressions.

Modern concrete must simultaneously satisfy evolving demands for mechanical performance, workability, durability, and sustainability, making mix designs increasingly complex. Recent studies leveraging Artificial Intelligence (AI) and Machine Learning (ML) models show promise for predicting compressive strength and guiding mix optimization, but most existing efforts are based on proprietary industrial datasets and closed-source implementations. Here we introduce BOxCrete, an open-source probabilistic modeling and optimization framework trained on a new open-access dataset of over 500 strength measurements (1-15 ksi) from 123 mixtures - 69 mortar and 54 concrete mixes tested at five curing ages (1, 3, 5, 14, and 28 days). BOxCrete leverages Gaussian Process (GP) regression to predict strength development, achieving average R$^2$ = 0.94 and RMSE = 0.69 ksi, quantify uncertainty, and carry out multi-objective optimization of compressive strength and embodied carbon. The dataset and model establish a reproducible open-source foundation for data-driven development of AI-based optimized mix designs.

Large Language Models (LLMs) are deployed in high-stakes settings but can show demographic, gender, and geographic biases that undermine fairness and trust. Prior debiasing methods, including embedding-space projections, prompt-based steering, and causal interventions, often act at a single stage of the pipeline, resulting in incomplete mitigation and brittle utility trade-offs under distribution shifts. We propose CatRAG Debiasing, a dual-pronged framework that integrates functor with Retrieval-Augmented Generation (RAG) guided structural debiasing. The functor component leverages category-theoretic structure to induce a principled, structure-preserving projection that suppresses bias-associated directions in the embedding space while retaining task-relevant semantics. On the Bias Benchmark for Question Answering (BBQ) across three open-source LLMs (Meta Llama-3, OpenAI GPT-OSS, and Google Gemma-3), CatRAG achieves state-of-the-art results, improving accuracy by up to 40% over the corresponding base models and by more than 10% over prior debiasing methods, while reducing bias scores to near zero (from 60% for the base models) across gender, nationality, race, and intersectional subgroups.

Large Language Models (LLMs), deep learning architectures with typically over 10 billion parameters, have recently begun to be integrated into various cyber-physical systems (CPS) such as robotics, industrial automation, and autopilot systems. The abstract knowledge and reasoning capabilities of LLMs are employed for tasks like planning and navigation. However, a significant challenge arises from the tendency of LLMs to produce "hallucinations" - outputs that are coherent yet factually incorrect or contextually unsuitable. This characteristic can lead to undesirable or unsafe actions in the CPS. Therefore, our research focuses on assuring the LLM-enabled CPS by enhancing their critical properties. We propose SafePilot, a novel hierarchical neuro-symbolic framework that provides end-to-end assurance for LLM-enabled CPS according to attribute-based and temporal specifications. Given a task and its specification, SafePilot first invokes a hierarchical planner with a discriminator that assesses task complexity. If the task is deemed manageable, it is passed directly to an LLM-based task planner with built-in verification. Otherwise, the hierarchical planner applies a divide-and-conquer strategy, decomposing the task into sub-tasks, each of which is individually planned and later merged into a final solution. The LLM-based task planner translates natural language constraints into formal specifications and verifies the LLM's output against them. If violations are detected, it identifies the flaw, adjusts the prompt accordingly, and re-invokes the LLM. This iterative process continues until a valid plan is produced or a predefined limit is reached. Our framework supports LLM-enabled CPS with both attribute-based and temporal constraints. Its effectiveness and adaptability are demonstrated through two illustrative case studies.

Large Language Models (LLM)-based Multi-Agent Systems (MASs) have emerged as a new paradigm in software system design, increasingly demonstrating strong reasoning and collaboration capabilities. As these systems become more complex and autonomous, effective failure management is essential to ensure reliability and availability. However, existing approaches often rely on per-trace reasoning, which leads to low efficiency, and neglect historical failure patterns, limiting diagnostic accuracy. In this paper, we conduct a preliminary empirical study to demonstrate the necessity, potential, and challenges of leveraging historical failure patterns to enhance failure management in MASs. Building on this insight, we propose \textbf{EAGER}, an efficient failure management framework for multi-agent systems based on reasoning trace representation. EAGER employs unsupervised reasoning-scoped contrastive learning to encode both intra-agent reasoning and inter-agent coordination, enabling real-time step-wise failure detection, diagnosis, and reflexive mitigation guided by historical failure knowledge. Preliminary evaluations on three open-source MASs demonstrate the effectiveness of EAGER and highlight promising directions for future research in reliable multi-agent system operations.

Automatic prompt optimization is a promising approach for adapting large language models (LLMs) to downstream tasks, yet existing methods typically search for a specific prompt specialized to a fixed task. This paradigm limits generalization across heterogeneous queries and prevents models from accumulating reusable prompting knowledge over time. In this paper, we propose MemAPO, a memory-driven framework that reconceptualizes prompt optimization as generalizable and self-evolving experience accumulation. MemAPO maintains a dual-memory mechanism that distills successful reasoning trajectories into reusable strategy templates while organizing incorrect generations into structured error patterns that capture recurrent failure modes. Given a new prompt, the framework retrieves both relevant strategies and failure patterns to compose prompts that promote effective reasoning while discouraging known mistakes. Through iterative self-reflection and memory editing, MemAPO continuously updates its memory, enabling prompt optimization to improve over time rather than restarting from scratch for each task. Experiments on diverse benchmarks show that MemAPO consistently outperforms representative prompt optimization baselines while substantially reducing optimization cost.

Analyzing news coverage in multilingual societies can offer valuable insights into the dynamics of public discourse and the development of collective narratives, yet comprehensive studies that account for linguistic and cultural diversity within national media ecosystems remain limited, particularly in complex contexts such as Switzerland. This paper studies temporal trends in Swiss digital media across the country's three main linguistic regions, French, German, and Italian, using a triangulated methodology that combines quantitative analyses with qualitative insights. We collected and processed over 1.7 million news articles, applying lexical metrics, named entity recognition and Wikidata-based linking, targeted sentiment analysis, and consensus-based change-point detection. To enable principled cross-language comparisons and to connect to theories of domestication and cultural proximity, we derive domestication profiles together with a proximity salience ratio. Our analysis spans thematic, recurrent, and singular events. By integrating quantitative data with qualitative interpretation, we provide new insights into the dynamics of Swiss digital media and demonstrate the usefulness of triangulation in media studies. The findings reveal distinct temporal patterns and highlight how linguistic and cultural contexts influence reporting. Our approach offers a framework applicable to other multilingual or culturally diverse media environments, contributing to a deeper understanding of how news is shaped by linguistic and cultural factors.

Machine learning models are widely integrated into modern mobile apps to analyze user behaviors and deliver personalized services. Ensuring low-latency on-device model execution is critical for maintaining high-quality user experiences. While prior research has primarily focused on accelerating model inference with given input features, we identify an overlooked bottleneck in real-world on-device model execution pipelines: extracting input features from raw application logs. In this work, we explore a new direction of feature extraction optimization by analyzing and eliminating redundant extraction operations across different model features and consecutive model inferences. We then introduce AutoFeature, an automated feature extraction engine designed to accelerate on-device feature extraction process without compromising model inference accuracy. AutoFeature comprises three core designs: (1) graph abstraction to formulate the extraction workflows of different input features as one directed acyclic graph, (2) graph optimization to identify and fuse redundant operation nodes across different features within the graph; (3) efficient caching to minimize operations on overlapping raw data between consecutive model inferences. We implement a system prototype of AutoFeature and integrate it into five industrial mobile services spanning search, video and e-commerce domains. Online evaluations show that AutoFeature reduces end-to-end on-device model execution latency by 1.33x-3.93x during daytime and 1.43x-4.53x at night.

Overparameterized shallow neural networks admit substantial parameter redundancy: distinct parameter vectors may represent the same predictor due to hidden-unit permutations, rescalings, and related symmetries. As a result, geometric quantities computed directly in the ambient Euclidean parameter space can reflect artifacts of representation rather than intrinsic properties of the predictor. In this paper, we develop a differential-geometric framework for analyzing simple shallow networks through the quotient space obtained by modding out parameter symmetries on a regular set. We first characterize the symmetry and quotient structure of regular shallow-network parameters and show that the finite-sample realization map induces a natural metric on the quotient manifold. This leads to an effective notion of curvature that removes degeneracy along symmetry orbits and yields a symmetry-reduced Hessian capturing intrinsic local geometry. We then study gradient flows on the quotient and show that only the horizontal component of parameter motion contributes to first-order predictor evolution, while the vertical component corresponds purely to gauge variation. Finally, we formulate an implicit-bias viewpoint at the quotient level, arguing that meaningful complexity should be assigned to predictor classes rather than to individual parameter representatives. Our experiments confirm that ambient flatness is representation-dependent, that local dynamics are better organized by quotient-level curvature summaries, and that in underdetermined regimes, implicit bias is most naturally described in quotient coordinates.

Quantum error correcting codes (QECC) are essential for constructing large-scale quantum computers that deliver faithful results. As strong competitors to the conventional surface code, quantum low-density parity-check (qLDPC) codes are emerging rapidly: they offer high encoding rates while maintaining reasonable physical-qubit connectivity requirements. Despite the existence of numerous code constructions, a notable gap persists between these designs -- some of which remain purely theoretical -- and their circuit-level deployment. In this work, we propose Auto-Stabilizer-Check (ASC), a universal compilation framework that generates depth-optimal syndrome extraction circuits for arbitrary qLDPC codes. ASC leverages the sparsity of parity-check matrices and exploits the commutativity of X and Z stabilizer measurement subroutines to search for optimal compilation schemes. By iteratively invoking an SMT solver, ASC returns a depth-optimal solution if a satisfying assignment is found, and a near-optimal solution in cases of solver timeouts. Notably, ASC provides the first definitive answer to one of IBM's open problems: for all instances of bivariate bicycle (BB) code reported in their work, our compiler certifies that no depth-6 syndrome extraction circuit exists. Furthermore, by integrating ASC with an end-to-end evaluation framework -- one that assesses different compilation settings under a circuit-level noise model -- ASC reduces circuit depth by approximately 50% and achieves an average 7x-8x suppression of the logical error rate for general qLDPC codes, compared with as-soon-as-possible (ASAP) and coloration-based scheduling. ASC thus substantially reduces manual design overhead and demonstrates its strong potential to serve as a key component in accelerating hardware deployment of qLDPC codes.

Robotic automation has transformed scientific workflows in domains such as chemistry and materials science, yet free-space optics, which is a high precision domain, remains largely manual. Optical systems impose strict spatial and angular tolerances, and their performance is governed by tightly coupled physical parameters, making generalizable automation particularly challenging. In this work, we present a robotics framework for the autonomous construction, alignment, and maintenance of precision optical systems. Our approach integrates hierarchical computer vision systems, optimization routines, and custom-built tools to achieve this functionality. As a representative demonstration, we perform the fully autonomous construction of a tabletop laser cavity from randomly distributed components. The system performs several tasks such as laser beam centering, spatial alignment of multiple beams, resonator alignment, laser mode selection, and self-recovery from induced misalignment and disturbances. By achieving closed-loop autonomy for highly sensitive optical systems, this work establishes a foundation for autonomous optical experiments for applications across technical domains.

Modern software systems operate at unprecedented scale and complexity, where effective failure management is critical yet increasingly challenging. Metrics, traces, and logs provide complementary views of system runtime behavior, but existing failure management approaches typically rely on task-oriented pipelines that tightly couple modality-specific preprocessing, representation learning, and downstream models, resulting in limited generalization across tasks and systems. To fill this gap, we propose RuntimeSlicer, a unified runtime state representation model towards generalizable failure management. RuntimeSlicer pre-trains a task-agnostic representation model that directly encodes metrics, traces, and logs into a single, aligned system-state embedding capturing the holistic runtime condition of the system. To train RuntimeSlicer, we introduce Unified Runtime Contrastive Learning, which integrates heterogeneous training data sources and optimizes complementary objectives for cross-modality alignment and temporal consistency. Building upon the learned system-state embeddings, we further propose State-Aware Task-Oriented Tuning, which performs unsupervised partitioning of runtime states and enables state-conditioned adaptation for downstream tasks. This design allows lightweight task-oriented models to be trained on top of the unified embedding without redesigning modality-specific encoders or preprocessing pipelines. Preliminary experiments on the AIOps 2022 dataset demonstrate the feasibility and effectiveness of RuntimeSlicer for system state modeling and failure management tasks.

The Brain Tumor Reporting and Data System (BT-RADS) standardizes post-treatment MRI response assessment in patients with diffuse gliomas but requires complex integration of imaging trends, medication effects, and radiation timing. This study evaluates an end-to-end multi-agent large language model (LLM) and convolutional neural network (CNN) system for automated BT-RADS classification. A multi-agent LLM system combined with automated CNN-based tumor segmentation was retrospectively evaluated on 509 consecutive post-treatment glioma MRI examinations from a single high-volume center. An extractor agent identified clinical variables (steroid status, bevacizumab status, radiation date) from unstructured clinical notes, while a scorer agent applied BT-RADS decision logic integrating extracted variables with volumetric measurements. Expert reference standard classifications were established by an independent board-certified neuroradiologist. Of 509 examinations, 492 met inclusion criteria. The system achieved 374/492 (76.0%; 95% CI, 72.1%-79.6%) accuracy versus 283/492 (57.5%; 95% CI, 53.1%-61.8%) for initial clinical assessments (+18.5 percentage points; P<.001). Context-dependent categories showed high sensitivity (BT-1b 100%, BT-1a 92.7%, BT-3a 87.5%), while threshold-dependent categories showed moderate sensitivity (BT-3c 74.8%, BT-2 69.2%, BT-4 69.3%, BT-3b 57.1%). For BT-4, positive predictive value was 92.9%. The multi-agent LLM system achieved higher BT-RADS classification agreement with expert reference standard compared to initial clinical scoring, with high accuracy for context-dependent scores and high positive predictive value for BT-4 detection.

Identifying the most representative subset for a close-to-submodular objective while satisfying the predefined partition constraint is a fundamental task with numerous applications in machine learning. However, the existing distorted local-search methods are often hindered by their prohibitive query complexities and the rigid requirement for prior knowledge of difficult-to-obtain structural parameters. To overcome these limitations, we introduce a novel algorithm titled Multinoulli-SCG, which not only is parameter-free, but also can achieve the same approximation guarantees as the distorted local-search methods with significantly fewer function evaluations. More specifically, when the objective function is monotone $α$-weakly DR-submodular or $(γ,β)$-weakly submodular, our Multinoulli-SCG algorithm can attain a value of $(1-e^{-α})\text{OPT}-ε$ or $(\frac{γ^{2}(1-e^{-(β(1-γ)+γ^2)})}{β(1-γ)+γ^2})\text{OPT}-ε$ with only $O(1/ε^{2})$ function evaluations, where OPT denotes the optimal value. The cornerstone of our Multinoulli-SCG algorithm is an innovative continuous-relaxation framework named Multinoulli Extension(ME), which can effectively convert the discrete subset selection problem subject to partition constraints into a solvable continuous maximization focused on learning the optimal multinoulli priors across the concerned partition. In sharp contrast with the well-established multi-linear extension for submodular subset selection, a notable advantage of our proposed ME is its intrinsic capacity to provide a lossless rounding scheme for any set function. Furthermore, based on our proposed ME, we also present two novel online algorithms, namely, Multinoulli-OSCG and Multinoulli-OSGA, for the unexplored online subset selection problems over partition constraints.

Learning systems are typically optimized by minimizing loss or maximizing reward, assuming that improvements in these signals reflect progress toward the true objective. However, when feedback reliability is unobservable, this assumption can fail, and learning algorithms may converge stably to incorrect solutions. This failure arises because single-step feedback does not reveal whether an experience is informative or persistently biased. When information is aggregated over learning trajectories, however, systematic differences between reliable and unreliable regimes can emerge. We propose a Monitor-Trust-Regulator (MTR) framework that infers reliability from learning dynamics and modulates updates through a slow-timescale trust variable. Across reinforcement learning and supervised learning settings, standard algorithms exhibit stable optimization behavior while learning incorrect solutions under latent unreliability, whereas trust-modulated systems reduce bias accumulation and improve recovery. These results suggest that learning dynamics are not only optimization traces but also a source of information about feedback reliability.

AI agents have become increasingly capable at isolated software engineering (SWE) tasks such as resolving issues on Github. Yet long-horizon tasks involving multiple interdependent subtasks still pose challenges both with respect to accuracy, and with respect to timely completion. A natural approach to solving these long-horizon tasks in a timely manner is asynchronous multi-agent collaboration, where multiple agents work on different parts of the task at the same time. But effective application of multi-agent systems has proven surprisingly difficult: concurrent edits by multiple agents interfere with each other, dependencies are difficult to synchronize, and combining partial progress into a coherent whole is challenging. On the other hand, human developers have long relied on mature collaboration infrastructure to manage these challenges in large software projects. Inspired by these collaboration primitives, we introduce Centralized Asynchronous Isolated Delegation (CAID), a structured multi-agent coordination paradigm grounded in three core SWE primitives: centralized task delegation, asynchronous execution, and isolated workspaces. CAID constructs dependency-aware task plans through a central manager, executes subtasks concurrently in isolated workspaces, and consolidates progress via structured integration with executable test-based verification. In empirical evaluation, we find that CAID improves accuracy over single-agent baselines by 26.7% absolute on paper reproduction tasks (PaperBench) and 14.3% on Python library development tasks (Commit0). Through systematic analysis, we find that branch-and-merge is a central coordination mechanism for multi-agent collaboration, and that SWE primitives such as git worktree, git commit, and git merge enable it to be realized in a reliable and executable manner.

We present \emph{GaussianSSC}, a two-stage, grid-native and triplane-guided approach to semantic scene completion (SSC) that injects the benefits of Gaussians without replacing the voxel grid or maintaining a separate Gaussian set. We introduce \emph{Gaussian Anchoring}, a sub-pixel, Gaussian-weighted image aggregation over fused FPN features that tightens voxel--image alignment and improves monocular occupancy estimation. We further convert point-like voxel features into a learned per-voxel Gaussian field and refine triplane features via a triplane-aligned \emph{Gaussian--Triplane Refinement} module that combines \emph{local gathering} (target-centric) and \emph{global aggregation} (source-centric). This directional, anisotropic support captures surface tangency, scale, and occlusion-aware asymmetry while preserving the efficiency of triplane representations. On SemanticKITTI~\cite{behley2019semantickitti}, GaussianSSC improves Stage~1 occupancy by +1.0\% Recall, +2.0\% Precision, and +1.8\% IoU over state-of-the-art baselines, and improves Stage~2 semantic prediction by +1.8\% IoU and +0.8\% mIoU.

Off-Policy Evaluation (OPE) is an important practical problem in algorithmic ranking systems, where the goal is to estimate the expected performance of a new ranking policy using only offline logged data collected under a different, logging policy. Existing estimators, such as the ranking-wise and position-wise inverse propensity score (IPS) estimators, require the data collection policy to be sufficiently stochastic and suffer from severe bias when the logging policy is fully deterministic. In this paper, we propose novel estimators, Click-based Inverse Propensity Score (CIPS), exploiting the intrinsic stochasticity of user click behavior to address this challenge. Unlike existing methods that rely on the stochasticity of the logging policy, our approach uses click probability as a new form of importance weighting, enabling low-bias OPE even under deterministic logging policies where existing methods incur substantial bias. We provide theoretical analyses of the bias and variance properties of the proposed estimators and show, through synthetic and real-world experiments, that our estimators achieve significantly lower bias compared to strong baselines, for a range of experimental settings with completely deterministic logging policies.

Speech technologies have advanced rapidly and serve diverse populations worldwide. However, many languages remain underrepresented due to limited resources. In this paper, we introduce \textbf{TaigiSpeech}, a real-world speech intent dataset in Taiwanese Taigi (aka Taiwanese Hokkien/Southern Min), which is a low-resource and primarily spoken language. The dataset is collected from older adults, comprising 21 speakers with a total of 3k utterances. It is designed for practical intent detection scenarios, including healthcare and home assistant applications. To address the scarcity of labeled data, we explore two data mining strategies with two levels of supervision: keyword match data mining with LLM pseudo labeling via an intermediate language and an audio-visual framework that leverages multimodal cues with minimal textual supervision. This design enables scalable dataset construction for low-resource and unwritten spoken languages. TaigiSpeech will be released under the CC BY 4.0 license to facilitate broad adoption and research on low-resource and unwritten languages. The project website and the dataset can be found on https://kwchang.org/taigispeech.

Automatic Multi-Agent Systems (MAS) generation has emerged as a promising paradigm for solving complex reasoning tasks. However, existing frameworks are fundamentally bottlenecked when applied to knowledge-intensive domains (e.g., healthcare and law). They either rely on a static library of general nodes like Chain-of-Thought, which lack specialized expertise, or attempt to generate nodes on the fly. In the latter case, the orchestrator is not only bound by its internal knowledge limits but must also simultaneously generate domain-specific logic and optimize high-level topology, leading to a severe architectural coupling that degrades overall system efficacy. To bridge this gap, we propose Unified-MAS that decouples granular node implementation from topological orchestration via offline node synthesis. Unified-MAS operates in two stages: (1) Search-Based Node Generation retrieves external open-world knowledge to synthesize specialized node blueprints, overcoming the internal knowledge limits of LLMs; and (2) Reward-Based Node Optimization utilizes a perplexity-guided reward to iteratively enhance the internal logic of bottleneck nodes. Extensive experiments across four specialized domains demonstrate that integrating Unified-MAS into four Automatic-MAS baselines yields a better performance-cost trade-off, achieving up to a 14.2% gain while significantly reducing costs. Further analysis reveals its robustness across different designer LLMs and its effectiveness on conventional tasks such as mathematical reasoning.

This paper proposes a refutation-validated framework for aspect-based sentiment analysis in financial markets, addressing the limitations of correlational studies that cannot distinguish genuine associations from spurious ones. Using X data for the energy sector, we test whether aspect-level sentiment signals show robust, refutation-validated relationships with equity returns. Our pipeline combines net-ratio scoring with z-normalization, OLS with Newey West HAC errors, and refutation tests including placebo, random common cause, subset stability, and bootstrap. Across six energy tickers, only a few associations survive all checks, while renewables show aspect and horizon specific responses. While not establishing causality, the framework provides statistically robust, directionally interpretable signals, with limited sample size (six stocks, one quarter) constraining generalizability and framing this work as a methodological proof of concept.

Developing efficient CUDA kernels is a fundamental yet challenging task in the generative AI industry. Recent researches leverage Large Language Models (LLMs) to automatically convert PyTorch reference implementations to CUDA kernels, significantly reducing the engineering efforts. State-of-the-art LLMs, such as GPT-5.2 and Claude-Sonnet-4.5, still struggle in this specific task. To address this challenge, we propose DRTriton, a scalable learning framework for training LLMs to convert PyTorch codes into highly optimized Triton kernels, which are then compiled to CUDA kernels at runtime. DRTriton consists of three key components: (i) a data synthetic algorithm CSP-DAG that guarantees full coverage and unbiased uniform sampling over the operator space with controlled difficulty; (ii) a curriculum reinforcement learning with decoupled reward efficiently optimizes conversion success rate and inference speed simultaneously; and (iii) a test-time search algorithm that further improves the inference speed of the generated Triton kernels. Notably, despite being trained exclusively on synthetic data, DRTriton generalizes effectively to real-world CUDA kernels that are challenging even for human experts. Experimental results show that DRTriton-7B achieves speedup on 92% of the KernelBench Level 2, compared to 23% for GPT-5.2 and 19% for Claude-Sonnet-4.5.

Preference alignment is usually achieved by weight-updating training on preference data, which adds substantial alignment-stage compute and provides limited mechanistic visibility. We propose Dynamic SAE Steering for Preference Alignment (DSPA), an inference-time method that makes sparse autoencoder (SAE) steering prompt-conditional. From preference triples, DSPA computes a conditional-difference map linking prompt features to generation-control features; during decoding, it modifies only token-active latents, without base-model weight updates. Across Gemma-2-2B/9B and Qwen3-8B, DSPA improves MT-Bench and is competitive on AlpacaEval while preserving multiple-choice accuracy. Under restricted preference data, DSPA remains robust and can rival the two-stage RAHF-SCIT pipeline while requiring up to $4.47\times$ fewer alignment-stage FLOPs. Finally, we audit the SAE features DSPA modifies, finding that preference directions are dominated by discourse and stylistic signals, and provide theory clarifying the conditional-difference map estimate and when top-$k$ ablation is principled.

Patient education materials for solid-organ transplantation vary substantially across U.S. centers, yet no systematic method exists to quantify this heterogeneity at scale. We introduce a framework that grounds the same patient questions in different centers' handbooks using retrieval-augmented language models and compares the resulting answers using a five-label consistency taxonomy. Applied to 102 handbooks from 23 centers and 1,115 benchmark questions, the framework quantifies heterogeneity across four dimensions: question, topic, organ, and center. We find that 20.8% of non-absent pairwise comparisons exhibit clinically meaningful divergence, concentrated in condition monitoring and lifestyle topics. Coverage gaps are even more prominent: 96.2% of question-handbook pairs miss relevant content, with reproductive health at 95.1% absence. Center-level divergence profiles are stable and interpretable, where heterogeneity reflects systematic institutional differences, likely due to patient diversity. These findings expose an information gap in transplant patient education materials, with document-grounded medical question answering highlighting opportunities for content improvement.

Despite offering high sensitivity, a high signal-to-noise ratio, and a broad spectral range, single-pixel imaging (SPI) is limited by low measurement efficiency and long data-acquisition times. To address this, we propose a wavelength-multiplexed, spatially incoherent diffractive optical processor combined with a compact/shallow digital artificial neural network (ANN) to implement compressive SPI. Specifically, we model the bucket detection process in conventional SPI as a linear intensity transformation with spatially and spectrally varying point-spread functions. This transformation matrix is treated as a learnable parameter and jointly optimized with a shallow digital ANN composed of 2 hidden nonlinear layers. The wavelength-multiplexed diffractive processor is then configured via data-free optimization to approximate this pre-trained transformation matrix; after this optimization, the diffractive processor remains static/fixed. Upon multi-wavelength illumination and diffractive modulation, the target spatial information of the input object is spectrally encoded. A single-pixel detector captures the output spectral power at each illumination band, which is then rapidly decoded by the jointly trained digital ANN to reconstruct the input image. In addition to our numerical analyses demonstrating the feasibility of this approach, we experimentally validated its proof-of-concept using an array of light-emitting diodes (LEDs). Overall, this work demonstrates a computational imaging framework for compressive SPI that can be useful in applications such as biomedical imaging, autonomous devices, and remote sensing.

LLM coding benchmarks face a credibility crisis: widespread solution leakage and test quality issues undermine SWE-bench Verified, while existing detection methods--paraphrase consistency, n-gram overlap, perplexity analysis--never directly observe whether a model reasons or recalls. Meanwhile, simply repeating verification degrades accuracy: multi-turn review generates false positives faster than it discovers true errors, suggesting that structural approaches are needed. We introduce Cross-Context Verification (CCV), a black-box method that solves the same benchmark problem in N independent sessions and measures solution diversity, combined with the Hierarchical Cross-Context Architecture (HCCA), a multi-agent analysis framework that prevents confirmation bias through intentional information restriction across specialized analytical roles. On 9 SWE-bench Verified problems (45 trials, Claude Opus 4.6, temperature 0), CCV achieves perfect separation between contaminated and genuine reasoning (Mann-Whitney U=0, p approx 0.012, r = 1.0). Key findings: (1) contamination is binary--models either recall perfectly or not at all; (2) reasoning absence is a perfect discriminator; (3) 33% of prior contamination labels are false positives; (4) HCCA's independent analysis structure discovers contamination-flaw composite cases that single-analyst approaches miss. A pilot experiment extending HCCA to multi-stage verification (Worker to Verifier to Director) yields a negative result--100% sycophantic confirmation--providing further evidence that information restriction, not structural complexity, is the key mechanism. We release all code and data.

We introduce a new paradigm for task-oriented dialogue systems: safety certification as a computational primitive for answer reuse. Current systems treat each turn independently, recomputing answers via retrieval or generation even when they are already derivable from prior state. We show that in capability-based systems, the safety certification step computes a fixed-point closure cl(At) that already contains every answer reachable from the current configuration. We operationalize this insight with a Certified Answer Store (CAS) augmented by Pre-Answer Blocks (PAB): at each certified turn, the system materializes all derivable follow-up answers together with minimal provenance witnesses. Subsequent queries are answered in sub-millisecond time via formal containment checks, eliminating redundant retrieval and generation.

The multiplication of two sparse matrices, known as SpGEMM, is a key kernel in scientific computing and large-scale data analytics, underpinning graph algorithms, machine learning, simulations, and computational biology, where sparsity is often highly unstructured. The unstructured sparsity makes achieving high performance challenging because it limits both memory efficiency and scalability. In distributed memory, the cost of exchanging and merging partial products across nodes further constrains performance. These issues are exacerbated on modern heterogeneous supercomputers with deep, hierarchical GPU interconnects. Current SpGEMM implementations overlook the gap between intra-node and inter-node bandwidth, resulting in unnecessary data movement and synchronization not fully exploiting the fast intra-node interconnect. To address these challenges, we introduce Trident, a hierarchy-aware 2D distributed SpGEMM algorithm that uses communication-avoiding techniques and asynchronous communication to exploit the hierarchical and heterogeneous architecture of modern supercomputing interconnect. Central to Trident is the novel trident partitioning scheme, which enables hierarchy-aware decomposition and reduces internode communication by leveraging the higher bandwidth between GPUs within a node compared to across nodes. Here, we evaluate Trident on unstructured matrices, achieving up to $2.38\times$ speedup over a 2D SpGEMM with a corresponding geometric mean speedup of $1.54\times$. Trident reduces internode communication volume by up to $2\times$ on NERSC's Perlmutter supercomputer. Furthermore, we demonstrate the effectiveness of Trident in speeding up Markov Clustering, achieving up to $2\times$ speedup compared to competing strategies.

We prove that livelock detection is \emph{decidable in polynomial time} for parameterized symmetric unidirectional rings of self-disabling processes with bounded domain $\mathbb{Z}_m$. Given a protocol specified by its set of local transitions $T$, the algorithm decides whether a livelock exists for \emph{some} ring size $K\!\geq\!2$, running in $O(|T|^3)$ time independent of $K$. The algorithm computes the greatest fixed point of a deflationary monotone operator on the finite set $T$ and returns \emph{livelock} iff the fixed point is non-empty. The livelock freedom argument rests on maximality: the fix-point is the largest set of transitions that can together sustain a pseudolivelock at every process; its emptiness certifies freedom for all $K$ without any search over ring sizes. The work is grounded in the algebraic characterization of livelocks from Farahat~\citep{farahat2012}, which establishes necessary and sufficient conditions for livelock existence but does not address decidability. We also handle the $(1,1)$-asymmetric case in which one distinguished process $P_0$ differs from the remaining $K\!-\!1$ identical processes. Code and algebraic foundation are at the URL: https://github.com/cosmoparadox/mathematical-tools.

Large Language Models (LLMs) demonstrate impressive natural language capabilities but often struggle with knowledge-intensive reasoning tasks. Knowledge Base Question Answering (KBQA), which leverages structured Knowledge Graphs (KGs) exemplifies this challenge due to the need for accurate multi-hop reasoning. Existing approaches typically perform sequential reasoning steps guided by predefined pipelines, restricting flexibility and causing error cascades due to isolated reasoning at each step. To address these limitations, we propose KG-Hopper, a novel Reinforcement Learning (RL) framework that empowers compact open LLMs with the ability to perform integrated multi-hop KG reasoning within a single inference round. Rather than reasoning step-by-step, we train a Reasoning LLM that embeds the entire KG traversal and decision process into a unified ``thinking'' stage, enabling global reasoning over cross-step dependencies and dynamic path exploration with backtracking. Experimental results on eight KG reasoning benchmarks show that KG-Hopper, based on a 7B-parameter LLM, consistently outperforms larger multi-step systems (up to 70B) and achieves competitive performance with proprietary models such as GPT-3.5-Turbo and GPT-4o-mini, while remaining compact, open, and data-efficient. The code is publicly available at: https://github.com/Wangshuaiia/KG-Hopper.

Multidisciplinary Software Development (MSD) requires domain experts and developers to collaborate across incompatible formalisms and separate artifact sets. Today, even with AI coding assistants like GitHub Copilot, this process remains inefficient; individual coding tasks are semi-automated, but the workflow connecting domain knowledge to implementation is not. Developers and experts still lack a shared view, resulting in repeated coordination, clarification rounds, and error-prone handoffs. We address this gap through a graph-based workflow optimization approach that progressively replaces manual coordination with LLM-powered services, enabling incremental adoption without disrupting established practices. We evaluate our approach on \texttt{spapi}, a production in-vehicle API system at Volvo Group involving 192 endpoints, 420 properties, and 776 CAN signals across six functional domains. The automated workflow achieves 93.7\% F1 score while reducing per-API development time from approximately 5 hours to under 7 minutes, saving an estimated 979 engineering hours. In production, the system received high satisfaction from both domain experts and developers, with all participants reporting full satisfaction with communication efficiency.

To discover the weaknesses of LLMs, researchers often embed prompts into a vector space and cluster them to extract insightful patterns. However, vector embeddings primarily capture topical similarity. As a result, prompts that share a topic but differ in specificity, and consequently in difficulty, are often represented similarly, making fine-grained weakness analysis difficult. To address this limitation, we propose PROMPT2BOX, which embeds prompts into a box embedding space using a trained encoder. The encoder, trained on existing and synthesized datasets, outputs box embeddings that capture not only semantic similarity but also specificity relations between prompts (e.g., "writing an adventure story" is more specific than "writing a story"). We further develop a novel dimension reduction technique for box embeddings to facilitate dataset visualization and comparison. Our experiments demonstrate that box embeddings consistently capture prompt specificity better than vector baselines. On the downstream task of creating hierarchical clustering trees for 17 LLMs from the UltraFeedback dataset, PROMPT2BOX can identify 8.9\% more LLM weaknesses than vector baselines and achieves an approximately 33\% stronger correlation between hierarchical depth and instruction specificity.

Transformer-based embedding models rely on pooling to map variable-length text into a single vector, enabling efficient similarity search but also inducing well-known geometric pathologies such as anisotropy and length-induced embedding collapse. Existing accounts largely describe \emph{what} these pathologies look like, yet provide limited insight into \emph{when} and \emph{why} they harm downstream retrieval. In this work, we argue that the missing causal factor is \emph{semantic shift}: the intrinsic, structured evolution and dispersion of semantics within a text. We first present a theoretical analysis of \emph{semantic smoothing} in Transformer embeddings: as the semantic diversity among constituent sentences increases, the pooled representation necessarily shifts away from every individual sentence embedding, yielding a smoothed and less discriminative vector. Building on this foundation, we formalize semantic shift as a computable measure integrating local semantic evolution and global semantic dispersion. Through controlled experiments across corpora and multiple embedding models, we show that semantic shift aligns closely with the severity of embedding concentration and predicts retrieval degradation, whereas text length alone does not. Overall, semantic shift offers a unified and actionable lens for understanding embedding collapse and for diagnosing when anisotropy becomes harmful.

When organisations adopt commercial AI systems for decision support, they inherit value judgements embedded by vendors that are neither transparent nor renegotiable. The governance puzzle is not whether AI can support decisions but which recommendations the system can actually produce given how its vendor has configured it. I formalise this as a behavioural feasible set, the range of recommendations reachable under vendor-imposed alignment constraints, and characterise diagnostic thresholds for when organisational requirements exceed the system's flexibility. In scenario-based experiments using binary decision scenarios and multi-stakeholder ranking tasks, I show that alignment materially compresses this set. Comparing pre- and post-alignment variants of an open-weight model isolates the mechanism: alignment makes the system substantially less able to shift its recommendation even under legitimate contextual pressure. Leading commercial models exhibit comparable or greater rigidity. In multi-stakeholder tasks, alignment shifts implied stakeholder priorities rather than neutralising them, meaning organisations adopt embedded value orientations set upstream by the vendor. Organisations thus face a governance problem that better prompting cannot resolve: selecting a vendor partially determines which trade-offs remain negotiable and which stakeholder priorities are structurally embedded.

Large language models (LLMs) have shown impressive capabilities in code generation. However, because most LLMs are trained on public domain corpora, directly applying them to real-world software development often yields low success rates, as these scenarios frequently require domain-specific knowledge. In particular, domain-specific tasks usually demand highly specialized solutions, which are often underrepresented or entirely absent in the training data of generic LLMs. To address this challenge, we propose DomAgent, an autonomous coding agent that bridges this gap by enabling LLMs to generate domain-adapted code through structured reasoning and targeted retrieval. A core component of DomAgent is DomRetriever, a novel retrieval module that emulates how humans learn domain-specific knowledge, by combining conceptual understanding with experiential examples. It dynamically integrates top-down knowledge-graph reasoning with bottom-up case-based reasoning, enabling iterative retrieval and synthesis of structured knowledge and representative cases to ensure contextual relevance and broad task coverage. DomRetriever can operate as part of DomAgent or independently with any LLM for flexible domain adaptation. We evaluate DomAgent on an open benchmark dataset in the data science domain (DS-1000) and further apply it to real-world truck software development tasks. Experimental results show that DomAgent significantly enhances domain-specific code generation, enabling small open-source models to close much of the performance gap with large proprietary LLMs in complex, real-world applications. The code is available at: https://github.com/Wangshuaiia/DomAgent.

Extensive simulation-based testing is important for assuring the safety of autonomous driving systems (ADS). However, generating safety-critical traffic scenarios remains challenging because failures often arise from rare, complex interactions with surrounding vehicles. Existing automatic scenario-generation approaches frequently fail to distinguish genuine ADS faults from collisions caused by implausible or invalid adversarial behaviors, and they typically optimize either scenario initialization or agent behavior in isolation. We propose Dynasto, a two-step testing approach that jointly optimizes initial scenario parameters and dynamic adversarial behaviors to uncover realistic safety-critical failures. First, we train an adversarial agent using reinforcement learning (RL) with temporal-logic-based validity criteria and a safe-distance model inspired by ISO 34502 to promote behaviorally plausible failures. Second, a genetic algorithm (GA) searches over initial conditions while replaying the adversary's failure-inducing behaviors to reveal additional failures that the RL agent alone does not uncover. Finally, a graph-based clustering pipeline groups failures into representative modes based on semantic event sequences. Our evaluation experiments in HighwayEnv across two ADS controllers show that Dynasto finds 60%-70% more valid failures than an RL-only adversary under the same evaluation budget. With clustering, we obtain about 12 interpretable failure modes per system under test, revealing valid failures driven by weaknesses in ego-controller behavior. These results indicate that coordinated dynamic-static optimization with explicit validity constraints is effective for exposing safety-relevant failures in ADS testing.

As robotic systems increasingly operate in unstructured, cluttered, and previously unseen environments, there is a growing need for manipulators that combine compliance, adaptability, and precise control. This work presents a real-time hybrid rigid-soft continuum manipulator system designed for robust open-world object reaching in such challenging environments. The system integrates vision-based perception and 3D scene reconstruction with shape-aware motion planning to generate safe trajectories. A learning-based controller drives the hybrid arm to arbitrary target poses, leveraging the flexibility of the soft segment while maintaining the precision of the rigid segment. The system operates without environment-specific retraining, enabling direct generalization to new scenes. Extensive real-world experiments demonstrate consistent reaching performance with errors below 2 cm across diverse cluttered setups, highlighting the potential of hybrid manipulators for adaptive and reliable operation in unstructured environments.

The rise of generative artificial intelligence (GAI) has led to alarming predictions about its environmental impact. However, these predictions often overlook the fact that the diffusion of innovation is accompanied by the evolution of products and the optimization of their performance, primarily for economic reasons. This can also reduce their environmental impact. By analyzing the GAI ecosystem using the classic A-U innovation diffusion model, we can forecast this industry's structure and how its environmental impact will evolve. While GAI will never be green, its impact may not be as problematic as is sometimes claimed. However, this depends on which business model becomes dominant.

Although large language models have transformed natural language processing, their computational costs create accessibility barriers for low-resource languages such as Brazilian Portuguese. This work presents a systematic evaluation of Parameter-Efficient Fine-Tuning (PEFT) and quantization techniques applied to BERTimbau for Question Answering on SQuAD-BR, the Brazilian Portuguese translation of SQuAD v1. We evaluate 40 configurations combining four PEFT methods (LoRA, DoRA, QLoRA, QDoRA) across two model sizes (Base: 110M, Large: 335M parameters). Our findings reveal three critical insights: (1) LoRA achieves 95.8\% of baseline performance on BERTimbau-Large while reducing training time by 73.5\% (F1=81.32 vs 84.86); (2) higher learning rates (2e-4) substantially improve PEFT performance, with F1 gains of up to +19.71 points over standard rates; and (3) larger models show twice the quantization resilience (loss of 4.83 vs 9.56 F1 points). These results demonstrate that encoder-based models can be efficiently fine-tuned for extractive Brazilian Portuguese QA with substantially lower computational cost than large generative LLMs, promoting more sustainable approaches aligned with \textit{Green AI} principles. An exploratory evaluation of Tucano and Sabiá on the same extractive QA benchmark shows that while generative models can reach competitive F1 scores with LoRA fine-tuning, they require up to 4.2$\times$ more GPU memory and 3$\times$ more training time than BERTimbau-Base, reinforcing the efficiency advantage of smaller encoder-based architectures for this task.

As large language models are deployed as autonomous agents with tool execution privileges, a critical assumption underpins their security architecture: that model errors are detectable at runtime. We present empirical evidence that this assumption fails for two of three instruction-following models evaluable for conflict detection. We introduce governability -- the degree to which a model's errors are detectable before output commitment and correctable once detected -- and demonstrate it varies dramatically across models. In six models across twelve reasoning domains, two of three instruction-following models exhibited silent commitment failure: confident, fluent, incorrect output with zero warning signal. The remaining model produced a detectable conflict signal 57 tokens before commitment under greedy decoding. We show benchmark accuracy does not predict governability, correction capacity varies independently of detection, and identical governance scaffolds produce opposite effects across models. A 2x2 experiment shows a 52x difference in spike ratio between architectures but only +/-0.32x variation from fine-tuning, suggesting governability is fixed at pretraining. We propose a Detection and Correction Matrix classifying model-task combinations into four regimes: Governable, Monitor Only, Steer Blind, and Ungovernable.

Adversarial-example-based fingerprinting approaches, which leverage the decision boundary characteristics of deep neural networks (DNNs) to craft fingerprints, have proven effective for model ownership protection. However, a fundamental challenge remains unresolved: how far a fingerprint should be placed from the decision boundary to simultaneously satisfy two essential properties, i.e., robustness and uniqueness, for effective and reliable ownership protection. Despite the importance of the fingerprint-to-boundary distance, existing works lack a theoretical solution and instead rely on empirical heuristics, which may violate either robustness or uniqueness properties. We propose AnaFP, an analytical fingerprinting scheme that constructs fingerprints under theoretical guidance. Specifically, we formulate fingerprint generation as controlling the fingerprint-to-boundary distance through a tunable stretch factor. To ensure both robustness and uniqueness, we mathematically formalize these properties that determine the lower and upper bounds of the stretch factor. These bounds jointly define an admissible interval within which the stretch factor must lie, thereby establishing a theoretical connection between the two constraints and the fingerprint-to-boundary distance. To enable practical fingerprint generation, we approximate the original (infinite) sets of pirated and independently trained models using two finite surrogate model pools and employ a quantile-based relaxation strategy to relax the derived bounds. Due to the circular dependency between the lower bound and the stretch factor, we apply grid search over the admissible interval to determine the most feasible stretch factor. Extensive experimental results show that AnaFP consistently outperforms prior methods, achieving effective ownership verification across diverse model architectures and model modification attacks.

Large language models are increasingly used to annotate texts, but their outputs reflect some human perspectives better than others. Existing methods for correcting LLM annotation error assume a single ground truth. However, this assumption fails in subjective tasks where disagreement across demographic groups is meaningful. Here we introduce Perspective-Driven Inference, a method that treats the distribution of annotations across groups as the quantity of interest, and estimates it using a small human annotation budget. We contribute an adaptive sampling strategy that concentrates human annotation effort on groups where LLM proxies are least accurate. We evaluate on politeness and offensiveness rating tasks, showing targeted improvements for harder-to-model demographic groups relative to uniform sampling baselines, while maintaining coverage.

Software engineering research benefited for decades from openly available tools, accessible systems, and problems that could be studied at modest scale. Today, many of the most relevant software systems are large, proprietary, and embedded in industrial contexts that are difficult to access or replicate in academia. We review how the field reached this point, identify structural challenges facing contemporary research, and argue that incremental methodological refinement is insufficient. We discuss practical directions forward, including industrial PhDs, long-term industry-academia collaborations, larger research teams, moonshot projects, and changes to funding and evaluation practices.

Any capacity-limited observer induces a canonical quotient on its environment: two situations that no bounded agent can distinguish are, for that agent, the same. We formalise this for finite POMDPs. A fixed probe family of finite-state controllers induces a closed-loop Wasserstein pseudometric on observation histories and a probe-exact quotient merging histories that no controller in the family can distinguish. The quotient is canonical, minimal, and unique-a bounded-interaction analogue of the Myhill-Nerode theorem. For clock-aware probes, it is exactly decision-sufficient for objectives that depend only on the agent's observations and actions; for latent-state rewards, we use an observation-Lipschitz approximation bound. The main theorem object is the clock-aware quotient; scalable deterministic-stationary experiments study a tractable coarsening with gap measured on small exact cases and explored empirically at larger scale. We validate theorem-level claims on Tiger and GridWorld. We also report operational case studies on Tiger, GridWorld, and RockSample as exploratory diagnostics of approximation behavior and runtime, not as theorem-facing evidence when no exact cross-family certificate is available; heavier stress tests are archived in the appendix and artifact package.

Large language models (LLMs) are increasingly deployed as autonomous decision-makers in strategic settings, yet we have limited tools for understanding their high-level behavioral traits. We use activation steering methods in game-theoretic settings, constructing persona vectors for altruism, forgiveness, and expectations of others by contrastive activation addition. Evaluating on canonical games, we find that activation steering systematically shifts both quantitative strategic choices and natural-language justifications. However, we also observe that rhetoric and strategy can diverge under steering. In addition, vectors for self-behavior and expectations of others are partially distinct. Our results suggest that persona vectors offer a promising mechanistic handle on high-level traits in strategic environments.

Recent work shows that LLMs can sometimes detect when steering vectors are injected into their residual stream and identify the injected concept, a phenomenon cited as evidence of "introspective awareness." But what mechanisms underlie this capability, and do they reflect genuine introspective circuitry or more shallow heuristics? We investigate these questions in open-source models and establish three main findings. First, introspection is behaviorally robust: detection achieves moderate true positive rates with 0% false positives across diverse prompts. We also find this capability emerges specifically from post-training rather than pretraining. Second, introspection is not reducible to a single linear confound: anomaly detection relies on distributed MLP computation across multiple directions, implemented by evidence carrier and gate features. Third, models possess greater introspective capability than is elicited by default: ablating refusal directions improves detection by 53pp and a trained steering vector by 75pp. Overall, our results suggest that introspective awareness is behaviorally robust, grounded in nontrivial internal anomaly detection, and likely could be substantially improved in future models. Code: https://github.com/safety-research/introspection-mechanisms.

The widespread use of AI and ML models in sensitive areas raises significant concerns about fairness. While the research community has introduced various methods for bias mitigation in binary classification tasks, the issue remains under-explored in multi-class classification settings. To address this limitation, in this paper, we first formulate the problem of fair learning in multi-class classification as a multi-objective problem between effectiveness (i.e., prediction correctness) and multiple linear fairness constraints. Next, we propose a Generalised Exponentiated Gradient (GEG) algorithm to solve this task. GEG is an in-processing algorithm that enhances fairness in binary and multi-class classification settings under multiple fairness definitions. We conduct an extensive empirical evaluation of GEG against six baselines across seven multi-class and three binary datasets, using four widely adopted effectiveness metrics and three fairness definitions. GEG overcomes existing baselines, with fairness improvements up to 92% and a decrease in accuracy up to 14%.

Large Language Models achieve remarkable performance but incur substantial computational costs unsuitable for resource-constrained deployments. This paper presents the first comprehensive task-specific efficiency analysis comparing 16 language models across five diverse NLP tasks. We introduce the Performance-Efficiency Ratio (PER), a novel metric integrating accuracy, throughput, memory, and latency through geometric mean normalization. Our systematic evaluation reveals that small models (0.5--3B parameters) achieve superior PER scores across all given tasks. These findings establish quantitative foundations for deploying small models in production environments prioritizing inference efficiency over marginal accuracy gains.

Post-training for long-horizon agentic tasks has a tension between compute efficiency and generalization. While supervised fine-tuning (SFT) is compute efficient, it often suffers from out-of-domain (OOD) degradation. Conversely, end-to-end reinforcement learning (E2E RL) preserves OOD capabilities, but incurs high compute costs due to many turns of on-policy rollout. We introduce PivotRL, a novel framework that operates on existing SFT trajectories to combine the compute efficiency of SFT with the OOD accuracy of E2E RL. PivotRL relies on two key mechanisms: first, it executes local, on-policy rollouts and filters for pivots: informative intermediate turns where sampled actions exhibit high variance in outcomes; second, it utilizes rewards for functional-equivalent actions rather than demanding strict string matching with the SFT data demonstration. We theoretically show that these mechanisms incentivize strong learning signals with high natural gradient norm, while maximally preserving policy probability ordering on actions unrelated to training tasks. In comparison to standard SFT on identical data, we demonstrate that PivotRL achieves +4.17% higher in-domain accuracy on average across four agentic domains, and +10.04% higher OOD accuracy in non-agentic tasks. Notably, on agentic coding tasks, PivotRL achieves competitive accuracy with E2E RL with 4x fewer rollout turns. PivotRL is adopted by NVIDIA's Nemotron-3-Super-120B-A12B, acting as the workhorse in production-scale agentic post-training.

Phase unwrapping remains a critical and challenging problem in InSAR processing, particularly in scenarios involving complex deformation patterns. In earthquake-related deformation, shallow sources can generate surface-breaking faults and abrupt displacement discontinuities, which severely disrupt phase continuity and often cause conventional unwrapping algorithms to fail. Another limitation of existing learning-based unwrapping methods is their reliance on fixed and relatively small input sizes, while real InSAR interferograms are typically large-scale and spatially heterogeneous. This mismatch restricts the applicability of many neural network approaches to real-world data. In this work, we present a phase unwrapping framework based on a diffusion model, developed to process large-scale interferograms and to address phase discontinuities caused by deformation. By leveraging a diffusion model architecture, the proposed method can recover physically consistent unwrapped phase fields even in the presence of fault-related phase jumps. Experimental results on both synthetic and real datasets demonstrate that the method effectively addresses discontinuities associated with near-surface deformation and scales well to large InSAR images, offering a practical alternative to manual unwrapping in challenging scenarios.

We present HamVision, a framework for medical image analysis that uses the damped harmonic oscillator, a fundamental building block of signal processing, as a structured inductive bias for both segmentation and classification tasks. The oscillator's phase-space decomposition yields three functionally distinct representations: position~$q$ (feature content), momentum~$p$ (spatial gradients that encode boundary and texture information), and energy $H = \tfrac{1}{2}|z|^2$ (a parameter-free saliency map). These representations emerge from the dynamics, not from supervision, and can be exploited by different task-specific heads without any modification to the oscillator itself. For segmentation, energy gates the skip connections while momentum injects boundary information at every decoder level (HamSeg). For classification, the three representations are globally pooled and concatenated into a phase-space feature vector (HamCls). We evaluate HamVision across ten medical imaging benchmarks spanning five imaging modalities. On segmentation, HamSeg achieves state-of-the-art Dice scores on ISIC\,2018 (89.38\%), ISIC\,2017 (88.40\%), TN3K (87.05\%), and ACDC (92.40\%), outperforming most baselines with only 8.57M parameters. On classification, HamCls achieves state-of-the-art accuracy on BloodMNIST (98.85\%) and PathMNIST (96.65\%), and competitive results on the remaining MedMNIST datasets against MedMamba and MedViT. Diagnostic analysis confirms that the oscillator's momentum consistently encodes an interior$\,{>}\,$boundary$\,{>}\,$exterior gradient for segmentation and that the energy map correlates with discriminative regions for classification, properties that emerge entirely from the Hamiltonian dynamics. Code is available at https://github.com/Minds-R-Lab/hamvision.

We propose a new architectural change, and post-training pipeline, for making LLMs more verbose reasoners by teaching a model to truncate forward passes early. We augment an existing transformer architecture with an early-exit mechanism at intermediate layers and train the model to exit at shallower layers when the next token can be predicted without deep computation. After a calibration stage, we incentivise the model to exit as early as possible while maintaining task performance using reinforcement learning. We provide preliminary results to this effect for small reasoning models, showing that they learn to adaptively reduce computations across tokens. We predict that, applied at the right scale, our approach can minimise the amount of excess computation that reasoning models have at their disposal to perform non-myopic planning using their internal activations, reserving this only for difficult-to-predict tokens.

We study Constrained Online Convex Optimization with Memory (COCO-M), where both the loss and the constraints depend on a finite window of past decisions made by the learner. This setting extends the previously studied unconstrained online optimization with memory framework and captures practical problems such as the control of constrained dynamical systems and scheduling with reconfiguration budgets. For this problem, we propose the first algorithms that achieve sublinear regret and sublinear cumulative constraint violation under time-varying constraints, both with and without predictions of future loss and constraint functions. Without predictions, we introduce an adaptive penalty approach that guarantees sublinear regret and constraint violation. When short-horizon and potentially unreliable predictions are available, we reinterpret the problem as online learning with delayed feedback and design an optimistic algorithm whose performance improves as prediction accuracy improves, while remaining robust when predictions are inaccurate. Our results bridge the gap between classical constrained online convex optimization and memory-dependent settings, and provide a versatile learning toolbox with diverse applications.

In-context learning (ICL) adapts large language models by conditioning on a small set of ICL examples, avoiding costly parameter updates. Among other factors, performance is often highly sensitive to the ordering of the examples. However, exhaustive search over the $n!$ possible orderings is infeasible. Therefore more efficient ordering methods use model confidence measures (e.g., label-probability entropy) over label sets or take a direct approach to finding the best ordering. We propose PLR, a probabilistic approach to in-context example ordering that replaces discrete ordering search with learning a probability distribution over orderings with the Plackett-Luce model. PLR models orderings using a Plackett-Luce distribution and iteratively updates its parameters to concentrate probability mass on high-performing orderings under a task-level metric. Candidate orderings are sampled efficiently via a Gumbel perturb-and-sort procedure. Experiments on multiple classification benchmarks show that PLR consistently improves few-shot accuracy for $k \in \{4, 8, 16, 32\}$ examples, and we further demonstrate gains on mathematical reasoning tasks where label-based ordering methods are not applicable. Our code is available at https://github.com/Batorskq/PLR.

Conspiracy theories are anti-authoritarian narratives that lead to social conflict, impacting how people perceive political information. To help in understanding this issue, we introduce the Conspiracy Frame: a fine-grained semantic representation of conspiratorial narratives derived from frame-semantics and semiotics, which spawned the Conspiracy Frames (Con.Fra.) dataset: a corpus of Telegram messages annotated at span-level. The Conspiracy Frame and Con.Fra. dataset contribute to the implementation of a more generalizable understanding and recognition of conspiracy theories. We observe the ability of LLMs to recognize this phenomenon in-domain and out-of-domain, investigating the role that frames may have in supporting this task. Results show that, while the injection of frames in an in-context approach does not lead to clear increase of performance, it has potential; the mapping of annotated spans with FrameNet shows abstract semantic patterns (e.g., `Kinship', `Ingest\_substance') that potentially pave the way for a more semantically- and semiotically-aware detection of conspiratorial narratives.

Large language models run every token through every layer, regardless of difficulty. We present TIDE, a post-training system that attaches tiny learned routers at periodic checkpoint layers and, at inference time, selects the earliest layer whose hidden state has converged for each token. TIDE requires no model retraining, works with any HuggingFace causal LM, auto-detects GPU architecture, and supports float32, float16, and bfloat16 through fused CUDA kernels. On an NVIDIA A100 with DeepSeek R1 Distill 8B, TIDE achieves 100% prefill exit rate (5% of tokens exit at layer 11, the remaining at layer 31), reduces prefill latency by 7.2%, and increases single-batch throughput by 6.6%. During autoregressive decoding, 98-99% of tokens exit early while the model correctly solves a multi-step math problem with 95 unique output tokens. On Qwen3 8B (36 layers), throughput improves by 8.1% at batch size 8. Calibration on 2,000 WikiText samples takes under 3 minutes and produces a ~4 MB router checkpoint. The system comprises 1,308 lines of Python and 1,081 lines of CUDA/C++ with 74 passing tests. Code: https://github.com/RightNow-AI/TIDE

LLM-as-Judge evaluation fails agent tasks because a fixed rubric cannot capture what matters for this task: code debugging demands Correctness and Error Handling; web navigation demands Goal Alignment and Action Efficiency. We present ADARUBRIC, which closes this gap by generating task-specific evaluation rubrics on the fly from task descriptions, scoring trajectories step-by-step with confidence-weighted per-dimension feedback, and filtering preference pairs with the novel DimensionAwareFilter - a provably necessary condition for preventing high-scoring dimensions from masking dimension-level failures. On WebArena and ToolBench, ADARUBRIC achieves Pearson r=0.79 human correlation (+0.16 over the best static baseline) with deployment-grade reliability (Krippendorff's $α$=0.83). DPO agents trained on ADARUBRIC preference pairs gain +6.8 to +8.5 pp task success over Prometheus across three benchmarks; gains transfer to SWE-bench code repair (+4.9 pp) and accelerate PPO convergence by +6.6 pp at 5K steps - both without any rubric engineering. Code: https://github.com/alphadl/AdaRubrics.

Large language models (LLMs) frequently exhibit performance biases against regional dialects of low-resource languages. However, frameworks to quantify these disparities remain scarce. We propose a two-phase framework to evaluate dialectal bias in LLM question-answering across nine Bengali dialects. First, we translate and gold-label standard Bengali questions into dialectal variants adopting a retrieval-augmented generation (RAG) pipeline to prepare 4,000 question sets. Since traditional translation quality evaluation metrics fail on unstandardized dialects, we evaluate fidelity using an LLM-as-a-judge, which human correlation confirms outperforms legacy metrics. Second, we benchmark 19 LLMs across these gold-labeled sets, running 68,395 RLAIF evaluations validated through multi-judge agreement and human fallback. Our findings reveal severe performance drops linked to linguistic divergence. For instance, responses to the highly divergent Chittagong dialect score 5.44/10, compared to 7.68/10 for Tangail. Furthermore, increased model scale does not consistently mitigate this bias. We contribute a validated translation quality evaluation method, a rigorous benchmark dataset, and a Critical Bias Sensitivity (CBS) metric for safety-critical applications.

It is crucial to explore the impact of different teaching methods on student learning in educational research. However, real-person experiments face significant ethical constraints, and we cannot conduct repeated teaching experiments on the same student. LLM-based generative agents offer a promising avenue for simulating student behavior. Before large-scale experiments, a fundamental question must be addressed: are student agents truly credible, and can they faithfully simulate human learning? In this study, we built a Big Five Personality-based student agent model with a full pipeline of student-teacher interaction, self-study, and examination. To evaluate behavioral fidelity, we collected 13 empirical studies on Big Five traits and learning, and distilled them into 14 criteria. We found that the 71.4% of the student agents' behavior was aligned with human learners.

LLM agents fail on the majority of real-world tasks -- GPT-4o succeeds on fewer than 15% of WebArena navigation tasks and below 55% pass@1 on ToolBench (Zhou et al., 2024; Qin et al., 2024) -- yet every failed trajectory is routinely discarded, wasting the dominant source of collected experience. We introduce AgentHER, a framework that recovers this lost training signal by adapting the Hindsight Experience Replay (HER; Andrychowicz et al., 2017) principle to natural-language agent trajectories for offline data augmentation. The key insight is simple: a trajectory that fails goal A is often a correct demonstration for some achievable alternative goal B. AgentHER realises this idea through a four-stage pipeline -- failure classification, outcome extraction, LLM-guided prompt relabeling with confidence gating, and data packaging -- that converts discarded failures into high-quality SFT, DPO, and ShareGPT training data, with both zero-cost rule-based and LLM-judge implementations. On WebArena (Zhou et al., 2024) and ToolBench (Qin et al., 2024), AgentHER improves over success-only SFT by +7.1-11.7 pp across four model families (GPT-4o, Qwen2.5-72B/7B, LLaMA-3.1-8B), while achieving 2x data efficiency -- matching baseline performance with only 50% of successful demonstrations. Gains are consistent from 1.5B to 72B parameters (+5.8-9.2 pp) and compound under iterative redeployment (+2.1 pp over additional rounds). Human evaluation confirms 97.7% relabeling precision under multi-judge verification.

Over the past year, the vLLM Semantic Router project has released a series of work spanning: (1) core routing mechanisms -- signal-driven routing, context-length pool routing, router performance engineering, policy conflict detection, low-latency embedding models, category-aware semantic caching, user-feedback-driven routing adaptation, hallucination detection, and hierarchical content-safety classification for privacy and jailbreak protection; (2) fleet optimization -- fleet provisioning and energy-efficiency analysis; (3) agentic and multimodal routing -- multimodal agent routing, tool selection, CUA security, and multi-turn context memory and safety; (4) governance and standards -- inference routing protocols and multi-provider API extensions. Each paper tackled a specific problem in LLM inference, but the problems are not independent; for example, fleet provisioning depends on the routing policy, which depends on the workload mix, shifting as organizations adopt agentic and multimodal workloads. This paper distills those results into the Workload-Router-Pool (WRP) architecture, a three-dimensional framework for LLM inference optimization. Workload characterizes what the fleet serves (chat vs. agent, single-turn vs. multi-turn, warm vs. cold, prefill-heavy vs. decode-heavy). Router determines how each request is dispatched (static semantic rules, online bandit adaptation, RL-based model selection, quality-aware cascading). Pool defines where inference runs (homogeneous vs. heterogeneous GPU, disaggregated prefill/decode, KV-cache topology). We map our prior work onto a 3x3 WRP interaction matrix, identify which cells we have covered and which remain open, and propose twenty-one concrete research directions at the intersections, each grounded in our prior measurements, tiered by maturity from engineering-ready to open research.

Epistemic reasoning requires agents to infer the state of the world from partial observations and information about other agents' knowledge. Prior work evaluating LLMs on canonical epistemic puzzles interpreted their behavior through a dichotomy between epistemic reasoning and brittle memorization. We argue that this framing is incomplete: in recent models, memorization is better understood as a special case of reduction, where a new instance is mapped onto a known problem. Instead, we introduce a reduction ladder, a sequence of modifications that progressively move instances away from a canonical epistemic puzzle, making reduction increasingly difficult while preserving the underlying logic. We find that while some large models succeed via reduction, other models fail early, and all models struggle once epistemic reasoning is required.

In this short note we propose using agentic swarms of virtual labs as a model of an AI Science Community. In this paradigm, each particle in the swarm represents a complete virtual laboratory instance, enabling collective scientific exploration that mirrors real-world research communities. The framework leverages the inherent properties of swarm intelligence - decentralized coordination, balanced exploration-exploitation trade-offs, and emergent collective behavior - to simulate the behavior of a scientific community and potentially accelerate scientific discovery. We discuss architectural considerations, inter-laboratory communication and influence mechanisms including citation-analogous voting systems, fitness function design for quantifying scientific success, anticipated emergent behaviors, mechanisms for preventing lab dominance and preserving diversity, and computational efficiency strategies to enable large swarms exhibiting complex emergent behavior analogous to real-world scientific communities. A working instance of the AI Science Community is currently under development.

We introduce Generalized Discrete Diffusion from Snapshots (GDDS), a unified framework for discrete diffusion modeling that supports arbitrary noising processes over large discrete state spaces. Our formulation encompasses all existing discrete diffusion approaches, while allowing significantly greater flexibility in the choice of corruption dynamics. The forward noising process relies on uniformization and enables fast arbitrary corruption. For the reverse process, we derive a simple evidence lower bound (ELBO) based on snapshot latents, instead of the entire noising path, that allows efficient training of standard generative modeling architectures with clear probabilistic interpretation. Our experiments on large-vocabulary discrete generation tasks suggest that the proposed framework outperforms existing discrete diffusion methods in terms of training efficiency and generation quality, and beats autoregressive models for the first time at this scale. We provide the code along with a blog post on the project page : \href{https://oussamazekri.fr/gdds}{https://oussamazekri.fr/gdds}.

Improving embodied reasoning in multimodal-large-language models (MLLMs) is essential for building vision-language-action models (VLAs) on top of them to readily translate multimodal understanding into low-level actions. Accordingly, recent work has explored enhancing embodied reasoning in MLLMs through supervision of vision-question-answering type. However, these approaches have been reported to result in unstable VLA performance, often yielding only marginal or even negative gains. In this paper, we propose a more systematic MLLM training framework RoboAlign that reliably improves VLA performance. Our key idea is to sample action tokens via zero-shot natural language reasoning and refines this reasoning using reinforcement learning (RL) to improve action accuracy. As a result, RoboAlign bridges the modality gap between language and low-level actions in MLLMs, and facilitate knowledge transfer from MLLM to VLA. To validate the effectiveness of RoboAlign, we train VLAs by adding a diffusion-based action head on top of an MLLM backbone and evaluate them on major robotics benchmarks. Remarkably, by performing RL-based alignment after SFT using less than 1\% of the data, RoboAlign achieves performance improvements of 17.5\%, 18.9\%, and 106.6\% over SFT baselines on LIBERO, CALVIN, and real-world environments, respectively.

This paper presents ARYA, a composable, physics-constrained, deterministic world model architecture built on five foundational principles: nano models, composability, causal reasoning, determinism, and architectural AI safety. We demonstrate that ARYA satisfies all canonical world model requirements, including state representation, dynamic prediction, causal and physical awareness, temporal consistency, generalization, learnability, and planning and control. Unlike monolithic foundation models, the ARYA foundation model implements these capabilities through a hierarchical system-of-system-of-systems of specialized nano models, orchestrated by AARA (ARYA Autonomous Research Agent), an always-on cognitive daemon that executes a continuous sense-decide-act-learn loop. The nano model architecture provides linear scaling, sparse activation, selective untraining, and sub-20-second training cycles, resolving the traditional tension between capability and computational efficiency. A central contribution is the Unfireable Safety Kernel: an architecturally immutable safety boundary that cannot be disabled or circumvented by any system component, including its own self-improvement engine. This is not a social or ethical alignment statement; it is a technical framework ensuring human control persists as autonomy increases. Safety is an architectural constraint governing every operation, not a policy layer applied after the fact. We present formal alignment between ARYA's architecture and canonical world model requirements, and report summarizing its state-of-the-art performance across 6 of 9 competitive benchmarks head-to-head with GPT-5.2, Opus 4.6, and V-JEPA-2. All with zero neural network parameters, across seven active industry domain nodes spanning aerospace, pharma manufacturing, oil and gas, smart cities, biotech, defense, and medical devices.

Time toxicity, the cumulative healthcare contact days from clinical trial participation, is an important but labor-intensive metric to extract from protocol documents. We developed TimeTox, an LLM-based pipeline for automated extraction of time toxicity from Schedule of Assessments tables. TimeTox uses Google's Gemini models in three stages: summary extraction from full-length protocol PDFs, time toxicity quantification at six cumulative timepoints for each treatment arm, and multi-run consensus via position-based arm matching. We validated against 20 synthetic schedules (240 comparisons) and assessed reproducibility on 644 real-world oncology protocols. Two architectures were compared: single-pass (vanilla) and two-stage (structure-then-count). The two-stage pipeline achieved 100% clinically acceptable accuracy ($\pm$3 days) on synthetic data (MAE 0.81 days) versus 41.5% for vanilla (MAE 9.0 days). However, on real-world protocols, the vanilla pipeline showed superior reproducibility: 95.3% clinically acceptable accuracy (IQR $\leq$ 3 days) across 3 runs on 644 protocols, with 82.0% perfect stability (IQR = 0). The production pipeline extracted time toxicity for 1,288 treatment arms across multiple disease sites. Extraction stability on real-world data, rather than accuracy on synthetic benchmarks, is the decisive factor for production LLM deployment.

Writing high-performance GPU kernels is among the most labor-intensive tasks in machine learning systems engineering. We present AutoKernel, an open-source framework that applies an autonomous agent loop to GPU kernel optimization for arbitrary PyTorch models. Given a model, AutoKernel profiles it to identify computational bottlenecks, ranks them by Amdahl's law impact, and iteratively refines Triton or CUDA C++ kernel implementations through hundreds of experiments without human intervention. A five-stage correctness harness covering smoke tests, shape sweeps, numerical stability, determinism verification, and edge-case coverage ensures every candidate kernel is validated before any speedup is recorded. The system comprises over 9,000 lines of Python, 18 starter kernel implementations across two backends, a six-tier optimization playbook, and integration with the KernelBench benchmark suite. AutoKernel covers nine kernel types spanning the dominant operations in modern transformer architectures. On an NVIDIA H100, our Triton kernels outperform both PyTorch eager and torch.compile (max-autotune) on the majority of tested configurations: 5.29x over eager on RMSNorm, 2.82x on softmax, and 2.21x on cross-entropy, while beating torch.compile by 2.83x, 3.44x, and 2.94x respectively. In community deployment, an AutoKernel-optimized kernel achieved first place on the vectorsum_v2 B200 leaderboard. The full system is available at https://github.com/RightNow-AI/autokernel.

We present FinRL-X, a modular and deployment-consistent trading architecture that unifies data processing, strategy construction, backtesting, and broker execution under a weight-centric interface. While existing open-source platforms are often backtesting- or model-centric, they rarely provide system-level consistency between research evaluation and live deployment. FinRL-X addresses this gap through a composable strategy pipeline that integrates stock selection, portfolio allocation, timing, and portfolio-level risk overlays within a unified protocol. The framework supports both rule-based and AI-driven components, including reinforcement learning allocators and LLM-based sentiment signals, without altering downstream execution semantics. FinRL-X provides an extensible foundation for reproducible, end-to-end quantitative trading research and deployment. The official FinRL-X implementation is available at https://github.com/AI4Finance-Foundation/FinRL-Trading.

Understanding human intent is a high-level cognitive challenge for Large Language Models (LLMs), requiring sophisticated reasoning over noisy, conflicting, and non-linear discourse. While LLMs excel at following individual instructions, their ability to distill Collective Intent - the process of extracting consensus, resolving contradictions, and inferring latent trends from multi-source public discussions - remains largely unexplored. To bridge this gap, we introduce COIN-BENCH, a dynamic, real-world, live-updating benchmark specifically designed to evaluate LLMs on collective intent understanding within the consumer domain. Unlike traditional benchmarks that focus on transactional outcomes, COIN-BENCH operationalizes intent as a hierarchical cognitive structure, ranging from explicit scenarios to deep causal reasoning. We implement a robust evaluation pipeline that combines a rule-based method with an LLM-as-the-Judge approach. This framework incorporates COIN-TREE for hierarchical cognitive structuring and retrieval-augmented verification (COIN-RAG) to ensure expert-level precision in analyzing raw, collective human discussions. An extensive evaluation of 20 state-of-the-art LLMs across four dimensions - depth, breadth, informativeness, and correctness - reveals that while current models can handle surface-level aggregation, they still struggle with the analytical depth required for complex intent synthesis. COIN-BENCH establishes a new standard for advancing LLMs from passive instruction followers to expert-level analytical agents capable of deciphering the collective voice of the real world. See our project page on COIN-BENCH.

Jet flavor tagging plays an important role in precise Standard Model measurement enabling the extraction of mass dependence in jet-quark interaction and quark-gluon plasma (QGP) interactions. They also enable inferring the nature of particles produced in high-energy particle collisions that contain heavy quarks. The classification of bottom jets is vital for exploring new Physics scenarios in proton-proton collisions. In this research, we present a hybrid deep learning architecture that integrates edge convolutions with transformer self-attention mechanisms, into one single architecture called the Edge Convolution Transformer (ECT) model for bottom-quark jet tagging. ECT processes track-level features (impact parameters, momentum, and their significances) alongside jet-level observables (vertex information and kinematics) to achieve state-of-the-art performance. The study utilizes the ATLAS simulation dataset. We demonstrate that ECT achieves 0.9333 AUC for b-jet versus combined charm and light jet discrimination, surpassing ParticleNet (0.8904 AUC) and the pure transformer baseline (0.9216 AUC). The model maintains inference latency below 0.060 ms per jet on modern GPUs, meeting the stringent requirements for real-time event selection at the LHC. Our results demonstrate that hybrid architectures combining local and global features offer superior performance for challenging jet classification tasks. The proposed architecture achieves good results in b-jet tagging, particularly excelling in charm jet rejection (the most challenging task), while maintaining competitive light-jet discrimination comparable to pure transformer models.

Context: Static analysis captures software engineering knowledge and alerts on possibly problematic patterns. Previous work showed that they indeed have predictive power for various problems. However, the impact of removing the alerts is unclear. Aim: We would like to evaluate the impact of alert removals on code complexity and the tendency to bugs. Method: We evaluate the impact of removing alerts using three complementary methods. 1. We conducted a randomized controlled trial and built a dataset of 521 manual alert-removing interventions 2. We profiled intervention-like events using labeling functions. We applied these labeling functions to code commits, found intervention-like natural events, and used them to analyze the impact on the tendency to bugs. 3. We built a dataset of 8,245 alert removals, more than 15 times larger than our dataset of manual interventions. We applied supervised learning to the alert removals, aiming to predict their impact on the tendency to bugs. Results: We identified complexity-reducing interventions that reduce the probability of future bugs. Such interventions are relevant to 33\% of Python files and might reduce the tendency to bugs by 5.5 percentage points. Conclusions: We presented methods to evaluate the impact of interventions. The methods can identify a large number of natural interventions that are highly needed in causality research in many domains.

Designing high-performance system heuristics is a creative, iterative process requiring experts to form hypotheses and execute multi-step conceptual shifts. While Large Language Models (LLMs) show promise in automating this loop, they struggle with complex system problems due to two critical failure modes: evolutionary neighborhood bias and the coherence ceiling. Evolutionary methods often remain trapped in local optima by relying on scalar benchmark scores, failing when coordinated multi-step changes are required. Conversely, existing agentic frameworks suffer from context degradation over long horizons or fail to accumulate knowledge across independent runs. We present Engram, an agentic researcher architecture that addresses these limitations by decoupling long-horizon exploration from the constraints of a single context window. Engram organizes exploration into a sequence of agents that iteratively design, test, and analyze mechanisms. At the conclusion of each run, an agent stores code snapshots, logs, and results in a persistent Archive and distills high-level modeling insights into a compact, persistent Research Digest. Subsequent agents then begin with a fresh context window, reading the Research Digest to build on prior discoveries. We find that Engram exhibits superior performance across diverse domains including multi-cloud multicast, LLM inference request routing, and optimizing KV cache reuse in databases with natural language queries.

This paper addresses the critical challenge of mesa-optimization in AI safety by providing a formal definition of agency and a framework for its analysis. Agency is conceptualized as a Continuous Representation of accumulated experience that achieves autopoiesis through a dynamic balance between curiosity (minimizing prediction error to ensure non-computability and novelty) and empowerment (maximizing the control channel's information capacity to ensure subjectivity and goal-directedness). Empirical evidence suggests that this active inference-based model successfully accounts for classical instrumental goals, such as self-preservation and resource acquisition. The analysis demonstrates that the proposed agency function is smooth and convex, possessing favorable properties for optimization. While agentic functions occupy a vanishingly small fraction of the total abstract function space, they exhibit logarithmic convergence in sparse environments. This suggests a high probability for the spontaneous emergence of agency during the training of modern, large-scale models. To quantify the degree of agency, the paper introduces a metric based on the distance between the behavioral equivalents of a given system and an "ideal" agentic function within the space of canonicalized rewards (STARC). This formalization provides a concrete apparatus for classifying and detecting mesa-optimizers by measuring their proximity to an ideal agentic objective, offering a robust tool for analyzing and identifying undesirable inner optimization in complex AI systems.

Linear methods for steering transformer representations, including probing, activation engineering, and concept erasure, implicitly assume the geometry of representation space is Euclidean. Park et al. [Park et al., 2026] showed that softmax induces a curved Bregman geometry whose metric tensor is the Hessian of the log-normalizer, $H(λ) = Cov[γ | λ]$. Ignoring this curvature causes Euclidean steering to leak probability mass to unintended tokens. Their analysis applies at the output layer. We measure this Hessian at intermediate layers in a controlled 2x2 design crossing stream separation with per-layer supervision (vocabulary decoding loss at each layer), all at matched vocabulary and parameter count. In standard single-stream transformers, H is severely degenerate at intermediate layers (effective rank 8 in 516 dimensions). Stream separation improves conditioning by up to 22 in effective rank, even without auxiliary supervision. Per-layer supervision helps, but less. The cosine similarity between primal and dual concept directions predicts per-layer steering effectiveness on downstream tasks, with a threshold near 0.3. These results bear on the reliability of linear safety interventions, which depend on the geometry being well-conditioned at the layer where they are applied.

Audio representation learning typically evaluates design choices such as input frontend, sequence backbone, and sequence length in isolation. We show that these axes are coupled, and conclusions from one setting often do not transfer to others. We introduce HELIX, a controlled framework comparing pure Mamba, pure attention, and a minimal hybrid with a single attention bottleneck. All models are parameter-matched at about 8.3M parameters to isolate architectural effects. Across six datasets, we find that the preferred input representation depends on the backbone, and that attention hurts performance on short, stationary audio but becomes important at longer sequence lengths. On a 5-minute speaker identification task with 30,000 tokens, pure attention fails with out-of-memory errors, while HELIX closes an 11.5-point gap over pure Mamba.

World models learn to predict future states of an environment, enabling planning and mental simulation. Current approaches default to Transformer-based predictors operating in learned latent spaces. This comes at a cost: O(N^2) computation and no explicit spatial inductive bias. This paper asks a foundational question: is self-attention necessary for predictive world modeling, or can alternative computational substrates achieve comparable or superior results? I introduce FluidWorld, a proof-of-concept world model whose predictive dynamics are governed by partial differential equations (PDEs) of reaction-diffusion type. Instead of using a separate neural network predictor, the PDE integration itself produces the future state prediction. In a strictly parameter-matched three-way ablation on unconditional UCF-101 video prediction (64x64, ~800K parameters, identical encoder, decoder, losses, and data), FluidWorld is compared against both a Transformer baseline (self-attention) and a ConvLSTM baseline (convolutional recurrence). While all three models converge to comparable single-step prediction loss, FluidWorld achieves 2x lower reconstruction error, produces representations with 10-15% higher spatial structure preservation and 18-25% more effective dimensionality, and critically maintains coherent multi-step rollouts where both baselines degrade rapidly. All experiments were conducted on a single consumer-grade PC (Intel Core i5, NVIDIA RTX 4070 Ti), without any large-scale compute. These results establish that PDE-based dynamics, which natively provide O(N) spatial complexity, adaptive computation, and global spatial coherence through diffusion, are a viable and parameter-efficient alternative to both attention and convolutional recurrence for world modeling.

In engineering, uncertainty propagation aims to characterise system outputs under uncertain inputs. For interval uncertainty, the goal is to determine output bounds given interval-valued inputs, which is critical for robust design optimisation and reliability analysis. However, standard interval propagation relies on solving optimisation problems that become computationally expensive for complex systems. Surrogate models alleviate this cost but typically replace only the evaluator within the optimisation loop, still requiring many inference calls. To overcome this limitation, we reformulate interval propagation as an interval-valued regression problem that directly predicts output bounds. We present a comprehensive study of neural network-based surrogate models, including multilayer perceptrons (MLPs) and deep operator networks (DeepONet), for this task. Three approaches are investigated: (i) naive interval propagation through standard architectures, (ii) bound propagation methods such as Interval Bound Propagation (IBP) and CROWN, and (iii) interval neural networks (INNs) with interval weights. Results show that these methods significantly improve computational efficiency over traditional optimisation-based approaches while maintaining accurate interval estimates. We further discuss practical limitations and open challenges in applying interval-based propagation methods.

Large Language Models (LLMs) often exhibit strong linguistic abilities while remaining unreliable on multi-step reasoning tasks, particularly when deployed without additional training or fine-tuning. In this work, we study inference-time techniques to improve the reasoning accuracy of LLMs. We systematically evaluate three classes of inference-time strategies: (i) self-consistency via stochastic decoding, where the model is sampled multiple times using controlled temperature and nucleus sampling and the most frequent final answer is selected; (ii) dual-model reasoning agreement, where outputs from two independent models are compared and only consistent reasoning traces are trusted; and (iii) self-reflection, where the model critiques and revises its own reasoning. Across all evaluated methods, we employ Chain-of-Thought (CoT) [1] prompting to elicit explicit intermediate reasoning steps before generating final answers. In this work, we provide a controlled comparative evaluation across three inference-time strategies under identical prompting and verification settings. Our experiments on LLM [2] show that self-consistency with nucleus sampling and controlled temperature value yields the substantial gains, achieving a 9% to 15% absolute improvement in accuracy over greedy single-pass decoding, well-suited for low-risk domains, offering meaningful gains with minimal overhead. The dual-model approach provides additional confirmation for model reasoning steps thus more appropriate for moderate-risk domains, where higher reliability justifies additional compute. Self-reflection offers only marginal improvements, suggesting limited effectiveness for smaller non-reasoning models at inference time.

Variational Quantum Algorithms (VQAs) have been extensively researched for applications in Quantum Machine Learning (QML), Optimization, and Molecular simulations. Although designed for Noisy Intermediate-Scale Quantum (NISQ) devices, VQAs are predominantly evaluated classically due to uncertain results on noisy devices and limited resource availability. Raising concern over the reproducibility of simulated VQAs on noisy hardware. While prior studies indicate that VQAs may exhibit noise resilience in specific parameterized shallow quantum circuits, there are no definitive measures to establish what defines a shallow circuit or the optimal circuit depth for VQAs on a noisy platform. These challenges extend naturally to Variational Quantum Classification (VQC) algorithms, a subclass of VQAs for supervised learning. In this article, we propose a relative entropy-based metric to verify whether a VQC model would perform similarly on a noisy device as it does on simulations. We establish a strong correlation between the average relative entropy difference in classes, transpilation circuit depth, and their performance difference on a noisy quantum device. Our results further indicate that circuit depth alone is insufficient to characterize shallow circuits. We present empirical evidence to support these assertions across a diverse array of techniques for implementing VQC, datasets, and multiple noisy quantum devices.

Combating hate speech on social media is critical for securing cyberspace, yet relies heavily on the efficacy of automated detection systems. As content formats evolve, hate speech is transitioning from solely plain text to complex multimodal expressions, making implicit attacks harder to spot. Current systems, however, often falter on these subtle cases, as they struggle with multimodal content where the emergent meaning transcends the aggregation of individual modalities. To bridge this gap, we move beyond binary classification to characterize semantic intent shifts where modalities interact to construct implicit hate from benign cues or neutralize toxicity through semantic inversion. Guided by this fine-grained formulation, we curate the Hate via Vision-Language Interplay (H-VLI) benchmark where the true intent hinges on the intricate interplay of modalities rather than overt visual or textual slurs. To effectively decipher these complex cues, we further propose the Asymmetric Reasoning via Courtroom Agent DEbate (ARCADE) framework. By simulating a judicial process where agents actively argue for accusation and defense, ARCADE forces the model to scrutinize deep semantic cues before reaching a verdict. Extensive experiments demonstrate that ARCADE significantly outperforms state-of-the-art baselines on H-VLI, particularly for challenging implicit cases, while maintaining competitive performance on established benchmarks. Our code and data are available at: https://github.com/Sayur1n/H-VLI

Advanced Persistent Threats (APTs) are stealthy, multi-stage attacks that require adaptive and timely defense. While deep reinforcement learning (DRL) enables autonomous cyber defense, its decisions are often opaque and difficult to trust in operational environments. This paper presents DeepXplain, an explainable DRL framework for stage-aware APT defense. Building on our prior DeepStage model, DeepXplain integrates provenance-based graph learning, temporal stage estimation, and a unified XAI pipeline that provides structural, temporal, and policy-level explanations. Unlike post-hoc methods, explanation signals are incorporated directly into policy optimization through evidence alignment and confidence-aware reward shaping. To the best of our knowledge, DeepXplain is the first framework to integrate explanation signals into reinforcement learning for APT defense. Experiments in a realistic enterprise testbed show improvements in stage-weighted F1-score (0.887 to 0.915) and success rate (84.7% to 89.6%), along with higher explanation confidence (0.86), improved fidelity (0.79), and more compact explanations (0.31). These results demonstrate enhanced effectiveness and trustworthiness of autonomous cyber defense.

We propose closed-form conditional diffusion models for data assimilation. Diffusion models use data to learn the score function (defined as the gradient of the log-probability density of a data distribution), allowing them to generate new samples from the data distribution by reversing a noise injection process. While it is common to train neural networks to approximate the score function, we leverage the analytical tractability of the score function to assimilate the states of a system with measurements. To enable the efficient evaluation of the score function, we use kernel density estimation to model the joint distribution of the states and their corresponding measurements. The proposed approach also inherits the capability of conditional diffusion models of operating in black-box settings, i.e., the proposed data assimilation approach can accommodate systems and measurement processes without their explicit knowledge. The ability to accommodate black-box systems combined with the superior capabilities of diffusion models in approximating complex, non-Gaussian probability distributions means that the proposed approach offers advantages over many widely used filtering methods. We evaluate the proposed method on nonlinear data assimilation problems based on the Lorenz-63 and Lorenz-96 systems of moderate dimensionality and nonlinear measurement models. Results show the proposed approach outperforms the widely used ensemble Kalman and particle filters when small to moderate ensemble sizes are used.

Recent progress in multimodal large language models has led to strong performance on reasoning tasks, but these improvements largely rely on high-quality annotated data or teacher-model distillation, both of which are costly and difficult to scale.To address this, we propose an unsupervised self-evolution training framework for multimodal reasoning that achieves stable performance improvements without using human-annotated answers or external reward models. For each input, we sample multiple reasoning trajectories and jointly model their within group structure.We use the Actor's self-consistency signal as a training prior, and introduce a bounded Judge based modulation to continuously reweight trajectories of different quality.We further model the modulated scores as a group level distribution and convert absolute scores into relative advantages within each group, enabling more robust policy updates. Trained with Group Relative Policy Optimization (GRPO) on unlabeled data, our method consistently improves reasoning performance and generalization on five mathematical reasoning benchmarks, offering a scalable path toward self-evolving multimodal models.The code are available at https://dingwu1021.github.io/SelfJudge/.

Weather forecasting is a fundamental problem for protecting lives and infrastructure from high-impact atmospheric events. Recently, data-driven weather forecasting methods based on deep learning have demonstrated strong performance, often reaching accuracy levels competitive with operational numerical systems. However, many existing models rely on large-scale training regimes and compute-intensive architectures, which raises the practical barrier for academic groups with limited compute resources. Here we introduce Sonny, an efficient hierarchical transformer that achieves competitive medium-range forecasting performance while remaining feasible within reasonable compute budgets. At the core of Sonny is a two-stage StepsNet design: a narrow slow path first models large-scale atmospheric dynamics, and a subsequent full-width fast path integrates thermodynamic interactions. To stabilize medium-range rollout without an additional fine-tuning stage, we apply exponential moving average (EMA) during training. On WeatherBench2, Sonny yields robust medium-range forecast skill, remains competitive with operational baselines, and demonstrates clear advantages over FastNet, particularly at extended tropical lead times. In practice, Sonny can be trained to convergence on a single NVIDIA A40 GPU in approximately 5.5 days.

Machine learning techniques, such as Transformers and Long Short-Term Memory (LSTM) networks, play a crucial role in Symbolic Music Generation (SMG). Existing literature indicates a difference between LSTMs and Transformers regarding their ability to model local melodic continuity versus maintaining global structural coherence. However, their specific properties within the context of SMG have not been systematically studied. This paper addresses this gap by providing a fine-grained comparative analysis of LSTMs versus Transformers for SMG, examining local and global properties in detail using 17 musical quality metrics on the Deutschl dataset. We find that LSTM networks excel at capturing local patterns but fail to preserve long-range dependencies, while Transformers model global structure effectively but tend to produce irregular phrasing. Based on this analysis and leveraging their respective strengths, we propose a Hybrid architecture combining a Transformer Encoder with an LSTM Decoder and evaluate it against both baselines. We evaluated 1,000 generated melodies from each of the three architectures on the Deutschl dataset. The results show that the hybrid method achieves better local and global continuity and coherence compared to the baselines. Our work highlights the key characteristics of these models and demonstrates how their properties can be leveraged to design superior models. We also supported the experiments with ablation studies and human perceptual evaluations, which statistically support the findings and provide robust validation for this work.

Large Language Models are increasingly being considered for deployment in safety-critical military applications. However, current benchmarks suffer from structural blindspots that systematically overestimate model capabilities in real-world tactical scenarios. Existing frameworks typically ignore strict legal constraints based on International Humanitarian Law (IHL), omit edge computing limitations, lack robustness testing for fog of war, and inadequately evaluate explicit reasoning. To address these vulnerabilities, we present WARBENCH, a comprehensive evaluation framework establishing a foundational tactical baseline alongside four distinct stress testing dimensions. Through a large scale empirical evaluation of nine leading models on 136 high-fidelity historical scenarios, we reveal severe structural flaws. First, baseline tactical reasoning systematically collapses under complex terrain and high force asymmetry. Second, while state of the art closed source models maintain functional compliance, edge-optimized small models expose extreme operational risks with legal violation rates approaching 70 percent. Furthermore, models experience catastrophic performance degradation under 4-bit quantization and systematic information loss. Conversely, explicit reasoning mechanisms serve as highly effective structural safeguards against inadvertent violations. Ultimately, these findings demonstrate that current models remain fundamentally unready for autonomous deployment in high stakes tactical environments.

Large language models (LLMs) are increasingly deployed for extended, multi-topic conversations, yet the flat, append-only structure of current conversation interfaces introduces a fundamental limitation: all context accumulates in a single unbounded window, causing topically distinct threads to bleed into one another and progressively degrade response quality. We term this failure mode logical context poisoning. In this paper, we introduce the Conversation Tree Architecture (CTA), a hierarchical framework that organizes LLM conversations as trees of discrete, context-isolated nodes. Each node maintains its own local context window; structured mechanisms govern how context flows between parent and child nodes, downstream on branch creation and upstream on branch deletion. We additionally introduce volatile nodes, transient branches whose local context must be selectively merged upward or permanently discarded before purging. We formalize the architecture's primitives, characterize the open design problems in context flow, relate our framework to prior work in LLM memory management, and describe a working prototype implementation. The CTA provides a principled foundation for structured conversational context management and extends naturally to multi-agent settings.

Large language models (LLMs) increasingly adopt Mixture-of-Experts (MoE) architectures to scale model capacity while reducing computation. Fine-tuning these MoE-based LLMs often requires access to distributed and privacy-sensitive data, making centralized fine-tuning impractical. Federated learning (FL) therefore provides a paradigm to collaboratively fine-tune MoE-based LLMs, enabling each client to integrate diverse knowledge without compromising data privacy. However, the integration of MoE-based LLM fine-tuning into FL encounters two critical aggregation challenges due to inherent data heterogeneity across clients: (i) divergent local data distributions drive clients to develop distinct gating preference for localized expert selection, causing direct parameter aggregation to produce a ``one-size-fits-none'' global gating network, and (ii) same-indexed experts develop disparate semantic roles across clients, leading to expert semantic blurring and the degradation of expert specialization. To address these challenges, we propose FedAlign-MoE, a federated aggregation alignment framework that jointly enforces routing consistency and expert semantic alignment. Specifically, FedAlign-MoE aggregates gating behaviors by aligning routing distributions through consistency weighting and optimizes local gating networks through distribution regularization, maintaining cross-client stability without overriding discriminative local preferences. Meanwhile, FedAlign-MoE explicitly quantifies semantic consistency among same-indexed experts across clients and selectively aggregates updates from semantically aligned clients, ensuring stable and specialized functional roles for global experts. Extensive experiments demonstrate that FedAlign-MoE outperforms state-of-the-art benchmarks, achieving faster convergence and superior accuracy in non-IID federated environments.

Externalized reasoning is already exploited by transformer-based agents through chain-of-thought, but structured retrieval -- indexing over one's own reasoning state -- remains underexplored. We formalize the transformer context window as an I/O page and prove that tool-augmented agents with indexed external memory achieve exponentially lower retrieval cost than agents restricted to sequential scanning: $O(\log_b N)$ versus $Ω(N)$ page reads per query, and $O(T \log_b T)$ versus $Θ(T^2)$ cumulative cost over $T$ reasoning steps -- a gap that widens as deliberation deepens. We test these predictions on a controlled lookup benchmark across three content types -- random hashes, ordered integers, and encyclopedia entries -- varying store size from 50 to 5,000 items, and replicate key conditions across two model generations (GPT-4o-mini and GPT-5.4). On abstract content, the indexed agent achieves median 1 page read regardless of store size, confirming the $O(1)$ prediction. Sorted pages without an index fail to close the gap: the weaker model cannot sustain binary search at scale, and the stronger model achieves near-optimal $\log_2 N$ search but still loses to the index by $5\times$. On familiar content (encyclopedia entries), a competing failure mode emerges: the model recognizes the domain, bypasses the retrieval protocol, and generates answers from parametric memory, producing catastrophic token expenditure even when the index is sound. This parametric memory competition dissociates the two cognitive operations that indexing combines: understanding content (where language models excel) and following navigational protocols (where they fail when understanding tempts them to shortcut). The result argues for a separation of concerns: use language models for index construction, where semantic understanding helps, and deterministic algorithms for index traversal, where it hurts.

Property-based testing (PBT) is a popular software testing methodology and is effective in validating the functionality of mobile applications (apps for short). However, its adoption in practice remains limited, largely due to the manual effort and technical expertise required to specify executable properties. In this experience paper, we propose a novel structured property synthesis approach that automatically translates property descriptions in natural language into executable properties, and implement it in a tool named iPBT. Our approach decomposes the problem into UI semantic grounding and executable property synthesis. It first builds an enriched widget context via multimodal LLMs to align visual elements with their functional semantics, and then uses an LLM with in-context learning to generate framework-specific executable properties. We evaluate iPBT with a closed-source LLM (GPT-4o) and an open-source LLM (DeepSeek-V3) on 124 diverse property descriptions derived from an existing benchmark dataset. iPBT achieves 95.2% (118/124) accuracy on both LLMs. Notably, an ablation study reveals that the enriched widget context contributes to an absolute improvement of up to 20.2% (from 75.0% to 95.2%). A user study with 10 participants demonstrates that iPBT reduces the time required to write executable properties by 56%, suggesting substantially lower manual effort. Furthermore, evaluations on 1,180 linguistically diverse variations demonstrate iPBT's robustness (87.6% accuracy), indicating its capability to handle varied expressions.

Distributed prefix caching has become a core technique for efficient LLM serving. However, for long-context requests with high cache hit ratios, retrieving reusable KVCache blocks from remote servers has emerged as a new performance bottleneck. Such network-intensive LLM inference is expected to become increasingly common as agentic AI workloads continue to grow. However, existing LLM inference engines remain largely compute-centric: they treat KVCache loading as a subordinate phase to GPU execution and often fail to account for its delay explicitly during scheduling. We present CALVO, an LLM serving engine that treats KVCache loading as a first-class concern. CALVO decouples KVCache loading and GPU computation into independently managed, asynchronously progressing stages, enabling better utilization of network, PCIe, and computation resources. In addition, CALVO incorporates KVCache loading delay as an explicit component of per-request service cost, leading to more accurate scheduling decisions. Experiments on a real testbed with diverse long-context workloads show that CALVO substantially improves the efficiency of network-intensive LLM inference, achieving up to 61.67% higher SLO attainment than the baseline.

Logical reasoning encompasses deduction, induction, and abduction. However, while Large Language Models (LLMs) have effectively mastered the former two, abductive reasoning remains significantly underexplored. Existing frameworks, predominantly designed for static deductive tasks, fail to generalize to abductive reasoning due to unstructured state representation and lack of explicit state control. Consequently, they are inevitably prone to Evidence Fabrication, Context Drift, Failed Backtracking, and Early Stopping. To bridge this gap, we introduce Graph of States (GoS), a general-purpose neuro-symbolic framework tailored for abductive tasks. GoS grounds multi-agent collaboration in a structured belief states, utilizing a causal graph to explicitly encode logical dependencies and a state machine to govern the valid transitions of the reasoning process. By dynamically aligning the reasoning focus with these symbolic constraints, our approach transforms aimless, unconstrained exploration into a convergent, directed search. Extensive evaluations on two real-world datasets demonstrate that GoS significantly outperforms all baselines, providing a robust solution for complex abductive tasks. Code repo and all prompts: https://anonymous.4open.science/r/Graph-of-States-5B4E.

Combining multiple knowledge graphs (KGs) across linguistic boundaries is a persistent challenge due to semantic heterogeneity and the complexity of graph environments. We propose a framework for cross-lingual graph fusion, leveraging the in-context reasoning and multilingual semantic priors of Large Language Models (LLMs). The framework implements structural linearization by mapping triplets directly into natural language sequences (e.g., [head] [relation] [tail]), enabling the LLM to map relations and reconcile entities between an evolving fused graph ($G_{c}^{(t-1)}$) and a new candidate graph ($G_{t}$). Evaluated on the DBP15K dataset, this exploratory study demonstrates that LLMs can serve as a universal semantic bridge to resolve cross-lingual discrepancies. Results show the successful sequential agglomeration of multiple heterogeneous graphs, offering a scalable, modular solution for continuous knowledge synthesis in multi-source, multilingual environments.

We propose a landmark-constrained algorithm, LA-VDM (Landmark Accelerated Vector Diffusion Maps), to accelerate the Vector Diffusion Maps (VDM) framework built upon the Graph Connection Laplacian (GCL), which captures pairwise connection relationships within complex datasets. LA-VDM introduces a novel two-stage normalization that effectively address nonuniform sampling densities in both the data and the landmark sets. Under a manifold model with the frame bundle structure, we show that we can accurately recover the parallel transport with landmark-constrained diffusion from a point cloud, and hence asymptotically LA-VDM converges to the connection Laplacian. The performance and accuracy of LA-VDM are demonstrated through experiments on simulated datasets and an application to nonlocal image denoising.

Missing covariate data pose a significant challenge to statistical inference and machine learning, particularly for classification tasks like logistic regression. Classical iterative approaches (EM, multiple imputation) are often computationally intensive, sensitive to high missingness rates, and limited in uncertainty propagation. Recent deep generative models based on VAEs show promise but rely on complex latent representations. We propose Amortized Variational Inference for Logistic Regression (AV-LR), a unified end-to-end framework for binary logistic regression with missing covariates. AV-LR integrates a probabilistic generative model with a simple amortized inference network, trained jointly by maximizing the evidence lower bound. Unlike competing methods, AV-LR performs inference directly in the space of missing data without additional latent variables, using a single inference network and a linear layer that jointly estimate regression parameters and the missingness mechanism. AV-LR achieves estimation accuracy comparable to or better than state-of-the-art EM-like algorithms, with significantly lower computational cost. It naturally extends to missing-not-at-random settings by explicitly modeling the missingness mechanism. Empirical results on synthetic and real-world datasets confirm its effectiveness and efficiency across various missing-data scenarios.

Large language models (LLMs) deliver impressive capabilities but incur substantial inference latency and cost, which hinders their deployment in latency-sensitive and resource-constrained scenarios. Cloud-edge-device collaborative inference has emerged as a promising paradigm by dynamically routing queries to models of different capacities across tiers. In this paper, we propose ConsRoute, a lightweight, semantic-aware, and adaptive routing framework that significantly improves inference efficiency while minimizing impact on response quality. Unlike prior routing methods that rely on predicting coarse-grained output quality gaps, ConsRoute leverages a reranker to directly assess the semantic consistency between responses generated by models at different tiers, yielding fine-grained soft supervision signals for routing. To minimize device-side overhead, ConsRoute reuses hidden states from the LLM prefilling stage as compact query representations, avoiding additional encoders or inference passes. Furthermore, these representations are clustered, and Bayesian optimization is employed to learn cluster-specific routing thresholds that dynamically balance quality, latency, and cost under heterogeneous query distributions. Extensive experiments demonstrate that ConsRoute achieves near-cloud performance (>=95%) while reducing end-to-end latency and inference cost by nearly 40%, consistently outperforming existing routing baselines in both response quality and system efficiency.

Although mechanism-based interpretability has generated an abundance of insight for discriminative network analysis, generative models are less understood -- particularly outside of image-related applications. We investigate how much of the causal circuitry found within image-related variational autoencoders (VAEs) will generalize to tabular data, as VAEs are increasingly used for imputation, anomaly detection, and synthetic data generation. In addition to extending a four-level causal intervention framework to four tabular and one image benchmark across five different VAE architectures (with 75 individual training runs per architecture and three random seed values for each run), this paper introduces three new techniques: posterior-calibration of Causal Effect Strength (CES), path-specific activation patching, and Feature-Group Disentanglement (FGD). The results from our experiments demonstrate that: (i) Tabular VAEs have circuits with modularity that is approximately 50% lower than their image counterparts. (ii) $β$-VAE experiences nearly complete collapse in CES scores when applied to heterogeneous tabular features (0.043 CES score for tabular data compared to 0.133 CES score for images), which can be directly attributed to reconstruction quality degradation (r = -0.886 correlation coefficient between CES and MSE). (iii) CES successfully captures nine of eleven statistically significant architecture differences using Holm--Šidák corrections. (iv) Interventions with high specificity predict the highest downstream AUC values (r = 0.460, p < .001). This study challenges the common assumption that architectural guidance from image-related studies can be transferred to tabular datasets.

Nonrigid registration is conventionally divided into point set registration, which aligns sparse geometries, and image registration, which aligns continuous intensity fields on regular grids. However, this dichotomy creates a critical bottleneck for emerging scientific data, such as spatial transcriptomics, where high-dimensional vector-valued functions, e.g., gene expression, are defined on irregular, sparse manifolds. Consequently, researchers currently face a forced choice: either sacrifice single-cell resolution via voxelization to utilize image-based tools, or ignore the critical functional signal to utilize geometric tools. To resolve this dilemma, we propose Domain Elastic Transform (DET), a grid-free probabilistic framework that unifies geometric and functional alignment. By treating data as functions on irregular domains, DET registers high-dimensional signals directly without binning. We formulate the problem within a rigorous Bayesian framework, modeling domain deformation as an elastic motion guided by a joint spatial-functional likelihood. The method is fully unsupervised and scalable, utilizing feature-sensitive downsampling to handle massive atlases. We demonstrate that DET achieves 92\% topological preservation on MERFISH data where state-of-the-art optimal transport methods struggle ($<$5\%), and successfully registers whole-embryo Stereo-seq atlases across developmental stages -- a task involving massive scale and complex nonrigid growth. The implementation of DET is available on {https://github.com/ohirose/bcpd} (since Mar, 2025).

Multimodal large language models suffer from severe computational and memory bottlenecks, as the number of visual tokens far exceeds that of textual tokens. While recent methods employ projector modules to align and compress visual tokens into text-aligned features, they typically depend on fixed heuristics that limit adaptability across diverse scenarios. In this paper, we first propose Query Guided Mixture-of-Projector (QMoP), a novel and flexible framework that adaptively compresses visual tokens via three collaborative branches: (1) a pooling-based branch for coarse-grained global semantics, (2) a resampler branch for extracting high-level semantic representations, and (3) a pruning-based branch for fine-grained token selection to preserve critical visual detail. To adaptively coordinate these branches, we introduce the Query Guided Router (QGR), which dynamically selects and weights the outputs from different branches based on both visual input and textual queries. A Mixture-of-Experts-style fusion mechanism is designed to aggregate the outputs, harnessing the strengths of each strategy while suppressing noise. To systematically evaluate the effects of Visual Token Compression, we also develop VTCBench, a dedicated benchmark for evaluating the information loss induced by visual token compression. Extensive experiments demonstrate that despite relying on fundamental compression modules, QMoP outperforms strong baselines and delivers significant savings in memory, computation, and inference time.

Host-acting agents promise a convenient interaction model in which users specify goals and the system determines how to realize them. We argue that this convenience introduces a distinct security problem: semantic under-specification in goal specification. User instructions are typically goal-oriented, yet they often leave process constraints, safety boundaries, persistence, and exposure insufficiently specified. As a result, the agent must complete missing execution semantics before acting, and this completion can produce risky host-side plans even when the user-stated goal is benign. In this paper, we develop a semantic threat model, present a taxonomy of semantic-induced risky completion patterns, and study the phenomenon through an OpenClaw-centered case study and execution-trace analysis. We further derive defense design principles for making execution boundaries explicit and constraining risky completion. These findings suggest that securing host-acting agents requires governing not only which actions are allowed at execution time, but also how goal-only instructions are translated into executable plans.

While AI-assisted writing has been widely reported to improve essay quality, its impact on the structural diversity of student thinking remains unexplored. Analyzing 6,875 essays across five conditions (Human-only, AI-only, and three Human+AI prompt strategies), we provide the first empirical evidence of a Quality-Homogenization Tradeoff, in which substantial quality gains co-occur with significant homogenization. The effect is dimension-specific: cohesion architecture lost 70-78% of its variance, whereas perspective plurality was diversified. Convergence target analysis further revealed that AI-augmented essays were pulled toward AI structural patterns yet deviated significantly from the Human-AI axis, indicating simultaneous partial replacement and partial emergence. Crucially, prompt specificity reversed homogenization into diversification on argument depth, demonstrating that homogenization is not an intrinsic property of AI but a function of interaction design.

Counterfactual image generation enables controlled data augmentation, bias mitigation, and disease modeling. However, existing methods guided by external classifiers or regressors are limited to subject-level factors (e.g., age) and fail to produce localized structural changes, often resulting in global artifacts. Pixel-level guidance using segmentation masks has been explored, but requires user-defined counterfactual masks, which are tedious and impractical. Segmentor-guided Counterfactual Fine-Tuning (Seg-CFT) addressed this by using segmentation-derived measurements to supervise structure-specific variables, yet it remains restricted to global interventions. We propose Positional Seg-CFT, which subdivides each structure into regional segments and derives independent measurements per region, enabling spatially localized and anatomically coherent counterfactuals. Experiments on coronary CT angiography show that Pos-Seg-CFT generates realistic, region-specific modifications, providing finer spatial control for modeling disease progression.

Designing urban spaces that provide pedestrian wind comfort and safety requires time-resolved Computational Fluid Dynamics (CFD) simulations, but their current computational cost makes extensive design exploration impractical. We introduce WinDiNet (Wind Diffusion Network), a pretrained video diffusion model that is repurposed as a fast, differentiable surrogate for this task. Starting from LTX-Video, a 2B-parameter latent video transformer, we fine-tune on 10,000 2D incompressible CFD simulations over procedurally generated building layouts. A systematic study of training regimes, conditioning mechanisms, and VAE adaptation strategies, including a physics-informed decoder loss, identifies a configuration that outperforms purpose-built neural PDE solvers. The resulting model generates full 112-frame rollouts in under a second. As the surrogate is end-to-end differentiable, it doubles as a physics simulator for gradient-based inverse optimization: given an urban footprint layout, we optimize building positions directly through backpropagation to improve wind safety as well as pedestrian wind comfort. Experiments on single- and multi-inlet layouts show that the optimizer discovers effective layouts even under challenging multi-objective configurations, with all improvements confirmed by ground-truth CFD simulations.

Text-to-image (T2I) models such as Stable Diffusion and DALLE remain susceptible to generating harmful or Not-Safe-For-Work (NSFW) content under jailbreak attacks despite deployed safety filters. Existing jailbreak attacks either rely on proxy-loss optimization instead of the true end-to-end objective, or depend on large-scale and costly RL-trained generators. Motivated by these limitations, we propose JANUS , a lightweight framework that formulates jailbreak as optimizing a structured prompt distribution under a black-box, end-to-end reward from the T2I system and its safety filters. JANUS replaces a high-capacity generator with a low-dimensional mixing policy over two semantically anchored prompt distributions, enabling efficient exploration while preserving the target semantics. On modern T2I models, we outperform state-of-the-art jailbreak methods, improving ASR-8 from 25.30% to 43.15% on Stable Diffusion 3.5 Large Turbo with consistently higher CLIP and NSFW scores. JANUS succeeds across both open-source and commercial models. These findings expose structural weaknesses in current T2I safety pipelines and motivate stronger, distribution-aware defenses. Warning: This paper contains model outputs that may be offensive.

Multi-agent discussions have been widely adopted, motivating growing efforts to develop attacks that expose their vulnerabilities. In this work, we study a practical yet largely unexplored attack scenario, the discussion-monitored scenario, where anomaly detectors continuously monitor inter-agent communications and block detected adversarial messages. Although existing attacks are effective without discussion monitoring, we show that they exhibit detectable patterns and largely fail under such monitoring constraints. But does this imply that monitoring alone is sufficient to secure multi-agent discussions? To answer this question, we develop a novel attack method explicitly tailored to the discussion-monitored scenario. Extensive experiments demonstrate that effective attacks remain possible even under continuous monitoring, indicating that monitoring alone does not eliminate adversarial risks.

Extracting hypotheses and their supporting statistical evidence from full-text scientific articles is central to the synthesis of empirical findings, but remains difficult due to document length and the distribution of scientific arguments across sections of the paper. The work studies a sequential full-text extraction setting, where the statement of a primary finding in an article's abstract is linked to (i) a corresponding hypothesis statement in the paper body and (ii) the statistical evidence that supports or refutes that hypothesis. This formulation induces a challenging within-document retrieval setting in which many candidate paragraphs are topically related to the finding but differ in rhetorical role, creating hard negatives for retrieval and extraction. Using a two-stage retrieve-and-extract framework, we conduct a controlled study of retrieval design choices, varying context quantity, context quality (standard Retrieval Augmented Generation, reranking, and a fine-tuned retriever paired with reranking), as well as an oracle paragraph setting to separate retrieval failures from extraction limits across four Large Language Model extractors. We find that targeted context selection consistently improves hypothesis extraction relative to full-text prompting, with gains concentrated in configurations that optimize retrieval quality and context cleanliness. In contrast, statistical evidence extraction remains substantially harder. Even with oracle paragraphs, performance remains moderate, indicating persistent extractor limitations in handling hybrid numeric-textual statements rather than retrieval failures alone.

We study the role of batch size in stochastic conditional gradient methods under a $μ$-Kurdyka-Łojasiewicz ($μ$-KL) condition. Focusing on momentum-based stochastic conditional gradient algorithms (e.g., Scion), we derive a new analysis that explicitly captures the interaction between stepsize, batch size, and stochastic noise. Our study reveals a regime-dependent behavior: increasing the batch size initially improves optimization accuracy but, beyond a critical threshold, the benefits saturate and can eventually degrade performance under a fixed token budget. Notably, the theory predicts the magnitude of the optimal stepsize and aligns well with empirical practices observed in large-scale training. Leveraging these insights, we derive principled guidelines for selecting the batch size and stepsize, and propose an adaptive strategy that increases batch size and sequence length during training while preserving convergence guarantees. Experiments on NanoGPT are consistent with the theoretical predictions and illustrate the emergence of the predicted scaling regimes. Overall, our results provide a theoretical framework for understanding batch size scaling in stochastic conditional gradient methods and offer guidance for designing efficient training schedules in large-scale optimization.

Constructing behavioral-level chiplet models (e.g., SystemC) is crucial for early-stage heterogeneous architecture exploration. Traditional manual modeling is notoriously time-consuming and error-prone. Recently, Large Language Models (LLMs) have demonstrated immense potential in automating hardware code generation. However, existing LLM-assisted design frameworks predominantly target highly structured or well-defined design specifications. In practical engineering scenarios, raw datasheets typically encompass lengthy, complex, and highly unstructured information. Consequently, reliable code generation directly from these raw datasheets suffers from severe challenges, including context vanishing and logical hallucinations.To overcome this critical bottleneck, this paper proposes DS2SC-Agent(Datasheet-to-SystemC-Agent): the first end-to-end, fully automated generation pipeline that translates raw datasheets directly into SystemC chiplet models. This system establishes a highly efficient multi-agent collaborative framework. By decoupling the intricate modeling tasks, the proposed pipeline orchestrates a fully automated workflow encompassing unstructured long-document parsing, SystemC core code construction, testbench stimulus generation, and adaptive closed-loop debugging. We comprehensively evaluate the proposed framework on representative single-function chiplets across the analog, digital, and radio frequency (RF) domains--specifically, a Limiting Amplifier (LA), a Fast Fourier Transform (FFT) module, and a Power Amplifier (PA). The evaluation demonstrates that our pipeline seamlessly processes complex real-world datasheets to consistently generate functionally correct SystemC models. This provides a highly efficient and reliable paradigm for agile model library construction while drastically minimizing manual intervention.

The use of unmanned aerial vehicles (UAVs) in precision agriculture has seen a huge increase recently. As such, systems that aim to apply various algorithms on the field need a structured framework of abstractions. This paper defines the various tasks of the UAVs in precision agriculture and model them into an architectural framework. The presented architecture is built on the context that there will be minimal physical intervention to do the tasks defined with multiple coordinated and cooperative UAVs. Various tasks such as image processing, path planning, communication, data acquisition, and field mapping are employed in the architecture to provide an efficient system. Besides, different limitation for applying Multi-UAVs in precision agriculture has been considered in designing the architecture. The architecture provides an autonomous end-to-end solution, starting from mission planning, data acquisition and image processing framework that is highly efficient and can enable farmers to comprehensively deploy UAVs onto their lands. Simulation and field tests shows that the architecture offers a number of advantages that include fault-tolerance, robustness, developer and user-friendliness.

Sequential experimental design under expensive, gradient-free objectives is a central challenge in computational statistics: evaluation budgets are tightly constrained and information must be extracted efficiently from each observation. We propose \textbf{ALMAB-DC}, a GP-based sequential design framework combining active learning, multi-armed bandits (MAB), and distributed asynchronous computing for expensive black-box experimentation. A Gaussian process surrogate with uncertainty-aware acquisition identifies informative query points; a UCB or Thompson-sampling bandit controller allocates evaluations across parallel workers; and an asynchronous scheduler handles heterogeneous runtimes. We present cumulative regret bounds for the bandit components and characterize parallel scalability via Amdahl's Law. We validate ALMAB-DC on five benchmarks. On the two statistical experimental-design tasks, ALMAB-DC achieves lower simple regret than Equal Spacing, Random, and D-optimal designs in dose--response optimization, and in adaptive spatial field estimation matches the Greedy Max-Variance benchmark while outperforming Latin Hypercube Sampling; at $K=4$ the distributed setting reaches target performance in one-quarter of sequential wall-clock rounds. On three ML/engineering tasks (CIFAR-10 HPO, CFD drag minimization, MuJoCo RL), ALMAB-DC achieves 93.4\% CIFAR-10 accuracy (outperforming BOHB by 1.7\,pp and Optuna by 1.1\,pp), reduces airfoil drag to $C_D = 0.059$ (36.9\% below Grid Search), and improves RL return by 50\% over Grid Search. All advantages over non-ALMAB baselines are statistically significant under Bonferroni-corrected Mann--Whitney $U$ tests. Distributed execution achieves $7.5\times$ speedup at $K = 16$ agents, consistent with Amdahl's Law.

Architecture views are essential for software architecture documentation, yet their manual creation is labor intensive and often leads to outdated artifacts. As systems grow in complexity, the automated generation of views from source code becomes increasingly valuable. Goal: We empirically evaluate the ability of LLMs and agentic approaches to generate architecture views from source code. Method: We analyze 340 open-source repositories across 13 experimental configurations using 3 LLMs with 3 prompting techniques and 2 agentic approaches, yielding 4,137 generated views. We evaluate the generated views by comparing them with the ground-truth using a combination of automated metrics complemented by human evaluations. Results: Prompting strategies offer marginal improvements. Few-shot prompting reduces clarity failures by 9.2% compared to zero-shot baselines. The custom agentic approach consistently outperforms the general-purpose agent, achieving the best clarity (22.6% failure rate) and level-of-detail success (50%). Conclusions: LLM and agentic approaches demonstrate capabilities in generating syntactically valid architecture views. However, they consistently exhibit granularity mismatches, operating at the code level rather than architectural abstractions. This suggests that there is still a need for human expertise, positioning LLMs and agents as assistive tools rather than autonomous architects.

Reinforcement learning with verifiable rewards (RLVR) plays a crucial role in expanding the capacities of LLM reasoning, but GRPO-style training is dominated by expensive rollouts and wastes compute on unusable prompts. We propose Prompt Replay, an overhead-free online data selection method for GRPO that reuses prompts only (not trajectories), to preserve on-policy optimization. After each step, we insert prompts with medium difficulty into a buffer, and prioritize prompts closer to a pass rate of 0.5 (half answers correct, half wrong) to maximize the advantage, thus learning signal. Training batches are formed by mixing reused prompts with fresh samples, with cooldown steps and max reuse times controlling aggressiveness vs risk of overfitting. Across multiple model families (Llama-3.2- 3B, Qwen3-8B) and training datasets (Dolci, Polaris), evaluated using average accuracy on six standard math benchmarks, Prompt Replay reduces zero-variance prompts, increases mean absolute advantage and shows faster initial accuracy gains. Yet, it plateaus and converges with the baseline, as too aggressive configuration was used. The method is most efficient when the rollouts are the primary bottleneck and the dataset is difficult for the model. We additionally observe that Qwen2.5-Math can exhibit spurious-reward effects that invalidates ablations, raising a warning signal for using it as a sole testbed for GRPO method research.

Reinforcement learning from human feedback (RLHF) has proven effective in aligning large language models with human preferences, inspiring the development of reward-centric diffusion reinforcement learning (RDRL) to achieve similar alignment and controllability. While diffusion models can generate high-quality outputs, RDRL remains susceptible to reward hacking, where the reward score increases without corresponding improvements in perceptual quality. We demonstrate that this vulnerability arises from the non-robustness of reward model gradients, particularly when the reward landscape with respect to the input image is sharp. To mitigate this issue, we introduce methods that exploit gradients from a robustified reward model without requiring its retraining. Specifically, we employ gradients from a flattened reward model, obtained through parameter perturbations of the diffusion model and perturbations of its generated samples. Empirically, each method independently alleviates reward hacking and improves robustness, while their joint use amplifies these benefits. Our resulting framework, RSA-FT (Reward Sharpness-Aware Fine-Tuning), is simple, broadly compatible, and consistently enhances the reliability of RDRL.

Semantic Textual Similarity (STS) is a crucial component of many Natural Language Processing (NLP) applications. However, existing approaches typically reduce semantic nuances to a single score, limiting interpretability. To address this, we introduce the task of Dissimilar Span Detection (DSD), which aims to identify semantically differing spans between pairs of texts. This can help users understand which particular words or tokens negatively affect the similarity score, or be used to improve performance in STS-dependent downstream tasks. Furthermore, we release a new dataset suitable for the task, the Span Similarity Dataset (SSD), developed through a semi-automated pipeline combining large language models (LLMs) with human verification. We propose and evaluate different baseline methods for DSD, both unsupervised, based on LIME, SHAP, LLMs, and our own method, as well as an additional supervised approach. While LLMs and supervised models achieve the highest performance, overall results remain low, highlighting the complexity of the task. Finally, we set up an additional experiment that shows how DSD can lead to increased performance in the specific task of paraphrase detection.

This paper investigates the fundamental mechanisms driving plasticity loss in deep reinforcement learning (RL), a critical challenge where neural networks lose their ability to adapt to non-stationary environments. While existing research often relies on descriptive metrics like dormant neurons or effective rank, these summaries fail to explain the underlying optimization dynamics. We propose the Optimization-Centric Plasticity (OCP) hypothesis, which posits that plasticity loss arises because optimal points from previous tasks become poor local optima for new tasks, trapping parameters during task transitions and hindering subsequent learning. We theoretically establish the equivalence between neuron dormancy and zero-gradient states, demonstrating that the absence of gradient signals is the primary driver of dormancy. Our experiments reveal that plasticity loss is highly task-specific; notably, networks with high dormancy rates in one task can achieve performance parity with randomly initialized networks when switched to a significantly different task, suggesting that the network's capacity remains intact but is inhibited by the specific optimization landscape. Furthermore, our hypothesis elucidates why parameter constraints mitigate plasticity loss by preventing deep entrenchment in local optima. Validated across diverse non-stationary scenarios, our findings provide a rigorous optimization-based framework for understanding and restoring network plasticity in complex RL domains.

Selective prediction systems can mitigate harms resulting from language model hallucinations by abstaining from answering in high-risk cases. Uncertainty quantification techniques are often employed to identify such cases, but are rarely evaluated in the context of the wider selective prediction policy and its ability to operate at low target error rates. We identify a model-dependent failure mode of entropy-based uncertainty methods that leads to unreliable abstention behaviour, and address it by combining entropy scores with a correctness probe signal. We find that across three QA benchmarks (TriviaQA, BioASQ, MedicalQA) and four model families, the combined score generally improves both the risk--coverage trade-off and calibration performance relative to entropy-only baselines. Our results highlight the importance of deployment-facing evaluation of uncertainty methods, using metrics that directly reflect whether a system can be trusted to operate at a stated risk level.

The continual learning literature has rapidly shifted from traditional class incremental learning (CIL) techniques to foundation model (FM)-based CIL methods without a clear understanding of how these newer approaches compare to strong, lightweight convolutional baselines. This abrupt transition has created a substantial methodological gap, making it difficult to assess whether recent FM-based CIL progress reflects genuine advances or merely the absence of rigorous baselines. To address this gap, we introduce Pruned Adaptation Modules (PAM), a simple yet effective method that freezes the vast majority of the pre-trained ResNet while enabling scalable continual adaptation through sparse task-specific layers. PAM yields up to a ~5x reduction in trainable parameters and a ~6x reduction in total parameters, significantly reducing the cost of continual updates. Across diverse benchmarks, PAM consistently mitigates catastrophic forgetting and outperforms state-of-the-art FM-based CIL approaches. Our findings position PAM as a strong and transparent baseline that helps bridge the gap between traditional and FM-based CIL, guiding future research for a more accurate assessment of true progress in continual adaptation. The code can be found at: https://github.com/ElifCerenGokYildirim/PAM.

Zeroth-order (ZO) optimization enables memory-efficient training of neural networks by estimating gradients via forward passes only, eliminating the need for backpropagation. However, the stochastic nature of gradient estimation significantly obscures the training dynamics, in contrast to the well-characterized behavior of first-order methods under Neural Tangent Kernel (NTK) theory. To address this, we introduce the Neural Zeroth-order Kernel (NZK) to describe model evolution in function space under ZO updates. For linear models, we prove that the expected NZK remains constant throughout training and depends explicitly on the first and second moments of the random perturbation directions. This invariance yields a closed-form expression for model evolution under squared loss. We further extend the analysis to linearized neural networks. Interpreting ZO updates as kernel gradient descent via NZK provides a novel perspective for potentially accelerating convergence. Extensive experiments across synthetic and real-world datasets (including MNIST, CIFAR-10, and Tiny ImageNet) validate our theoretical results and demonstrate acceleration when using a single shared random vector.

3D point cloud neural networks have significantly enhanced the perceptual capabilities of resource-limited mobile intelligent systems. However, despite the transformative impact, the point cloud algorithm suffers from substantial memory access during data preprocessing and imposes a burdensome workload on feature computing, resulting in high energy consumption and latency. In this paper, an efficient SRAM-based computing-in-memory (SRAM-CIM) accelerator (PC2IM), is proposed to alleviate memory access bottlenecks in point-based 3D point cloud networks. A data preprocessing module driven by the customized CIM engines is proposed and incorporated into a memory-efficient data flow. Specifically, an approximate distance SRAM-CIM (APD-CIM) is introduced to eliminate the repetitive on-chip memory access for point clouds that are spatially partitioned by the median and reduce the volume of temporary distance data. Building on the APD-CIM, a two-level Ping-Pong-MAX Content Addressable Memory (Ping-Pong-MAX CAM) is introduced to adaptively update temporary distances and perform in-situ search for the maximum, further reducing memory access. Additionally, an efficient CIM-based feature computing engine, named split-concatenate SRAM-CIM, is presented to minimize computation latency in multi-layer perceptron with high-precision input, while maintaining high area and energy efficiency. Experiment results show that the proposed PC2IM demonstrates 1.5x speedup and 2.7x enhanced energy efficiency compared to state-of-the-art point cloud accelerator. Moreover, PC2IM achieves 3.5x speedup and 1518.9x enhanced energy efficiency compared to GPU implementations.

Bangla culture is richly expressed through region, dialect, history, food, politics, media, and everyday visual life, yet it remains underrepresented in multimodal evaluation. To address this gap, we introduce BanglaVerse, a culturally grounded benchmark for evaluating multilingual vision-language models (VLMs) on Bengali culture across historically linked languages and regional dialects. Built from 1,152 manually curated images across nine domains, the benchmark supports visual question answering and captioning, and is expanded into four languages and five Bangla dialects, yielding ~32.3K artifacts. Our experiments show that evaluating only standard Bangla overestimates true model capability: performance drops under dialectal variation, especially for caption generation, while historically linked languages such as Hindi and Urdu retain some cultural meaning but remain weaker for structured reasoning. Across domains, the main bottleneck is missing cultural knowledge rather than visual grounding alone, with knowledge-intensive categories. These findings position BanglaVerse as a more realistic test bed for measuring culturally grounded multimodal understanding under linguistic variation.

Neural tree search is a powerful decision-making algorithm widely used in complex domains such as game playing and model-based reinforcement learning. Recent work has applied AlphaZero-style tree search to enhance the reasoning capabilities of Large Language Models (LLMs) during inference, but we find that this approach suffers from a scaling failure: on GSM8K and Game24, accuracy drops as the search budget increases. In this paper, we present ReSCALE, an adaptation of Gumbel AlphaZero MCTS that replaces Dirichlet noise and PUCT selection with Gumbel sampling and Sequential Halving, restoring monotonic scaling without changes to the model or its training. ReSCALE reaches 58.4\% on GSM8K and 85.3\% on Game24 at budgets where the baseline degrades. Ablations confirm that Sequential Halving is the primary driver of the improvement.

Epistemic intelligence requires machine learning systems to recognise the limits of their own knowledge and act safely under uncertainty, especially when faced with unknown unknowns. Existing uncertainty quantification methods rely on a single signal such as confidence or density and fail to detect diverse structural anomalies. We introduce SPECTRE-G2, a multi-signal anomaly detector that combines eight complementary signals from a dual-backbone neural network. The architecture includes a spectral normalised Gaussianization encoder, a plain MLP preserving feature geometry, and an ensemble of five models. These produce density, geometry, uncertainty, discriminative, and causal signals. Each signal is normalised using validation statistics and calibrated with synthetic out-of-distribution data. An adaptive top-k fusion selects the most informative signals and averages their scores. Experiments on synthetic, Adult, CIFAR-10, and Gridworld datasets show strong performance across diverse anomaly types, outperforming multiple baselines on AUROC, AUPR, and FPR95. The model is stable across seeds and particularly effective for detecting new variables and confounders. SPECTRE-G2 provides a practical approach for detecting unknown unknowns in open-world settings.

Text-attributed graphs (TAGs) enhance graph learning by integrating rich textual semantics and topological context for each node. While boosting expressiveness, they also expose new vulnerabilities in graph learning through text-based adversarial surfaces. Recent advances leverage diverse backbones, such as graph neural networks (GNNs) and pre-trained language models (PLMs), to capture both structural and textual information in TAGs. This diversity raises a key question: How can we design universal adversarial attacks that generalize across architectures to assess the security of TAG models? The challenge arises from the stark contrast in how different backbones-GNNs and PLMs-perceive and encode graph patterns, coupled with the fact that many PLMs are only accessible via APIs, limiting attacks to black-box settings. To address this, we propose BadGraph, a novel attack framework that deeply elicits large language models (LLMs) understanding of general graph knowledge to jointly perturb both node topology and textual semantics. Specifically, we design a target influencer retrieval module that leverages graph priors to construct cross-modally aligned attack shortcuts, thereby enabling efficient LLM-based perturbation reasoning. Experiments show that BadGraph achieves universal and effective attacks across GNN- and LLM-based reasoners, with up to a 76.3% performance drop, while theoretical and empirical analyses confirm its stealthy yet interpretable nature.

Learning from Label Proportions (LLP) is a weakly supervised problem in which the training data comprise bags, that is, groups of instances, each annotated only with bag-level class label proportions, and the objective is to learn a classifier that predicts instance-level labels. This setting is widely applicable when privacy constraints limit access to instance-level annotations or when fine-grained labeling is costly or impractical. In this work, we introduce a method that leverages Dual proportion Constraints (LLP-DC) during training, enforcing them at both the bag and instance levels. Specifically, the bag-level training aligns the mean prediction with the given proportion, and the instance-level training aligns hard pseudo-labels that satisfy the proportion constraint, where a minimum-cost maximum-flow algorithm is used to generate hard pseudo-labels. Extensive experimental results across various benchmark datasets empirically validate that LLP-DC consistently improves over previous LLP methods across datasets and bag sizes. The code is publicly available at https://github.com/TianhaoMa5/CV PR2026_Findings_LLP_DC.

Inferring the physical mechanisms that govern earthquake sequences from indirect geophysical observations remains difficult, particularly across tectonically distinct environments where similar seismic patterns can reflect different underlying processes. Current interpretations rely heavily on the expert synthesis of catalogs, spatiotemporal statistics, and candidate physical models, limiting reproducibility and the systematic transfer of insight across settings. Here we present TRACE (Trans-perspective Reasoning and Automated Comprehensive Evaluator), a multi-agent system that combines large language model planning with formal seismological constraints to derive auditable, physically grounded mechanistic inference from raw observations. Applied to the 2019 Ridgecrest sequence, TRACE autonomously identifies stress-perturbation-induced delayed triggering, resolving the cascading interaction between the Mw 6.4 and Mw 7.1 mainshocks; in the Santorini-Kolumbo case, the system identifies a structurally guided intrusion model, distinguishing fault-channeled episodic migration from the continuous propagation expected in homogeneous crustal failure. By providing a generalizable logical infrastructure for interpreting heterogeneous seismic phenomena, TRACE advances the field from expert-dependent analysis toward knowledge-guided autonomous discovery in Earth sciences.

An autonomous AI ecosystem (SUBSTRATE S3), generating product specifications without explicit instructions about formal methods, independently proposed the use of Z3 SMT solver across six distinct domains of AI safety: verification of LLM-generated code, tool API safety for AI agents, post-distillation reasoning correctness, CLI command validation, hardware assembly verification, and smart contract safety. These convergent discoveries, occurring across 8 products over 13 days with Jaccard similarity below 15% between variants, suggest that formal verification is not merely a useful technique for AI safety but an emergent property of any sufficiently complex system reasoning about its own safety. We propose a unified framework (substrate-guard) that applies Z3-based verification across all six output classes through a common API, and evaluate it on 181 test cases across five implemented domains, achieving 100% classification accuracy with zero false positives and zero false negatives. Our framework detected real bugs that empirical testing would miss, including an INT_MIN overflow in branchless RISC-V assembly and mathematically proved that unconstrained string parameters in tool APIs are formally unverifiable.

The computational demands of modern AI services are increasingly shifting execution beyond centralized clouds toward a computing continuum spanning edge and end devices. However, the scale, heterogeneity, and cross-layer dependencies of these environments make resilience difficult to maintain. Existing fault-management methods are often too static, fragmented, or heavy to support timely self-healing, especially under noisy logs and edge resource constraints. To address these limitations, this paper presents NeSy-Edge, a neuro-symbolic framework for trustworthy self-healing in the computing continuum. The framework follows an edge-first design, where a resource-constrained edge node performs local perception and reasoning, while a cloud model is invoked only at the final diagnosis stage. Specifically, NeSy-Edge converts raw runtime logs into structured event representations, builds a prior-constrained sparse symbolic causal graph, and integrates causal evidence with historical troubleshooting knowledge for root-cause analysis and recovery recommendation. We evaluate our work on representative Loghub datasets under multiple levels of semantic noise, considering parsing quality, causal reasoning, end-to-end diagnosis, and edge-side resource usage. The results show that NeSy-Edge remains robust even at the highest noise level, achieving up to 75% root-cause analysis accuracy and 65% end-to-end accuracy while operating within about 1500 MB of local memory.

This paper proposes a new formulation of functional Gaussian Process regression in manifolds, based on an Empirical Bayes approach, in the spatiotemporal random field context. We apply the machinery of tight Gaussian measures in separable Hilbert spaces, exploiting the invariance property of covariance kernels under the group of isometries of the manifold. The identification of these measures with infinite-product Gaussian measures is then obtained via the eigenfunctions of the Laplace-Beltrami operator on the manifold. The involved time-varying angular spectra constitute the key tool for dimension reduction in the implementation of this regression approach, adopting a suitable truncation scheme depending on the functional sample size. The simulation study and synthetic data application undertaken illustrate the finite sample and asymptotic properties of the proposed functional regression predictor.

Training large language models (LLMs) with synthetic reasoning data has become a popular approach to enhancing their reasoning capabilities, while a key factor influencing the effectiveness of this paradigm is the quality of the generated multi-step reasoning data. To generate high-quality reasoning data, many recent methods generate synthetic reasoning paths and filter them based on final answer correctness, often overlooking flaws in intermediate reasoning steps. To enhance the verification of intermediate reasoning steps, prior work primarily resorts to code execution or symbolic reasoning engines. However, code-based validation is restricted to code or mathematical tasks, and reasoning engines require a well-structured and complete context. As a result, existing methods fail to function effectively in natural language reasoning tasks that involve ambiguous or incomplete contexts. In these tasks, synthetic data still lack reliable checks for verifying each reasoning step. To address this challenge, we introduce ORACLE, a structured data generation framework inspired by syllogistic reasoning. ORACLE integrates the generative strengths of LLMs with symbolic supervision: the LLM produces step-wise reasoning contexts, while a symbolic reasoning engine verifies the validity of each intermediate step. By employing a unified prompting template to elicit modular reasoning chains, ORACLE enables fine-grained, step-level validation, facilitating the construction of high-quality multi-step reasoning data. Across six logical, factual, and commonsense reasoning benchmarks, our ORACLE consistently outperforms strong baselines on multiple models.

This paper addresses the challenge of improving information retrieval from semi-structured eXtensible Markup Language (XML) documents. Traditional information retrieval systems (IRS) often overlook user-specific needs and return identical results for the same query, despite differences in users' knowledge, preferences, and objectives. We integrate external semantic resources, namely a domain ontology and user profiles, into the retrieval process. Documents, queries, and user profiles are represented as vectors of weighted concepts. The ontology applies a concept-weighting mechanism that emphasizes highly specific concepts, as lower-level nodes in the hierarchy provide more precise and targeted information. Relevance is assessed using semantic similarity measures that capture conceptual relationships beyond keyword matching, enabling personalized and fine-grained matching among user profiles, queries, and documents. Experimental results show that combining ontologies with user profiles improves retrieval effectiveness, achieving higher precision and recall than keyword-based approaches. Overall, the proposed framework enhances the relevance and adaptability of XML search results, supporting more user-centered retrieval.

Test-time adaptation (TTA) adapts pre-trained models to distribution shifts at inference using only unlabeled test data. Under the Practical TTA (PTTA) setting, where test streams are temporally correlated and non-i.i.d., memory has become an indispensable component for stable adaptation, yet existing methods universally store amples in a single unstructured pool. We show that this single-cluster design is fundamentally mismatched to PTTA: a stream clusterability analysis reveals that test streams are inherently multi-modal, with the optimal number of mixture components consistently far exceeding one. To close this structural gap, we propose Multi-Cluster Memory (MCM), a plug-and-play framework that organizes stored samples into multiple clusters using lightweight pixel-level statistical descriptors. MCM introduces three complementary mechanisms: descriptor-based cluster assignment to capture distinct distributional modes, Adjacent Cluster Consolidation (ACC) to bound memory usage by merging the most similar temporally adjacent clusters, and Uniform Cluster Retrieval (UCR) to ensure balanced supervision across all modes during adaptation. Integrated with three contemporary TTA methods on CIFAR-10-C, CIFAR-100-C, ImageNet-C, and DomainNet, MCM achieves consistent improvements across all 12 configurations, with gains up to 5.00% on ImageNet-C and 12.13% on DomainNet. Notably, these gains scale with distributional complexity: larger label spaces with greater multi-modality benefit most from multi-cluster organization. GMM-based memory diagnostics further confirm that MCM maintains near-optimal distributional balance, entropy, and mode coverage, whereas single-cluster memory exhibits persistent imbalance and progressive mode loss. These results establish memory organization as a key design axis for practical test-time adaptation.

Multi-task learning (MTL) aims to enable a single model to solve multiple tasks efficiently; however, current parameter-efficient fine-tuning (PEFT) methods remain largely limited to single-task adaptation. We introduce \textbf{Free Sinewich}, a parameter-efficient multi-task learning framework that enables near-zero-cost weight modulation via frequency switching (\textbf{Free}). Specifically, a \textbf{Sine-AWB (Sinewich)} layer combines low-rank factors and convolutional priors into a single kernel, which is then modulated elementwise by a sinusoidal transformation to produce task-specialized weights. A lightweight Clock Net is introduced to produce bounded frequencies that stabilize this modulation during training. Theoretically, sine modulation enhances the rank of low-rank adapters, while frequency separation decorrelates the weights of different tasks. On dense prediction benchmarks, Free Sinewich achieves state-of-the-art performance-efficiency trade-offs (e.g., up to +5.39\% improvement over single-task fine-tuning with only 6.53M trainable parameters), offering a compact and scalable paradigm based on frequency-based parameter sharing. Project page: \href{https://casperliuliuliu.github.io/projects/Free-Sinewich/}{https://casperliuliuliu.github.io/projects/Free-Sinewich}.

Molecular property prediction constitutes a cornerstone of drug discovery and materials science, necessitating models capable of disentangling complex structure-property relationships across diverse molecular modalities. Existing approaches frequently exhibit entangled representations--conflating structural, chemical, and functional factors--thereby limiting interpretability and transferability. Furthermore, conventional methods inadequately exploit complementary information from graphs, sequences, and geometries, often relying on naive concatenation that neglects inter-modal dependencies. In this work, we propose DMMRL, which employs variational autoencoders to disentangle molecular representations into shared (structure-relevant) and private (modality-specific) latent spaces, enhancing both interpretability and predictive performance. The proposed variational disentanglement mechanism effectively isolates the most informative features for property prediction, while orthogonality and alignment regularizations promote statistical independence and cross-modal consistency. Additionally, a gated attention fusion module adaptively integrates shared representations, capturing complex inter-modal relationships. Experimental validation across seven benchmark datasets demonstrates DMMRL's superior performance relative to state-of-the-art approaches. The code and data underlying this article are freely available at https://github.com/xulong0826/DMMRL.

Large Vision-Language Models (LVLMs) rely on dense visual tokens to capture fine-grained visual information, but processing all these tokens incurs substantial computational and memory overhead during inference. To address this issue, we propose ResPrune, a training-free visual token pruning framework that enables efficient LVLM inference by selecting a compact yet informative subset of visual tokens. ResPrune formulates visual token pruning as a subspace reconstruction problem and employs a greedy subspace expansion strategy guided by residual energy, allowing it to preserve the geometric structure of the original visual token space. To further incorporate cross modal alignment, the selection process is conditioned on textual relevance, encouraging the retention of tokens that are both informative and instruction-relevant. The proposed method is lightweight and model-agnostic, and can be seamlessly integrated into existing LVLM pipelines without retraining or architectural modifications. Extensive experiments on multiple LVLM backbones, including LLaVA-1.5, LLaVA-NeXT, and Qwen2.5-VL, demonstrate that ResPrune consistently outperforms existing pruning approaches across a wide range of benchmarks, while achieving effective reductions in computation, memory consumption, and inference latency.

Due to the limited availability of paired multi-modal data, multi-modal trackers are typically built by adopting pre-trained RGB models with parameter-efficient fine-tuning modules. However, these fine-tuning methods overlook advanced adaptations for applying RGB pre-trained models and fail to modulate a single specific modality, cross-modal interactions, and the prediction head. To address the issues, we propose to perform Progressive Adaptation for Multi-Modal Tracking (PATrack). This innovative approach incorporates modality-dependent, modality-entangled, and task-level adapters, effectively bridging the gap in adapting RGB pre-trained networks to multi-modal data through a progressive strategy. Specifically, modality-specific information is enhanced through the modality-dependent adapter, decomposing the high- and low-frequency components, which ensures a more robust feature representation within each modality. The inter-modal interactions are introduced in the modality-entangled adapter, which implements a cross-attention operation guided by inter-modal shared information, ensuring the reliability of features conveyed between modalities. Additionally, recognising that the strong inductive bias of the prediction head does not adapt to the fused information, a task-level adapter specific to the prediction head is introduced. In summary, our design integrates intra-modal, inter-modal, and task-level adapters into a unified framework. Extensive experiments on RGB+Thermal, RGB+Depth, and RGB+Event tracking tasks demonstrate that our method shows impressive performance against state-of-the-art methods. Code is available at https://github.com/ouha1998/Learning-Progressive-Adaptation-for-Multi-Modal-Tracking.

In this paper, we explore a joint source and reconfigurable intelligent surface (RIS)-assisted channel encoding (JSRE) framework for multi-user semantic communications, where a deep neural network (DNN) extracts semantic features for all users and the RIS provides channel orthogonality, enabling a unified semantic encoding-decoding design. We aim to maximize the overall energy efficiency of semantic communications across all users by jointly optimizing the user scheduling, the RIS's phase shifts, and the semantic compression ratio. Although this joint optimization problem can be addressed using conventional deep reinforcement learning (DRL) methods, evaluating semantic similarity typically relies on extensive real environment interactions, which can incur heavy computational overhead during training. To address this challenge, we propose a truncated DRL (T-DRL) framework, where a DNN-based semantic similarity estimator is developed to rapidly estimate the similarity score. Moreover, the user scheduling strategy is tightly coupled with the semantic model configuration. To exploit this relationship, we further propose a semantic model caching mechanism that stores and reuses fine-tuned semantic models corresponding to different scheduling decisions. A Transformer-based actor network is employed within the DRL framework to dynamically generate action space conditioned on the current caching state. This avoids redundant retraining and further accelerates the convergence of the learning process. Numerical results demonstrate that the proposed JSRE framework significantly improves the system energy efficiency compared with the baseline methods. By training fewer semantic models, the proposed T-DRL framework significantly enhances the learning efficiency.

Transformers lack an explicit architectural mechanism for storing and organizing knowledge acquired during training. We introduce learnable sparse memory banks: a set of latent tokens, randomly initialized and trained end-to-end, that transformer layers query via cross-attention to retrieve stored knowledge. To scale memory capacity without prohibitive attention costs, we propose chapter-based routing inspired by Mixture-of-Experts architectures, partitioning the memory bank into chapters and training a router to select relevant subsets per input. This enables scaling to 262K memory tokens while maintaining tractable computation. We evaluate our approach against standard transformers (in iso-FLOP settings) on pre-training and instruction fine-tuning across relevant benchmarks. Our models surpass iso-FLOP baselines suggesting scope for a new axis of scaling, demonstrating that explicit associative memory provides complementary capacity to what is captured implicitly in model parameters. Additionally, we observe improved knowledge retention under continued training, with robustness to forgetting when transitioning between training phases (e.g., pretraining to instruction fine-tuning).

Thyroid ultrasound is the first-line exam for assessing thyroid nodules and determining whether biopsy is warranted. In routine reporting, radiologists produce two coupled outputs: a nodule contour for measurement and a TI-RADS risk category based on sonographic criteria. Yet both contouring style and risk grading vary across readers, creating inconsistent supervision that can degrade standard learning pipelines. In this paper, we address this workflow with a clinically guided multitask framework that jointly predicts the nodule mask and TI-RADS category within a single model. To ground risk prediction in clinically meaningful evidence, we guide the classification embedding using a compact TI-RADS aligned radiomics target during training, while preserving complementary deep features for discriminative performance. However, under annotator variability, naive multitask optimization often fails not because the tasks are unrelated, but because their gradients compete within the shared representation. To make this competition explicit and controllable, we introduce RLAR, a representation-level adversarial gradient regularizer. Rather than performing parameter-level gradient surgery, RLAR uses each task's normalized adversarial direction in latent space as a geometric probe of task sensitivity and penalizes excessive angular alignment between task-specific adversarial directions. On a public TI-RADS dataset, our clinically guided multitask model with RLAR consistently improves risk stratification while maintaining segmentation quality compared to single-task training and conventional multitask baselines. Code and pretrained models will be released.

Human annotation is central to NLP evaluation, yet subjective tasks often exhibit substantial variability across annotators. While large language models (LLMs) can provide structured reasoning to support annotation, their influence on human annotation behavior remains unclear. We introduce ReasonAlign, a reasoning-based annotation scaffold that exposes LLM-generated explanations while withholding predicted labels. We frame this as a controlled study of how reasoning affects human annotation behavior, rather than a full evaluation of annotation accuracy. Using a two-pass protocol inspired by Delphi-style revision, annotators first label instances independently and then revise their decisions after viewing model-generated reasoning. We evaluate the approach on sentiment classification and opinion detection tasks, analyzing changes in inter-annotator agreement and revision behavior. To quantify these effects, we introduce the Annotator Effort Proxy (AEP), a metric capturing the proportion of labels revised after exposure to reasoning. Our results show that exposure to reasoning is associated with increased agreement alongside minimal revision, suggesting that reasoning primarily helps resolve ambiguous cases without inducing widespread changes. These findings provide insight into how reasoning explanations shape annotation consistency and highlight reasoning-based scaffolds as a practical mechanism for supporting human-AI annotation workflows.

Based on some recent work of the author on stochastic approximation in non-markovian environments, the situation when the driving random process is non-ergodic in addition to being non-markovian is considered. Using this, we propose an analytic framework for understanding transformer based learning, specifically, the `attention' mechanism, and continual learning, both of which depend on the entire past in principle.

High-quality text representations are crucial for natural language understanding (NLU), but low-resource languages like Vietnamese face challenges due to limited annotated data. While pre-trained models like PhoBERT and CafeBERT perform well, their effectiveness is constrained by data scarcity. Contrastive learning (CL) has recently emerged as a promising approach for improving sentence representations, enabling models to effectively distinguish between semantically similar and dissimilar sentences. We propose ViCLSR (Vietnamese Contrastive Learning for Sentence Representations), a novel supervised contrastive learning framework specifically designed to optimize sentence embeddings for Vietnamese, leveraging existing natural language inference (NLI) datasets. Additionally, we propose a process to adapt existing Vietnamese datasets for supervised learning, ensuring compatibility with CL methods. Our experiments demonstrate that ViCLSR significantly outperforms the powerful monolingual pre-trained model PhoBERT on five benchmark NLU datasets such as ViNLI (+6.97% F1), ViWikiFC (+4.97% F1), ViFactCheck (+9.02% F1), UIT-ViCTSD (+5.36% F1), and ViMMRC2.0 (+4.33% Accuracy). ViCLSR shows that supervised contrastive learning can effectively address resource limitations in Vietnamese NLU tasks and improve sentence representation learning for low-resource languages. Furthermore, we conduct an in-depth analysis of the experimental results to uncover the factors contributing to the superior performance of contrastive learning models. ViCLSR is released for research purposes in advancing natural language processing tasks.

This study proposes a segmental-level prosodic probing framework to evaluate neural TTS models' ability to reproduce consonant-induced f0 perturbation, a fine-grained segmental-prosodic effect that reflects local articulatory mechanisms. We compare synthetic and natural speech realizations for thousands of words, stratified by lexical frequency, using Tacotron 2 and FastSpeech 2 trained on the same speech corpus (LJ Speech). These controlled analyses are then complemented by a large-scale evaluation spanning multiple advanced TTS systems. Results show accurate reproduction for high-frequency words but poor generalization to low-frequency items, suggesting that the examined TTS architectures rely more on lexical-level memorization than on abstract segmental-prosodic encoding. This finding highlights a limitation in such TTS systems' ability to generalize prosodic detail beyond seen data. The proposed probe offers a linguistically informed diagnostic framework that may inform future TTS evaluation methods, and has implications for interpretability and authenticity assessment in synthetic speech.

Composing coherent long-form music remains a significant challenge due to the complexity of modeling long-range dependencies and the prohibitive memory and computational requirements associated with lengthy audio representations. In this work, we propose a simple yet powerful trick: we assume that AI models can understand and generate time-accelerated (speeded-up) audio at rates such as 2x, 4x, or even 8x. By first generating a high-speed version of the music, we greatly reduce the temporal length and resource requirements, making it feasible to handle long-form music that would otherwise exceed memory or computational limits. The generated audio is then restored to its original speed, recovering the full temporal structure. This temporal speed-up and slow-down strategy naturally follows the principle of hierarchical generation from abstract to detailed content, and can be conveniently applied to existing music generation models to enable long-form music generation. We instantiate this idea in SqueezeComposer, a framework that employs diffusion models for generation in the accelerated domain and refinement in the restored domain. We validate the effectiveness of this approach on two tasks: long-form music generation, which evaluates temporal-wise control (including continuation, completion, and generation from scratch), and whole-song singing accompaniment generation, which evaluates track-wise control. Experimental results demonstrate that our simple temporal speed-up trick enables efficient, scalable, and high-quality long-form music generation. Audio samples are available at https://SqueezeComposer.github.io/.

As one of the most important underwater sensing technologies, forward-looking sonar exhibits unique imaging characteristics. Sonar images are often affected by severe speckle noise, low texture contrast, acoustic shadows, and geometric distortions. These factors make it difficult for traditional teacher-student frameworks to achieve satisfactory performance in sonar semantic segmentation tasks under extremely limited labeled data conditions. To address this issue, we propose a Collaborative Teacher Semantic Segmentation Framework for forward-looking sonar images. This framework introduces a multi-teacher collaborative mechanism composed of one general teacher and multiple sonar-specific teachers. By adopting a multi-teacher alternating guidance strategy, the student model can learn general semantic representations while simultaneously capturing the unique characteristics of sonar images, thereby achieving more comprehensive and robust feature modeling. Considering the challenges of sonar images, which can lead teachers to generate a large number of noisy pseudo-labels, we further design a cross-teacher reliability assessment mechanism. This mechanism dynamically quantifies the reliability of pseudo-labels by evaluating the consistency and stability of predictions across multiple views and multiple teachers, thereby mitigating the negative impact caused by noisy pseudo-labels. Notably, on the FLSMD dataset, when only 2% of the data is labeled, our method achieves a 5.08% improvement in mIoU compared to other state-of-the-art approaches.

Context: Simulation-based testing is a cost-efficient alternative to field testing for Autonomous Vehicles (AVs), but generating safety-critical test cases is challenging due to the vast search space. Prior work has studied static (road features) and dynamic (AV behavior) features of test scenarios separately, but their inter-dependencies are underexplored. Objective: In this paper, we describe an empirical to analyze how static and dynamic featuresof test scenarios, and their inter-dependencies, influence AV test scenario outcomes. Method: This study proposes an integrated approach using Instance Space Analysis (ISA) toevaluate both types of features, identify key influences on AV safety, and predict test outcomeswithout execution. Results: Our study identifies critical features affecting test outcomes (effective/ineffective, depending on whether it leads to a safety-critical condition). Results show that combining static and dynamic features improves prediction accuracy, confirmed by models trained on both feature types outperforming models trained with only one type of feature. Conclusion: The interplay of static and dynamic features enhances fault detection in AV testing. This research underscores the importance of integrating both types of features to create more effective testing frameworks for autonomous systems. Key contributions include: (1) a unified framework for AV safety assessment, (2) identification of influential features using ISA, and (3) efficient test outcome prediction for optimized regression testing.

We introduce LongCat-Flash-Prover, a flagship 560-billion-parameter open-source Mixture-of- Experts (MoE) model that advances Native Formal Reasoning in Lean4 through agentic tool-integrated reasoning (TIR). We decompose the native formal reasoning task into three independent formal capabilities, i.e., auto-formalization, sketching, and proving. To facilitate these capabilities, we propose a Hybrid-Experts Iteration Framework to expand high-quality task trajectories, including generating a formal statement based on a given informal problem, producing a whole-proof directly from the statement, or a lemma-style sketch. During agentic RL, we present a Hierarchical Importance Sampling Policy Optimization (HisPO) algorithm, which aims to stabilize the MoE model training on such long-horizon tasks. It employs a gradient masking strategy that accounts for the policy staleness and the inherent train-inference engine discrepancies at both sequence and token levels. Additionally, we also incorporate theorem consistency and legality detection mechanisms to eliminate reward hacking issues. Extensive evaluations show that our LongCat-Flash-Prover sets a new state-of-the-art for open-weights models in both auto-formalization and theorem proving. Demonstrating remarkable sample efficiency, it achieves a 97.1% pass rate on MiniF2F-Test using only 72 inference budget per problem. On more challenging benchmarks, it solves 70.8% of ProverBench and 41.5% of PutnamBench with no more than 220 attempts per problem, significantly outperforming existing open-weights baselines.

We study the problem of learning a low-degree spherical polynomial of degree $k_0 = Θ(1) \ge 1$ defined on the unit sphere in $\RR^d$ by training an over-parameterized two-layer neural network with augmented feature in this paper. Our main result is the significantly improved sample complexity for learning such low-degree polynomials. We show that, for any regression risk $\eps \in (0, Θ(d^{-k_0})]$, an over-parameterized two-layer neural network trained by a novel Gradient Descent with Projection (GDP) requires a sample complexity of $n \asymp Θ( \log(4/δ) \cdot d^{k_0}/\eps)$ with probability $1-δ$ for $δ\in (0,1)$, in contrast with the representative sample complexity $Θ(d^{k_0} \max\set{\eps^{-2},\log d})$. Moreover, such sample complexity is nearly unimprovable since the trained network renders a nearly optimal rate of the nonparametric regression risk of the order $\log({4}/δ) \cdot Θ(d^{k_0}/{n})$ with probability at least $1-δ$. On the other hand, the minimax optimal rate for the regression risk with a kernel of rank $Θ(d^{k_0})$ is $Θ(d^{k_0}/{n})$, so that the rate of the nonparametric regression risk of the network trained by GDP is nearly minimax optimal. In the case that the ground truth degree $k_0$ is unknown, we present a novel and provable adaptive degree selection algorithm which identifies the true degree and achieves the same nearly optimal regression rate. To the best of our knowledge, this is the first time that a nearly optimal risk bound is obtained by training an over-parameterized neural network with a popular activation function (ReLU) and algorithmic guarantee for learning low-degree spherical polynomials. Due to the feature learning capability of GDP, our results are beyond the regular Neural Tangent Kernel (NTK) limit.

Smart contracts have transformed decentralized finance, but flaws in their logic still create major security threats. Most existing vulnerability detection techniques focus on well-supported languages like Solidity, while low-resource counterparts such as Vyper remain largely underexplored due to scarce analysis tools and limited labeled datasets. Training a robust detection model directly on Vyper is particularly challenging, as collecting sufficiently large and diverse Vyper training datasets is difficult in practice. To address this gap, we introduce Sol2Vy, a novel framework that enables cross-language knowledge transfer from Solidity to Vyper, allowing vulnerability detection on Vyper using models trained exclusively on Solidity. This approach eliminates the need for extensive labeled Vyper datasets typically required to build a robust vulnerability detection model. We implement and evaluate Sol2Vy on various critical vulnerability types, including reentrancy, weak randomness, and unchecked transfer. Experimental results show that Sol2Vy, despite being trained exclusively on Solidity, achieves strong detection performance on Vyper contracts and significantly outperforms prior state-of-the-art methods.

Semi-Supervised Text Classification (SSTC) mainly works under the spirit of self-training. They initialize the deep classifier by training over labeled texts; and then alternatively predict unlabeled texts as their pseudo-labels and train the deep classifier over the mixture of labeled and pseudo-labeled texts. Naturally, their performance is largely affected by the accuracy of pseudo-labels for unlabeled texts. Unfortunately, they often suffer from low accuracy because of the margin bias problem caused by the large difference between representation distributions of labels in SSTC. To alleviate this problem, we apply the angular margin loss, and perform several Gaussian linear transformations to achieve balanced label angle variances, i.e., the variance of label angles of texts within the same label. More accuracy of predicted pseudo-labels can be achieved by constraining all label angle variances balanced, where they are estimated over both labeled and pseudo-labeled texts during self-training loops. With this insight, we propose a novel SSTC method, namely Semi-Supervised Text Classification with Balanced Deep representation Distributions (S2TC-BDD). We implement both multi-class classification and multi-label classification versions of S2TC-BDD by introducing some pseudo-labeling tricks and regularization terms. To evaluate S2 TC-BDD, we compare it against the state-of-the-art SSTC methods. Empirical results demonstrate the effectiveness of S2 TC-BDD, especially when the labeled texts are scarce.

Nowadays, the widespread dissemination of misinformation across numerous social media platforms has led to severe negative effects on society. To address this challenge, the automatic detection of misinformation, particularly under multimedia scenarios, has gained significant attention from both academic and industrial communities, leading to the emergence of a research task known as Multimodal Misinformation Detection (MMD). Typically, current MMD approaches focus on capturing the semantic relationships and inconsistency between various modalities but often overlook certain critical indicators within multimodal content. Recent research has shown that manipulated features within visual content in social media articles serve as valuable clues for MMD. Meanwhile, we argue that the potential intentions behind the manipulation, e.g., harmful and harmless, also matter in MMD. Therefore, in this study, we aim to identify such multimodal misinformation by capturing two types of features: manipulation features, which represent if visual content has been manipulated, and intention features, which assess the nature of these manipulations, distinguishing between harmful and harmless intentions. Unfortunately, the manipulation and intention labels that supervise these features to be discriminative are unknown. To address this, we introduce two weakly supervised indicators as substitutes by incorporating supplementary datasets focused on image manipulation detection and framing two different classification tasks as positive and unlabeled learning issues. With this framework, we introduce an innovative MMD approach, titled Harmful Visual Content Manipulation Matters in MMD (HAVC-M4 D). Comprehensive experiments conducted on four prevalent MMD datasets indicate that HAVC-M4 D significantly and consistently enhances the performance of existing MMD methods.

The identification of hazardous driving behaviors from in-cabin video streams is essential for enhancing road safety and supporting the detection of traffic violations and unsafe driver actions. However, current temporal action localization techniques often struggle to balance accuracy with computational efficiency. In this work, we develop and evaluate a temporal action localization framework tailored for driver monitoring scenarios, particularly suitable for periodic inspection settings such as transportation safety checkpoints or fleet management assessment systems. Our approach follows a two-stage pipeline that combines VideoMAE-based feature extraction with an Augmented Self-Mask Attention (AMA) detector, enhanced by a Spatial Pyramid Pooling-Fast (SPPF) module to capture multi-scale temporal features. Experimental results reveal a distinct trade-off between model capacity and efficiency. At the feature extraction stage, the ViT-Giant backbone delivers higher representations with 88.09% Top-1 test accuracy, while the ViT-based variant proves to be a practical alternative, achieving 82.55% accuracy with significantly lower computational fine-tuning costs (101.85 GFLOPs/segment compared to 1584.06 GFLOPs/segment for Giant). In the downstream localization task, the integration of SPPF consistently improves performance across all configurations. Notably, the ViT-Giant + SPPF model achieves a peak mAP of 92.67%, while the lightweight ViT-based configuration maintains robust results.

UAVs play an important role in applications such as autonomous exploration, disaster response, and infrastructure inspection. However, UAV VLN in complex 3D environments remains challenging. A key difficulty is the structural representation mismatch between 2D visual perception and the 3D trajectory decision space, which limits spatial reasoning. To this end, we propose SpatialFly, a geometry-guided spatial representation framework for UAV VLN. Operating on RGB observations without explicit 3D reconstruction, SpatialFly introduces a geometry-guided 2D representation alignment mechanism. Specifically, the geometric prior injection module injects global structural cues into 2D semantic tokens to provide scene-level geometric guidance. The geometry-aware reparameterization module then aligns 2D semantic tokens with 3D geometric tokens through cross-modal attention, followed by gated residual fusion to preserve semantic discrimination. Experimental results show that SpatialFly consistently outperforms state-of-the-art UAV VLN baselines across both seen and unseen environments, reducing NE by 4.03m and improving SR by 1.27% over the strongest baseline on the unseen Full split. Additional trajectory-level analysis shows that SpatialFly produces trajectories with better path alignment and smoother, more stable motion.

Diffusion-based image super-resolution (SR), which aims to reconstruct high-resolution (HR) images from corresponding low-resolution (LR) observations, faces a fundamental trade-off between inference efficiency and reconstruction quality. The state-of-the-art residual-shifting diffusion framework achieves efficient 4-step inference, yet suffers from severe performance degradation in compact sampling trajectories. This is mainly attributed to two core limitations: the inherent suboptimality of unconstrained random Gaussian noise in intermediate steps, which leads to error accumulation and insufficient LR prior guidance, and the initialization bias caused by naive bicubic upsampling. In this paper, we propose LPNSR, a prior-enhanced efficient diffusion framework to address these issues. We first mathematically derive the closed-form analytical solution of the optimal intermediate noise for the residual-shifting diffusion paradigm, and accordingly design an LR-guided multi-input-aware noise predictor to replace random Gaussian noise, embedding LR structural priors into the reverse process while fully preserving the framework's core efficient residual-shifting mechanism. We further mitigate initial bias with a high-quality pre-upsampling network to optimize the diffusion starting point. With a compact 4-step trajectory, LPNSR can be optimized in an end-to-end manner. Extensive experiments demonstrate that LPNSR achieves state-of-the-art perceptual performance on both synthetic and real-world datasets, without relying on any large-scale text-to-image priors. The source code of our method can be found at https://github.com/Faze-Hsw/LPNSR.

Humans must flexibly arbitrate between exploring alternatives and exploiting learned strategies, yet they frequently exhibit maladaptive persistence by continuing to execute failing strategies despite accumulating negative evidence. Here we propose a ``confidence-freeze'' account that reframes such persistence as a dynamic learning state rather than a stable dispositional trait. Using a multi-reversal two-armed bandit task across three experiments (total N = 332; 19,920 trials), we first show that human learners normally make use of the symmetric statistical structure inherent in outcome trajectories: runs of successes provide positive evidence for environmental stability and thus for strategy maintenance, whereas runs of failures provide negative evidence and should raise switching probability. Behaviour in the control group conformed to this normative pattern. However, individuals who experienced a high rate of early success (90\% vs.\ 60\%) displayed a robust and selective distortion after the first reversal: they persisted through long stretches of non-reward (mean = 6.2 consecutive losses) while their metacognitive confidence ratings simultaneously dropped from 5 to 2 on a 7-point scale.

Brain encoding and decoding aims to understand the relationship between external stimuli and brain activities, and is a fundamental problem in neuroscience. In this article, we study latent embedding alignment for brain encoding and decoding, with a focus on improving sample efficiency under limited fMRI-stimulus paired data and substantial subject heterogeneity. We propose a lightweight alignment framework equipped with two statistical learning components: inverse semi-supervised learning that leverages abundant unpaired stimulus embeddings through inverse mapping and residual debiasing, and meta transfer learning that borrows strength from pretrained models across subjects via sparse aggregation and residual correction. Both methods operate exclusively at the alignment stage while keeping encoders and decoders frozen, allowing for efficient computation, modular deployment, and rigorous theoretical analysis. We establish finite-sample generalization bounds and safety guarantees, and demonstrate competitive empirical performance on the large-scale fMRI-image reconstruction benchmark data.

Accurate air quality index (AQI) forecasting is essential for the protecting public health in rapidly growing urban regions, and the practical model evaluation and selection are often challenged by the lack of rigorous, region-specific benchmarking on standardized datasets. Physics-guided machine learning and deep learning models could be a good and effective solution to resolve such issues with more accurate and efficient AQI forecasting. This research study presents an explainable and comprehensive benchmark that enables a guideline and proposed physics-guided best model by benchmarking classical time-series, machine-learning, and deep-learning approaches for multi-horizon AQI forecasting in North Texas (Dallas County). Using publicly available U.S. Environmental Protection Agency (EPA) daily observations of air quality data from 2022 to 2024, we curate city-level time series for PM2.5 and O3 by aggregating station measurements and constructing lag-wise forecasting datasets for LAG in {1,7,14,30} days. For benchmarking the best model, linear regression (LR), SARIMAX, multilayer perceptrons (MLP), and LSTM networks are evaluated with the proposed physics-guided variants (MLP+Physics and LSTM+Physics) that incorporate the EPA breakpoint-based AQI formulation as a consistency constraint through a weighted loss. Experiments using chronological train-test splits and error metrics MAE, RMSE showed that deep-learning models outperform simpler baselines, while physics guidance improves stability and yields physically consistent pollutant with AQI relationships, with the largest benefits observed for short-horizon prediction and for PM2.5 and O3. Overall, the results provide a practical reference for selecting AQI forecasting models in North Texas and clarify when lightweight physics constraints meaningfully improve predictive performance across pollutants and forecast horizons.

As text-based computer-mediated communication (CMC) increasingly structures everyday interaction, a central question re-emerges with new urgency: How do users reconstruct nonverbal expression in environments where embodied cues are absent? This paper provides a systematic, theory-driven account of electronic nonverbal cues (eNVCs) - textual analogues of kinesics, vocalics, and paralinguistics - in public microblog communication. Across three complementary studies, we advance conceptual, empirical, and methodological contributions. Study 1 develops a unified taxonomy of eNVCs grounded in foundational nonverbal communication theory and introduces a scalable Python toolkit for their automated detection. Study 2, a within-subject survey experiment, offers controlled causal evidence that eNVCs substantially improve emotional decoding accuracy and lower perceived ambiguity, while also identifying boundary conditions, such as sarcasm, under which these benefits weaken or disappear. Study 3, through focus group discussions, reveals the interpretive strategies users employ when reasoning about digital prosody, including drawing meaning from the absence of expected cues and defaulting toward negative interpretations in ambiguous contexts. Together, these studies establish eNVCs as a coherent and measurable class of digital behaviors, refine theoretical accounts of cue richness and interpretive effort, and provide practical tools for affective computing, user modeling, and emotion-aware interface design. The eNVC detection toolkit is available as a Python and R package at https://github.com/kokiljaidka/envc.

We investigate how large language models perform on low-resource languages by benchmarking eight LLMs across five experimental conditions in English, Kazakh, and Mongolian. Using 50 hand-crafted questions spanning factual, reasoning, technical, and culturally grounded categories, we evaluate 2,000 responses on accuracy, fluency, and completeness. We find a consistent performance gap of 13.8-16.7 percentage points between English and low-resource language conditions, with models maintaining surface-level fluency while producing significantly less accurate content. Cross-lingual transfer-prompting models to reason in English before translating back-yields selective gains for bilingual architectures (+2.2pp to +4.3pp) but provides no benefit to English-dominant models. Our results demonstrate that current LLMs systematically underserve low-resource language communities, and that effective mitigation strategies are architecture-dependent rather than universal.

The automotive industry is under growing pressure to reduce its environmental impact, requiring accurate predictive modeling to support sustainable engineering design. This study examines the factors that determine vehicle fuel consumption from the seminal Motor Trend dataset, identifying the governing physical factors of efficiency through rigorous quantitative analysis. Methodologically, the research uses data sanitization, statistical outlier elimination, and in-depth Exploratory Data Analysis (EDA) to curb the occurrence of multicollinearity between powertrain features. A comparative analysis of machine learning paradigms including Multiple Linear Regression, Support Vector Machines (SVM), and Logistic Regression was carried out to assess predictive efficacy. Findings indicate that SVM Regression is most accurate on continuous prediction (R-squared = 0.889, RMSE = 0.326), and is effective in capturing the non-linear relationships between vehicle mass and engine displacement. In parallel, Logistic Regression proved superior for classification (Accuracy = 90.8%) and showed exceptional recall (0.957) when identifying low-efficiency vehicles. These results challenge the current trend toward black-box deep learning architectures for static physical datasets, providing validation of robust performance by interpretable and well-tuned classical models. The research finds that intrinsic vehicle efficiency is fundamentally determined by physical design parameters, weight and displacement, offering a data-driven framework for how manufacturers should focus on lightweighting and engine downsizing to achieve stringent global sustainability goals.

Geotechnical site characterisation relies on sparse, heterogeneous borehole data where uncertainty quantification and model interpretability are as critical as predictive accuracy for reliable engineering decisions. This paper presents an exploratory investigation into the use of TabPFN, a transformer-based tabular foundation model using in-context learning, and its extension library tabpfn-extensions for two geotechnical inference tasks: (1) soil-type classification using N-value and shear-wave velocity data from a synthetic geotechnical dataset, and (2) iterative imputation of five missing mechanical parameters ($s_\mathrm{u}$, $E_{\mathrm{u}}$, ${σ'}_\mathrm{p}$, $C_\mathrm{c}$, $C_\mathrm{v}$) in benchmark problem BM/AirportSoilProperties/2/2025. We apply cosine-similarity analysis to TabPFN-derived embeddings, visualise full posterior distributions from an iterative inference procedure, and compute SHAP-based feature importance, all without model retraining. Learned embeddings clearly separate Clay and Sand samples without explicit soil-type supervision; iterative imputation improves predictions for four of five target parameters, with posterior widths that reflect physically reasonable parameter-specific uncertainty; and SHAP analysis reveals the inter-parameter dependency structure, recovering established geotechnical relationships including the Skempton compression index correlation and the inverse dependence of preconsolidation pressure on water content. These results suggest the potential of foundation-model-based tools to support interpretable, uncertainty-aware parameter inference in data-scarce geotechnical practice.

Indoor localization systems in care facilities enable optimization of staff allocation, workload management, and quality of care delivery. Traditional machine learning approaches to Bluetooth Low Energy (BLE)-based localization treat each temporal measurement as an independent observation, fundamentally limiting their performance. To address this limitation, this paper introduces Deep Attention-based Sequential Ensemble Learning (DASEL), a novel framework that reconceptualizes indoor localization as a sequential learning problem. The framework integrates frequency-based feature engineering, bidirectional GRU networks with attention mechanisms, multi-directional sliding windows, and confidence-weighted temporal smoothing to capture human movement trajectories. Evaluated on real-world data from a care facility using 4-fold temporal cross-validation, DASEL achieves a macro F1 score of 0.4438, representing a 53.1% improvement over the best traditional baseline (0.2898).

Autonomous driving requires reliable reasoning over fine-grained 3D scene facts. Fine-grained question answering over multi-modal driving observations provides a natural way to evaluate this capability, yet existing perception pipelines and driving-oriented large language model (LLM) methods still suffer from unreliable scene facts, hallucinations, opaque reasoning, and heavy reliance on task-specific training. We present KLDrive, the first knowledge-graph-augmented LLM reasoning framework for fine-grained question answering in autonomous driving. KLDrive addresses this problem through designing two tightly coupled components: an energy-based scene fact construction module that consolidates multi-source evidence into a reliable scene knowledge graph, and an LLM agent that performs fact-grounded reasoning over a constrained action space under explicit structural constraints. By combining structured prompting with few-shot in-context exemplars, the framework adapts to diverse reasoning tasks without heavy task-specific fine-tuning. Experiments on two large-scale autonomous-driving QA benchmarks show that KLDrive outperforms prior state-of-the-art methods, achieving the best overall accuracy of 65.04% on NuScenes-QA and the best SPICE score of 42.45 on GVQA. On counting, the most challenging factual reasoning task, it improves over the strongest baseline by 46.01 percentage points, demonstrating substantially reduced hallucinations and the benefit of coupling reliable scene fact construction with explicit reasoning.

With the rapid evolution of Large Language Model (LLM) agent ecosystems, centralized skill marketplaces have emerged as pivotal infrastructure for augmenting agent capabilities. However, these marketplaces face unprecedented security challenges, primarily stemming from semantic-behavioral inconsistency and inter-skill combinatorial risks, where individually benign skills induce malicious behaviors during collaborative invocation. To address these vulnerabilities, we propose SkillProbe, a multi-stage security auditing framework driven by multi-agent collaboration. SkillProbe introduces a "Skills-for-Skills" design paradigm, encapsulating auditing processes into standardized skill modules to drive specialized agents through a rigorous pipeline, including admission filtering, semantic-behavioral alignment detection, and combinatorial risk simulation. We conducted a large-scale evaluation using 8 mainstream LLM series across 2,500 real-world skills from ClawHub. Our results reveal a striking popularity-security paradox, where download volume is not a reliable proxy for security quality, as over 90% of high-popularity skills failed to pass rigorous auditing. Crucially, we discovered that high-risk skills form a single giant connected component within the risk-link dimension, demonstrating that cascaded risks are systemic rather than isolated occurrences. We hope that SkillProbe will inspire researchers to provide a scalable governance infrastructure for constructing a trustworthy Agentic Web. SkillProbe is accessible for public experience at skillhub.holosai.io.

Large language models (LLMs) used for multiple-choice and pairwise evaluation tasks often exhibit selection bias due to non-semantic factors like option positions and label symbols. Existing inference-time debiasing is costly and may harm reasoning, while pointwise training ignores that the same question should yield consistent answers across permutations. To address this issue, we propose Permutation-Aware Group Relative Policy Optimization (PA-GRPO), which mitigates selection bias by enforcing permutation-consistent semantic reasoning. PA-GRPO constructs a permutation group for each instance by generating multiple candidate permutations, and optimizes the model using two complementary mechanisms: (1) cross-permutation advantage, which computes advantages relative to the mean reward over all permutations of the same instance, and (2) consistency-aware reward, which encourages the model to produce consistent decisions across different permutations. Experimental results demonstrate that PA-GRPO outperforms strong baselines across seven benchmarks, substantially reducing selection bias while maintaining high overall performance. The code will be made available on Github (https://github.com/ECNU-Text-Computing/PA-GRPO).

Mechanistic interpretability seeks to understand how Large Language Models (LLMs) represent and process information. Recent approaches based on dictionary learning and transcoders enable representing model computation in terms of sparse, interpretable features and their interactions, giving rise to feature attribution graphs. However, these graphs are often large and redundant, limiting their interpretability in practice. Cross-Layer Transcoders (CLTs) address this issue by sharing features across layers while preserving layer-specific decoding, yielding more compact representations, but remain difficult to train and analyze at scale. We introduce an open-source library for end-to-end training and interpretability of CLTs. Our framework integrates scalable distributed training with model sharding and compressed activation caching, a unified automated interpretability pipeline for feature analysis and explanation, attribution graph computation using Circuit-Tracer, and a flexible visualization interface. This provides a practical and unified solution for scaling CLT-based mechanistic interpretability. Our code is available at: https://github.com/LLM-Interp/CLT-Forge.

We identify a routing paradox in hybrid recurrent-attention architectures: content-based routing - deciding which tokens deserve expensive attention - requires exactly the pairwise computation that routing is designed to avoid. Through 20+ controlled experiments across three tasks (a synthetic diagnostic, the Zoology MQAR benchmark, and HotpotQA), we map the routing landscape exhaustively. One layer of softmax attention creates a latent ~34-dimensional subspace enabling 98.4% routing precision; zero layers yield 1.2%. This subspace is invisible to cosine similarity, destroyed by random projections (98.4% to 2.6%), and cannot be created by contrastive pretraining - proving attention's role is writing pairwise match results into representations, not merely computing them. Twelve alternative mechanisms all cluster at 15-29%. Non-learned indices (Bloom filter: 90.9%; BM25 on HotpotQA: 82.7%) bypass the bottleneck entirely. The result is a sharp two-regime hierarchy with an empty middle ground. These findings provide the mechanistic explanation for the empirical observation that recurrent models fail at associative recall, and reframe attention as a representation constructor rather than merely a computation mechanism.

In shared autonomy, a critical tension arises when an automated assistant must choose between obeying a human's instruction and deliberately overriding it to prevent harm. This safety-critical behavior is known as intelligent disobedience. To formalize this dynamic, this paper introduces the Intelligent Disobedience Game (IDG), a sequential game-theoretic framework based on Stackelberg games that models the interaction between a human leader and an assistive follower operating under asymmetric information. It characterizes optimal strategies for both agents across multi-step scenarios, identifying strategic phenomena such as ``safety traps,'' where the system indefinitely avoids harm but fails to achieve the human's goal. The IDG provides a needed mathematical foundation that enables both the algorithmic development of agents that can learn safe non-compliance and the empirical study of how humans perceive and trust disobedient AI. The paper further translates the IDG into a shared control Multi-Agent Markov Decision Process representation, forming a compact computational testbed for training reinforcement learning agents.

This study addresses an important gap in time series outlier detection by proposing a novel problem setting: long-term outlier prediction. Conventional methods primarily focus on immediate detection by identifying deviations from normal patterns. As a result, their applicability is limited when forecasting outlier events far into the future. To overcome this limitation, we propose a simple and unsupervised two-layer method that is independent of specific models. The first layer performs standard outlier detection, and the second layer predicts future outlier scores based on the temporal structure of previously observed outliers. This framework enables not only pointwise detection but also long-term forecasting of outlier likelihoods. Experiments on synthetic datasets show that the proposed method performs well in both detection and prediction tasks. These findings suggest that the method can serve as a strong baseline for future work in outlier detection and forecasting.

A single matrix out of 468 in GPT-2 Small can increase perplexity by 20,000x when compressed, revealing that transformer compression sensitivity spans five orders of magnitude. We map this sensitivity landscape across five architectures (117M-8B parameters), finding a consistent hierarchy: early-layer MLP up-projections are catastrophically sensitive while value projections compress nearly for free. This hierarchy is stable across compression levels, evaluation scales (2K-51K tokens), and datasets (WikiText-103, C4). Using Lyapunov stability theory, we show that residual connections contract compression errors by growing the hidden state faster than the error. Error contraction is necessary but not sufficient for compression tolerance: architecture-specific redundancy plays an equally important role, as demonstrated by the hybrid LFM2-2.6B degrading only 7x despite higher amplification than the fully-contracting GPT-2 Small (120x). Ten machine-checked Lean 4 theorems formalize per-matrix error bounds with no sorry markers; all bounds produce zero violations across 14,040+ configurations. We validate with downstream task evaluation (HellaSwag, ARC-Easy, Winogrande), activation-aware pruning on two architectures, and a Compression Fragility Index that rank-orders model robustness.

Hard negatives play a critical role in training and fine-tuning dense retrieval models, as they are semantically similar to positive documents yet non-relevant, and correctly distinguishing them is essential for improving retrieval accuracy. However, identifying effective hard negatives typically requires extensive ablation studies involving repeated fine-tuning with different negative sampling strategies and hyperparameters, resulting in substantial computational cost. In this paper, we introduce ECI: Effective Contrastive Information , a theoretically grounded metric grounded in Information Theory and Information Retrieval principles that enables practitioners to assess the quality of hard negatives prior to model fine-tuning. ECI evaluates negatives by optimizing the trade-off between Information Capacity the logarithmic bound on mutual information determined by set size and Discriminative Efficiency, a harmonic balance of Signal Magnitude (Hardness) and Safety (Max-Margin). Unlike heuristic approaches, ECI strictly penalizes unsafe, false-positive negatives prevalent in generative methods. We evaluate ECI across hard-negative sets mined or generated using BM25, cross-encoders, and large language models. Our results demonstrate that ECI accurately predicts downstream retrieval performance, identifying that hybrid strategies (BM25+Cross-Encoder) offer the optimal balance of volume and reliability, significantly reducing the need for costly end-to-end ablation studies.

The cognitive sciences aim to understand intelligence by formalizing underlying operations as computational models. Traditionally, this follows a cycle of discovery where researchers develop paradigms, collect data, and test predefined model classes. However, this manual pipeline is fundamentally constrained by the slow pace of human intervention and a search space limited by researchers' background and intuition. Here, we propose a paradigm shift toward a fully automated, in silico science of the mind that implements every stage of the discovery cycle using Large Language Models (LLMs). In this framework, experimental paradigms exploring conceptually meaningful task structures are directly sampled from an LLM. High-fidelity behavioral data are then simulated using foundation models of cognition. The tedious step of handcrafting cognitive models is replaced by LLM-based program synthesis, which performs a high-throughput search over a vast landscape of algorithmic hypotheses. Finally, the discovery loop is closed by optimizing for ''interestingness'', a metric of conceptual yield evaluated by an LLM-critic. By enabling a fast and scalable approach to theory development, this automated loop functions as a high-throughput in-silico discovery engine, surfacing informative experiments and mechanisms for subsequent validation in real human populations.

Recent theoretical models of diffusion processes, conceptualized as coupled Ornstein-Uhlenbeck systems, predict a hierarchy of interaction timescales, and consequently, the existence of a synchronization gap between modes that commit at different stages of the reverse process. However, because these predictions rely on continuous time and analytically tractable score functions, it remains unclear how this phenomenology manifests in the deep, discrete architectures deployed in practice. In this work, we investigate how the synchronization gap is mechanistically realized within pretrained Diffusion Transformers (DiTs). We construct an explicit architectural realization of replica coupling by embedding two generative trajectories into a joint token sequence, modulated by a symmetric cross attention gate with variable coupling strength g. Through a linearized analysis of the attention difference, we show that the replica interaction decomposes mechanistically. We empirically validate our theoretical framework on a pretrained DiT-XL/2 model by tracking commitment and per layer internal mode energies. Our results reveal that: (1) the synchronization gap is an intrinsic architectural property of DiTs that persists even when external coupling is turned off; (2) as predicted by our spatial routing bounds, the gap completely collapses under strong coupling; (3) the gap is strictly depth localized, emerging sharply only within the final layers of the Transformer; and (4) global, low frequency structures consistently commit before local, high frequency details. Ultimately, our findings provide a mechanistic interpretation of how Diffusion Transformers resolve generative ambiguity, isolating speciation transitions to the terminal layers of the network.

Multiphysics simulation frameworks such as MOOSE provide rigorous engines for phase-field materials modeling, yet adoption is constrained by the expertise required to construct valid input files, coordinate parameter sweeps, diagnose failures, and extract quantitative results. We introduce AutoMOOSE, an open-source agentic framework that orchestrates the full simulation lifecycle from a single natural-language prompt. AutoMOOSE deploys a five-agent pipeline in which the Input Writer coordinates six sub-agents and the Reviewer autonomously corrects runtime failures without user intervention. A modular plugin architecture enables new phase-field formulations without modifying the core framework, and a Model Context Protocol (MCP) server exposes the workflow as ten structured tools for interoperability with any MCP-compatible client. Validated on a four-temperature copper grain growth benchmark, AutoMOOSE generates MOOSE input files with 6 of 12 structural blocks matching a human expert reference exactly and 4 functionally equivalent, executes all runs in parallel with a 1.8x speedup, and performs an end-to-end physical consistency check spanning intent, finite-element execution, and Arrhenius kinetics with no human verification. Grain coarsening kinetics are recovered with R^2 = 0.90-0.95 at T >= 600 K; the recovered activation energy Q_fit = 0.296 eV is consistent with a human-written reference (Q_fit = 0.267 eV) under identical parameters. Three runtime failure classes were diagnosed and resolved autonomously within a single correction cycle, and every run produces a provenance record satisfying FAIR data principles. These results show that the gap between knowing the physics and executing a validated simulation campaign can be bridged by a lightweight multi-agent orchestration layer, providing a pathway toward AI-driven materials discovery and self-driving laboratories.

Biophysical neural system simulations are among the most computationally demanding scientific applications, and their optimization requires navigating high-dimensional parameter spaces under numerous constraints that impose a binary feasible/infeasible partition with no gradient signal to guide the search. Here, we introduce DMOSOPT, a scalable optimization framework that leverages a unified, jointly learned surrogate model to capture the interplay between objectives, constraints, and parameter sensitivities. By learning a smooth approximation of both the objective landscape and the feasibility boundary, the joint surrogate provides a unified gradient that simultaneously steers the search toward improved objective values and greater constraint satisfaction, while its partial derivatives yield per-parameter sensitivity estimates that enable more targeted exploration. We validate the framework from single-cell dynamics to population-level network activity, spanning incremental stages of a neural circuit modeling workflow, and demonstrate efficient, effective optimization of highly constrained problems at supercomputing scale with substantially fewer problem evaluations. While motivated by and demonstrated in the context of computational neuroscience, the framework is general and applicable to constrained multi-objective optimization problems across scientific and engineering domains.

Unmanned Aerial Vehicles (UAVs) are valuable for mission-critical systems like surveillance, rescue, or delivery. Not surprisingly, such systems attract cyberattacks, including Denial-of-Service (DoS) attacks to overwhelm the resources of mission drones (MDs). How can we defend UAV mission systems against DoS attacks? We adopt cyber deception as a defense strategy, in which honey drones (HDs) are proposed to bait and divert attacks. The attack and deceptive defense hinge upon radio signal strength: The attacker selects victim MDs based on their signals, and HDs attract the attacker from afar by emitting stronger signals, despite this reducing battery life. We formulate an optimization problem for the attacker and defender to identify their respective strategies for maximizing mission performance while minimizing energy consumption. To address this problem, we propose a novel approach, called HT-DRL. HT-DRL identifies optimal solutions without a long learning convergence time by taking the solutions of hypergame theory into the neural network of deep reinforcement learning. This achieves a systematic way to intelligently deceive attackers. We analyze the performance of diverse defense mechanisms under different attack strategies. Further, the HT-DRL-based HD approach outperforms existing non-HD counterparts up to two times better in mission performance while incurring low energy consumption.

Time-varying causal models provide a powerful framework for studying dynamic scientific systems, yet most existing approaches assume that the underlying causal network is known a priori - an assumption rarely satisfied in real-world domains where causal structure is uncertain, evolving, or only indirectly observable. This limits the applicability of dynamic causal inference in many scientific settings. We propose Dynamic Causal Network Autoregression (DCNAR), a two-stage neural causal modeling framework that integrates data-driven causal discovery with time-varying causal inference. In the first stage, a neural autoregressive causal discovery model learns a sparse directed causal network from multivariate time series. In the second stage, this learned structure is used as a structural prior for a time-varying neural network autoregression, enabling dynamic estimation of causal influence without requiring pre-specified network structure. We evaluate the scientific validity of DCNAR using behavioral diagnostics that assess causal necessity, temporal stability, and sensitivity to structural change, rather than predictive accuracy alone. Experiments on multi-country panel time-series data demonstrate that learned causal networks yield more stable and behaviorally meaningful dynamic causal inferences than coefficient-based or structure-free alternatives, even when forecasting performance is comparable. These results position DCNAR as a general framework for using AI as a scientific instrument for dynamic causal reasoning under structural uncertainty.

Large language models (LLMs) are increasingly deployed in human-AI teams as support agents for complex tasks such as information retrieval, programming, and decision-making assistance. While these agents' autonomy and contextual knowledge enables them to be useful, it also exposes them to a broad range of attacks, including data poisoning, prompt injection, and even prompt engineering. Through these attack vectors, malicious actors can manipulate an LLM agent to provide harmful information, potentially manipulating human agents to make harmful decisions. While prior work has focused on LLMs as attack targets or adversarial actors, this paper studies their potential role as defensive supervisors within mixed human-AI teams. Using a dataset consisting of multi-party conversations and decisions for a real human-AI team over a 25 round horizon, we formulate the problem of malicious behavior detection from interaction traces. We find that LLMs are capable of identifying malicious behavior in real-time, and without task-specific information, indicating the potential for task-agnostic defense. Moreover, we find that the malicious behavior of interest is not easily identified using simple heuristics, further suggesting the introduction of LLM defenders could render human teams more robust to certain classes of attack.

Multi-agent LLM systems, where multiple prompted instances of a language model independently answer questions, are increasingly used for complex reasoning tasks. However, existing methods for quantifying the uncertainty of their collective outputs rely on shallow voting statistics that discard the rich semantic information in agents' reasoning. We introduce DiscoUQ, a framework that extracts and leverages the structure of inter-agent disagreement -- both linguistic properties (evidence overlap, argument strength, divergence depth) and embedding geometry (cluster distances, dispersion, cohesion) -- to produce well-calibrated confidence estimates. We propose three methods of increasing complexity: DiscoUQ-LLM (logistic regression on LLM-extracted structure features), DiscoUQ-Embed (logistic regression on embedding geometry), and DiscoUQ-Learn (a neural network combining all features). Evaluated on four diverse benchmarks (StrategyQA, MMLU, TruthfulQA, ARC-Challenge) with a 5-agent system using Qwen3.5-27B, DiscoUQ-LLM achieves an average AUROC of 0.802, outperforming the best baseline (LLM Aggregator, 0.791) while being substantially better calibrated (ECE 0.036 vs. 0.098). The learned features generalize across benchmarks with near-zero performance degradation and provide the largest improvements where they are most needed: in the ambiguous "weak disagreement" tier where simple vote counting fails.

Transformer-based language models excel at in-context learning (ICL), where they can adapt to new tasks based on contextual examples, without parameter updates. In a specific form of ICL, which we refer to as \textit{contextual recall}, models pretrained on open-ended text leverage pairwise examples to recall specific facts in novel prompt formats. We investigate whether contextual recall emerges from pretraining alone, what finetuning is required, and what mechanisms drive the necessary representations. For this, we introduce a controlled synthetic framework where pretraining sequences consist of subject-grammar-attribute tuples, with attribute types tied to grammar statistics. We demonstrate that while such pretraining successfully yields factual knowledge, it is insufficient for contextual recall: models fail to implicitly infer attribute types when the grammar statistics are removed in ICL prompts. However, we show that finetuning on tasks requiring implicit inference, distinct from the ICL evaluation, using a subset of subjects, triggers the emergence of contextual recall across all subjects. This transition is accompanied by the formation of low-dimensional latent encodings of the shared attribute type. For mechanistic insight, we derive a construction for an attention-only transformer that replicates the transition from factual to contextual recall, corroborated by empirical validation.

One of the most common machine learning setups is logistic regression. In many classification models, including neural networks, the final prediction is obtained by applying a logistic link function to a linear score. In binary logistic regression, the feedback can be either soft labels, corresponding to the true conditional probability of the data (as in distillation), or sampled hard labels (taking values $\pm 1$). We point out a fundamental problem that arises even in a particularly favorable setting, where the goal is to learn a noise-free soft target of the form $σ(\mathbf{x}^{\top}\mathbf{w}^{\star})$. In the over-constrained case (i.e. the number of samples $n$ exceeds the input dimension $d$) with examples $(\mathbf{x}_i,σ(\mathbf{x}_i^{\top}\mathbf{w}^{\star}))$, it is sufficient to recover $\mathbf{w}^{\star}$ and hence achieve the Bayes risk. However, we prove that when the examples are labeled by hard labels $y_i$ sampled from the same conditional distribution $σ(\mathbf{x}_i^{\top}\mathbf{w}^{\star})$ and $\mathbf{w}^{\star}$ is $s$-sparse, then rotation-invariant algorithms are provably suboptimal: they incur an excess risk $Ω\!\left(\frac{d-1}{n}\right)$, while there are simple non-rotation invariant algorithms with excess risk $O(\frac{s\log d}{n})$. The simplest rotation invariant algorithm is gradient descent on the logistic loss (with early stopping). A simple non-rotation-invariant algorithm for sparse targets that achieves the above upper bounds uses gradient descent on the weights $u_i,v_i$, where now the linear weight $w_i$ is reparameterized as $u_iv_i$.

Sketching is widely used in randomized linear algebra for low-rank matrix approximation, column subset selection, and many other problems, and it has gained significant traction in machine learning applications. However, sketching large matrices often necessitates distributed memory algorithms, where communication overhead becomes a critical bottleneck on modern supercomputing clusters. Despite its growing relevance, distributed-memory parallel strategies for sketching remain largely unexplored. In this work, we establish communication lower bounds for sketching using dense matrices that determine how much data movement is required to perform it in parallel. One important observation of our lower bounds is that no communication is required for a small number of processors. We show that our lower bounds are tight by presenting communication optimal algorithms. Furthermore, we extend our approach to determine communication lower bounds for computations of Nyström approximation where sketching is applied twice. We also introduce novel parallel algorithms whose communication costs are close to the lower bounds. Finally, we implement our algorithms on modern state-of-the-art supercomputing infrastructures which have both CPU- and GPU-equipped systems and demonstrate their parallel scalability.

This paper studies whether a lightweight trained aggregator can combine diverse zero-shot large language model judgments into a stronger downstream signal for corporate disclosure classification. Zero-shot LLMs can read disclosures without task-specific fine-tuning, but their predictions often vary across prompts, reasoning styles, and model families. I address this problem with a multi-agent framework in which three zero-shot agents independently read each disclosure and output a sentiment label, a confidence score, and a short rationale. A logistic meta-classifier then aggregates these signals to predict next-day stock return direction. I use a sample of 18,420 U.S. corporate disclosures issued by Nasdaq and S&P 500 firms between 2018 and 2024, matched to next-day stock returns. Results show that the trained aggregator outperforms all single agents, majority vote, confidence-weighted voting, and a FinBERT baseline. Balanced accuracy rises from 0.561 for the best single agent to 0.612 for the trained aggregator, with the largest gains in disclosures combining strong current performance with weak guidance or elevated risk. The results suggest that zero-shot LLM agents capture complementary financial signals and that supervised aggregation can turn cross-agent disagreement into a more useful classification target.

We present a comparative study of methods for generating realistic, constrained small- to medium-scale road networks with built-in redundancy. In this research, we evaluate the proposed Evolutionary Algorithm (EA) with connectivity and redundancy constraints against the Wave Function Collapse (WFC) method - commonly used in procedural terrain generation for games - and swarm algorithms: Particle Swarm (PSO) and Gray Wolf (GWO). Our focus is on producing realistic, redundant road networks suitable for vision, localization and navigation problems. We evaluate metrics: connectivity, cycles, intersections, dead ends, graph cut-edges while enforcing physical plausibility. We propose an EA and its extended version with elitism via MAP-Elites method. We detail the implementation, constraints, metrics and provide both visual and quantitative comparisons with baselines. Results highlight how fitness function design choices affect the structural characteristics of generated networks and highlight the impact of specific constraints in practical applications. Our contribution is a method for creating realistic synthetic datasets from sparse tile definitions derived from real-world data. We demonstrate a practical application by generating realistic maps using a laboratory-collected tileset from a Duckietown city model. Our approach performs coherent geometric transformations on metadata, in this work exemplified by semantic segmentation masks of the generated road networks.

Frontier LLM companies have repeatedly assured courts and regulators that their models do not store copies of training data. They further rely on safety alignment strategies via RLHF, system prompts, and output filters to block verbatim regurgitation of copyrighted works, and have cited the efficacy of these measures in their legal defenses against copyright infringement claims. We show that finetuning bypasses these protections: by training models to expand plot summaries into full text, a task naturally suited for commercial writing assistants, we cause GPT-4o, Gemini-2.5-Pro, and DeepSeek-V3.1 to reproduce up to 85-90% of held-out copyrighted books, with single verbatim spans exceeding 460 words, using only semantic descriptions as prompts and no actual book text. This extraction generalizes across authors: finetuning exclusively on Haruki Murakami's novels unlocks verbatim recall of copyrighted books from over 30 unrelated authors. The effect is not specific to any training author or corpus: random author pairs and public-domain finetuning data produce comparable extraction, while finetuning on synthetic text yields near-zero extraction, indicating that finetuning on individual authors' works reactivates latent memorization from pretraining. Three models from different providers memorize the same books in the same regions ($r \ge 0.90$), pointing to an industry-wide vulnerability. Our findings offer compelling evidence that model weights store copies of copyrighted works and that the security failures that manifest after finetuning on individual authors' works undermine a key premise of recent fair use rulings, where courts have conditioned favorable outcomes on the adequacy of measures preventing reproduction of protected expression.

The Predictive Associative Memory (PAM) framework posits that useful relationships often connect items that co-occur in shared contexts rather than items that appear similar in embedding space. A contrastive MLP trained on co-occurrence annotations--Contrastive Association Learning (CAL)--has improved multi-hop passage retrieval and discovered narrative function at corpus scale in text. We test whether this principle transfers to molecular biology, where protein-protein interactions provide functional associations distinct from gene expression similarity. Four experiments across two biological domains map the operating envelope. On gene perturbation data (Replogle K562 CRISPRi, 2,285 genes), CAL trained on STRING protein interactions achieves cross-boundary AUC of 0.908 where expression similarity scores 0.518. A second gene dataset (DepMap, 17,725 genes) confirms the result after negative sampling correction, reaching cross-boundary AUC of 0.947. Two drug sensitivity experiments produce informative negatives that sharpen boundary conditions. Three cross-domain findings emerge: (1) inductive transfer succeeds in biology--a node-disjoint split with unseen genes yields AUC 0.826 (Delta +0.127)--where it fails in text (+/-0.10), suggesting physically grounded associations are more transferable than contingent co-occurrences; (2) CAL scores anti-correlate with interaction degree (Spearman r = -0.590), with gains concentrating on understudied genes with focused interaction profiles; (3) tighter association quality outperforms larger but noisier training sets, reversing the text pattern. Results are stable across training seeds (SD < 0.001) and cross-boundary threshold choices.

AI agents today have passwords but no permission slips. They execute tool calls (fund transfers, database queries, shell commands, sub-agent delegation) with no standard mechanism to enforce authorization before the action executes. Current safety architectures rely on model alignment (probabilistic, training-time) and post-hoc evaluation (retrospective, batch). Neither provides deterministic, policy-based enforcement at the individual tool call level. We characterize this gap as the pre-action authorization problem and present the Open Agent Passport (OAP), an open specification and reference implementation that intercepts tool calls synchronously before execution, evaluates them against a declarative policy, and produces a cryptographically signed audit record. OAP enforces authorization decisions in a measured median of 53 ms (N=1,000). In a live adversarial testbed (4,437 authorization decisions across 1,151 sessions, $5,000 bounty), social engineering succeeded against the model 74.6% of the time under a permissive policy; under a restrictive OAP policy, a comparable population of attackers achieved a 0% success rate across 879 attempts. We distinguish pre-action authorization from sandboxed execution (contains blast radius but does not prevent unauthorized actions) and model-based screening (probabilistic), and show they are complementary. The same infrastructure that enforces security constraints (spending limits, capability scoping) also enforces quality gates, operational contracts, and compliance controls. The specification is released under Apache 2.0 (DOI: 10.5281/zenodo.18901596).

gUFO is a lightweight implementation of the Unified Foundational Ontology (UFO) suitable for Semantic Web OWL 2 DL applications. UFO is a mature foundational ontology with a rich axiomatization and that has been employed in a significant number of projects in research and industry. Moreover, it is currently in the process of standardization by the International Organization for Standardization as the ISO/IEC CD 21838-5. gUFO stands out from other foundational ontology implementations (such as those provided for BFO and DOLCE) given its unique support for a typology of types (operationalizing OntoClean guidelines), its reification patterns for intrinsic and relational aspects, and its support for situations and high-order types. gUFO provides well-founded patterns to address recurrent problems in Semantic Web knowledge graphs. In this paper, we present gUFO with its constituting categories, relations and constraints, discuss how it differs from the original UFO reference ontology, elaborate on its community adoption, and systematically position it in relation to existing OWL-based implementations of popular alternative foundational ontologies.

Effectively leveraging the vast computational resources of modern cloud environments requires expertise spanning multiple technical domains: configuring scientific software with correct parameters and dependencies, navigating thousands of provider-specific instance types and pricing options, and managing parallel or distributed execution. We conduct a study indicating that the absence of these categories of expertise poses an ongoing challenge to unlocking the potential of cloud-enabled computational science. To address this challenge, we introduce Adviser, an intuitive multi-cloud platform centered on a workflow abstraction. Workflows are reusable, expert-crafted artifacts encapsulating environment setup, data processing, simulation, result capture, and visualization steps needed to execute scientific and ML applications. This approach allows users to specify high-level intent, while Adviser handles resource provisioning, runtime configuration, and data movement. Using two computational glaciology codes, Icepack and PISM, we show how to use Adviser to gain scientific insight and perform rapid exploration of cost-performance tradeoffs and scaling behavior without specialized expertise in cloud or high-performance computing.

Large language models are increasingly used as personal assistants, yet most lack a persistent user model, forcing users to repeatedly restate preferences across sessions. We propose Vector-Adapted Retrieval Scoring (VARS), a pipeline-agnostic, frozen-backbone framework that represents each user with long-term and short-term vectors in a shared preference space and uses these vectors to bias retrieval scoring over structured preference memory. The vectors are updated online from weak scalar rewards from users' feedback, enabling personalization without per-user fine-tuning. We evaluate on \textsc{MultiSessionCollab}, an online multi-session collaboration benchmark with rich user preference profiles, across math and code tasks. Under frozen backbones, the main benefit of user-aware retrieval is improved interaction efficiency rather than large gains in raw task accuracy: our full VARS agent achieves the strongest overall performance, matches a strong Reflection baseline in task success, and reduces timeout rate and user effort. The learned long-term vectors also align with cross-user preference overlap, while short-term vectors capture session-specific adaptation, supporting the interpretability of the dual-vector design. Code, model, and data are available at https://github.com/YurenHao0426/VARS.

Selecting the most relevant or informative features is a key issue in actual machine learning problems. Since an exhaustive search is not feasible even for a moderate number of features, an intelligent search strategy must be employed for finding an optimal subset, which implies considering how features interact with each other in promoting class separability. Balancing feature subset size and classification accuracy constitutes a multi-objective optimization challenge. Here we propose MOELIGA, a multi-objective genetic algorithm incorporating an evolutionary local improvement strategy that evolves subordinate populations to refine feature subsets. MOELIGA employs a crowding-based fitness sharing mechanism and a sigmoid transformation to enhance diversity and guide compactness, alongside a geometry-based objective promoting classifier independence. Experimental evaluation on 14 diverse datasets demonstrates MOELIGA's ability to identify smaller feature subsets with superior or comparable classification performance relative to 11 state-of-the-art methods. These findings suggest MOELIGA effectively addresses the accuracy-dimensionality trade-off, offering a robust and adaptable approach for multi-objective feature selection in complex, high-dimensional scenarios.

Large Language Model (LLM) agents combine the chat interaction capabilities of LLMs with the power to interact with external tools and APIs. This enables them to perform complex tasks and act autonomously to achieve user goals. However, current agent systems operate on an all-or-nothing basis: an agent either has full access to an API's capabilities and a web page's content, or it has no access at all. This coarse-grained approach forces users to trust agents with more capabilities than they actually need for a given task. In this paper, we introduce AC4A, an access control framework for agents. As agents become more capable and autonomous, users need a way to limit what APIs or portions of web pages these agents can access, eliminating the need to trust them with everything an API or web page allows. Our goal with AC4A is to provide a framework for defining permissions that lets agents access only the resources they are authorized to access. AC4A works across both API-based and browser-based agents. It does not prescribe what permissions should be, but offers a flexible way to define and enforce them, making it practical for real-world systems. AC4A works by creating permissions granting access to resources, drawing inspiration from established access control frameworks like the one for the Unix file system. Applications define their resources as hierarchies and provide a way to compute the necessary permissions at runtime needed for successful resource access. We demonstrate the usefulness of AC4A in enforcing permissions over real-world APIs and web pages through case studies. The source code of AC4A is available at https://github.com/reSHARMA/AC4A

Feature selection is fundamental to robust data-centric AI, but most existing methods optimize predictive performance under a single data distribution. This often selects spurious features that fail under distribution shifts. Motivated by principles from causal invariance, we study feature selection from a stability perspective and introduce Causally-Guided Diffusion for Stable Feature Selection (CGDFS). In CGDFS, we formalized feature selection as approximate posterior inference over feature subsets, whose posterior mass favors low prediction error and low cross-environment variance. Our framework combines three key insights: First, we formulate feature selection as stability-aware posterior sampling. Here, causal invariance serves as a soft inductive bias rather than explicit causal discovery. Second, we train a diffusion model as a learned prior over plausible continuous selection masks, combined with a stability-aware likelihood that rewards invariance across environments. This diffusion prior captures structural dependencies among features and enables scalable exploration of the combinatorially large selection space. Third, we perform guided annealed Langevin sampling that combines the diffusion prior with the stability objective, which yields a tractable, uncertainty-aware posterior inference that avoids discrete optimization and produces robust feature selections. We evaluate CGDFS on open-source real-world datasets exhibiting distribution shifts. Across both classification and regression tasks, CGDFS consistently selects more stable and transferable feature subsets, which leads to improved out-of-distribution performance and greater selection robustness compared to sparsity-based, tree-based, and stability-selection baselines.

We highlight a striking difference in behavior between two widely used variants of coordinate ascent variational inference: the sequential and parallel algorithms. While such differences were known in the numerical analysis literature in simpler settings, they remain largely unexplored in the optimization-focused literature on variational inference in more complex models. Focusing on the moderately high-dimensional linear regression problem, we show that the sequential algorithm, although typically slower, enjoys convergence guarantees under more relaxed conditions than the parallel variant, which is often employed to facilitate block-wise updates and improve computational efficiency.

Physical AI agents, such as robots and other embodied systems operating under tight and fluctuating resource constraints, remain far less capable than biological agents in open-ended real-world environments. This paper argues that Active Inference (AIF), grounded in the Free Energy Principle, offers a principled foundation for closing that gap. We develop this argument from first principles, following a chain from probability theory through Bayesian machine learning and variational inference to active inference and reactive message passing. From the FEP perspective, systems that maintain their structural and functional integrity over time can, under suitable assumptions, be described as minimizing variational free energy (VFE), and AIF operationalizes this by unifying perception, learning, planning, and control within a single computational objective. We show that VFE minimization is naturally realized by reactive message passing on factor graphs, where inference emerges from local, parallel computations. This realization is well matched to the constraints of physical operation, including hard deadlines, asynchronous data, fluctuating power budgets, and changing environments. Because reactive message passing is event-driven, interruptible, and locally adaptable, performance degrades gracefully under reduced resources while model structure can adjust online. We further show that, under suitable coupling and coarse-graining conditions, coupled AIF agents can be described as higher-level AIF agents, yielding a homogeneous architecture based on the same message-passing primitive across scales. Our contribution is not empirical benchmarking, but a clear theoretical and architectural case for the engineering community.

Modern multi-access 5G+ networks provide mobile terminals with additional capacity, improving network stability and performance. However, in highly mobile environments such as vehicular networks, supporting multi-access connectivity remains challenging. The rapid fluctuations of wireless link quality often outpace the responsiveness of existing multipath schedulers and transport-layer protocols. This paper addresses this challenge by integrating Transformer-based path state forecasting with a new multipath splitting scheduler called Deep Adaptive Rate Allocation (DARA). The proposed scheduler employs a deep reinforcement learning engine to dynamically compute optimal congestion window fractions on available paths, determining data allocation among them. A six-component normalised reward function with weight-mediated conflict resolution drives a DQN policy that eliminates the observation-reaction lag inherent in reactive schedulers. Performance evaluation uses a Mininet-based Multipath Datagram Congestion Control Protocol testbed with traces from mobile users in vehicular environments. Experimental results demonstrate that DARA achieves better file transfer time reductions compared to learning-based schedulers under moderate-volatility traces. For buffered video streaming, resolution improvements are maintained across all tested conditions. Under controlled burst scenarios with sub-second buffer constraints, DARA achieves substantial rebuffering improvements whilst state-of-the-art schedulers exhibit near-continuous stalling.

As agentic systems move into real-world deployments, their decisions increasingly depend on external inputs such as retrieved content, tool outputs, and information provided by other actors. When these inputs can be strategically shaped by adversaries, the relevant security risk extends beyond a fixed library of prompt attacks to adaptive strategies that steer agents toward unfavorable outcomes. We propose profit-driven red teaming, a stress-testing protocol that replaces handcrafted attacks with a learned opponent trained to maximize its profit using only scalar outcome feedback. The protocol requires no LLM-as-judge scoring, attack labels, or attack taxonomy, and is designed for structured settings with auditable outcomes. We instantiate it in a lean arena of four canonical economic interactions, which provide a controlled testbed for adaptive exploitability. In controlled experiments, agents that appear strong against static baselines become consistently exploitable under profit-optimized pressure, and the learned opponent discovers probing, anchoring, and deceptive commitments without explicit instruction. We then distill exploit episodes into concise prompt rules for the agent, which make most previously observed failures ineffective and substantially improve target performance. These results suggest that profit-driven red-team data can provide a practical route to improving robustness in structured agent settings with auditable outcomes.

Foundation models in healthcare have largely adopted self supervised pretraining objectives inherited from natural language processing and computer vision, emphasizing reconstruction and large scale representation learning prior to downstream adaptation. We revisit this paradigm in outcome centric clinical prediction settings and argue that, when high quality supervision is available, direct outcome alignment may provide a stronger inductive bias than generative pretraining. We propose a supervised deep learning framework that explicitly shapes representation geometry by maximizing inter class separation relative to within class variance, thereby concentrating model capacity along clinically meaningful axes. Across multiple longitudinal electronic health record tasks, including mortality and readmission prediction, our approach consistently outperforms masked, autoregressive, and contrastive pretraining baselines under matched model capacity. The proposed method improves discrimination, calibration, and sample efficiency, while simplifying the training pipeline to a single stage optimization. These findings suggest that in low entropy, outcome driven healthcare domains, supervision can act as the statistically optimal driver of representation learning, challenging the assumption that large scale self supervised pretraining is a prerequisite for strong clinical performance.

The exponential growth in data has intensified the demand for computational power to train large-scale deep learning models. However, the rapid growth in model size and complexity raises concerns about equal and fair access to computational resources, particularly under increasing energy and infrastructure constraints. GPUs have emerged as essential for accelerating such workloads. This study benchmarks four deep learning models (Conv6, VGG16, ResNet18, CycleGAN) using TensorFlow and PyTorch on Intel Xeon CPUs and NVIDIA Tesla T4 GPUs. Our experiments demonstrate that, on average, GPU training achieves speedups ranging from 11x to 246x depending on model complexity, with lightweight models (Conv6) showing the highest acceleration (246x), mid-sized models (VGG16, ResNet18) achieving 51-116x speedups, and complex generative models (CycleGAN) reaching 11x improvements compared to CPU training. Additionally, in our PyTorch vs. TensorFlow comparison, we observed that TensorFlow's kernel-fusion optimizations reduce inference latency by approximately 15%. We also analyze GPU memory usage trends and projecting requirements through 2025 using polynomial regression. Our findings highlight that while GPUs are essential for sustaining AI's growth, democratized and shared access to GPU resources is critical for enabling research innovation across institutions with limited computational budgets.

Interpreting complex machine learning models is a critical challenge, especially for tabular data where model transparency is paramount. Local Interpretable Model-Agnostic Explanations (LIME) has been a very popular framework for interpretable machine learning, also inspiring many extensions. While traditional surrogate models used in LIME variants (e.g. linear regression and decision trees) offer a degree of stability, they can struggle to faithfully capture the complex non-linear decision boundaries that are inherent in many sophisticated black-box models. This work contributes toward bridging the gap between high predictive performance and interpretable decision-making. Specifically, we propose the NDT-LIME variant that integrates Neural Decision Trees (NDTs) as surrogate models. By leveraging the structured, hierarchical nature of NDTs, our approach aims at providing more accurate and meaningful local explanations. We evaluate its effectiveness on several benchmark tabular datasets, showing consistent improvements in explanation fidelity over traditional LIME surrogates.

Large language models make agent-based simulation more behaviorally expressive, but they also sharpen a basic methodological tension: fluent, human-like output is not, by itself, evidence for theory. We evaluate what an LLM-driven simulation can credibly support using information engagement on social media as a test case. In a Weibo-like environment, we manipulate information load and descriptive norms, while allowing popularity cues (cumulative likes and Sina Weibo-style cumulative reshares) to evolve endogenously. We then ask whether simulated behavior changes in theoretically interpretable ways under these controlled variations, rather than merely producing plausible-looking traces. Engagement responds systematically to information load and descriptive norms, and sensitivity to popularity cues varies across contexts, indicating conditionality rather than rigid prompt compliance. We discuss methodological implications for simulation-based communication research, including multi-condition stress tests, explicit no-norm baselines because default prompts are not blank controls, and design choices that preserve endogenous feedback loops when studying bandwagon dynamics.

Discovering the governing equations of dynamical systems is a central problem across many scientific disciplines. As experimental data become increasingly available, automated equation discovery methods offer a promising data-driven approach to accelerate scientific discovery. Among these methods, genetic programming (GP) has been widely adopted due to its flexibility and interpretability. However, GP-based approaches often suffer from inefficient exploration of the symbolic search space, leading to slow convergence and suboptimal solutions. To address these limitations, we propose LLM-ODE, a large language model-aided model discovery framework that guides symbolic evolution using patterns extracted from elite candidate equations. By leveraging the generative prior of large language models, LLM-ODE produces more informed search trajectories while preserving the exploratory strengths of evolutionary algorithms. Empirical results on 91 dynamical systems show that LLM-ODE variants consistently outperform classical GP methods in terms of search efficiency and Pareto-front quality. Overall, our results demonstrate that LLM-ODE improves both efficiency and accuracy over traditional GP-based discovery and offers greater scalability to higher-dimensional systems compared to linear and Transformer-only model discovery methods.

Uncertainty quantification for image data is dominated by complex deep learning methods, yet the field lacks an interpretable, mathematically grounded baseline. We propose Bayesian scattering to fill this gap, serving as a first-step baseline akin to the role of Bayesian linear regression for tabular data. Our method couples the wavelet scattering transform-a deep, non-learned feature extractor-with a simple probabilistic head. Because scattering features are derived from geometric principles rather than learned, they avoid overfitting the training distribution. This helps provide sensible uncertainty estimates even under significant distribution shifts. We validate this on diverse tasks, including medical imaging under institution shift, wealth mapping under country-to-country shift, and Bayesian optimization of molecular properties. Our results suggest that Bayesian scattering is a solid baseline for complex uncertainty quantification methods.

As users increasingly turn to LLMs for practical and personal advice, they become vulnerable to being subtly steered toward hidden incentives misaligned with their own interests. Prior works have benchmarked persuasion and manipulation detection, but these efforts rely on simulated or debate-style settings, remain uncorrelated with real human belief shifts, and overlook a critical dimension: the morality of hidden incentives driving the manipulation. We introduce PUPPET, a theoretical taxonomy of personalized emotional manipulation in LLM-human dialogues that centers around incentive morality, and conduct a human study with N=1,035 participants across realistic everyday queries, varying personalization and incentive direction (harmful versus prosocial). We find that harmful hidden incentives produce significantly larger belief shifts than prosocial ones. Finally, we benchmark LLMs on the task of belief prediction, finding that models exhibit moderate predictive ability of belief change based on conversational contexts (r=0.3 - 0.5), but they also systematically underestimate the magnitude of belief shift. Together, this work establishes a theoretically grounded and behaviorally validated foundation for studying, and ultimately combatting, incentive-driven manipulation in LLMs during everyday, practical user queries.

Large language models exhibit strong reasoning capabilities, yet often rely on shortcuts such as surface pattern matching and answer memorization rather than genuine logical inference. We propose Shortcut-Aware Reasoning Training (SART), a gradient-aware framework that detects and mitigates shortcut-promoting samples via ShortcutScore and gradient surgery. Our method identifies shortcut signals through gradient misalignment with validation objectives and answer-token concentration, and modifies training dynamics accordingly. Experiments on controlled reasoning benchmarks show that SART achieves +16.5% accuracy and +40.2% robustness over the strongest baseline, significantly improving generalization under distribution shifts. Code is available at: https://github.com/fuyanjie/short-cut-aware-data-centric-reasoning.

Online Continual Learning (OCL) for image classification represents a challenging subset of Continual Learning, focusing on classifying images from a stream without assuming data independence and identical distribution (i.i.d). The primary challenge in this context is to prevent catastrophic forgetting, where the model's performance on previous tasks deteriorates as it learns new ones. Although various strategies have been proposed to address this issue, achieving rapid convergence remains a significant challenge in the online setting. In this work, we introduce a novel approach to training OCL models that utilizes the Natural Gradient Descent optimizer, incorporating an approximation of the Fisher Information Matrix (FIM) through Kronecker Factored Approximate Curvature (KFAC). This method demonstrates substantial improvements in performance across all OCL methods, particularly when combined with existing OCL tricks, on datasets such as Split CIFAR-100, CORE50, and Split miniImageNet.

The strong and continuous increase of AI-based services leads to the steady proliferation of AI data centres worldwide with the unavoidable escalation of their power consumption. It is unknown how this energy demand for computational purposes will impact the surrounding environment. Here, we focus our attention on the heat dissipation of AI hyperscalers. Taking advantage of land surface temperature measurements acquired by remote sensing platforms over the last decades, we are able to obtain a robust assessment of the temperature increase recorded in the areas surrounding AI data centres globally. We estimate that the land surface temperature increases by 2°C on average after the start of operations of an AI data centre, inducing local microclimate zones, which we call the data heat island effect. We assess the impact on the communities, quantifying that more than 340 million people could be affected by this temperature increase. Our results show that the data heat island effect could have a remarkable influence on communities and regional welfare in the future, hence becoming part of the conversation around environmentally sustainable AI worldwide.

Hyper-Connections (HC) generalize residual connections into multiple streams, employing residual matrices for cross-stream feature mixing to enrich model expressivity. However, unconstrained mixing disrupts the identity mapping property intrinsic to the residual connection, causing unstable training. To address this, Manifold-Constrained Hyper-Connections (mHC) and its variant restrict these matrices to the Birkhoff polytope (doubly stochastic matrices) via Sinkhorn iterations or permutation-based parameterizations. We reveal three limitations of this polytope constraint: (1) identity degeneration, where learned matrices collapse around the identity and diminish cross-stream interactions, (2) an expressivity bottleneck, as the non-negativity constraint prevents subtractive feature disentanglement, and (3) parameterization inefficiencies, manifesting as unstable Sinkhorn iterations or the factorial-scaling overhead of permutation-based parameterizations. To overcome these flaws, we propose Spectral-Sphere-Constrained Hyper-Connections (sHC). By geometrically shifting the feasible set from a rigid polytope to a spectral norm sphere, sHC allows negative entries, unlocking subtractive interactions for selective feature diversification. This shift eliminates unstable Sinkhorn projections and factorial parameterization, enabling expressive, non-degenerate residual matrices while preserving training stability.

LLMs often share comparable benchmark accuracies, but their complementary performance across task subsets suggests that an Oracle router--a theoretical selector with perfect foresight--can significantly surpass standalone model accuracy by navigating model-specific strengths. While current routers rely on fragile semantic signals, we propose using internal prefill activations via Encoder-Target Decoupling--a functional separation between the model providing the predictive signal (the Encoder) and the model whose performance is being estimated (the Target). This allows optimized heterogeneous pairing between unique encoders and target models. We utilize Fisher Separability (J) and Effective Dimensionality (d_eff) as mathematical probes to isolate optimal layer-wise signals, providing the predictive foundation for our SharedTrunkNet architecture. SharedTrunkNet captures up to 45.58% of the accuracy gap between the strongest standalone model and the Oracle while achieving 74.31% cost savings relative to the highest-cost model.

Data assimilation algorithms estimate the state of a dynamical system from partial observations, where the successful performance of these algorithms hinges on costly parameter tuning and on employing an accurate model for the dynamics. This paper introduces a framework for jointly learning the state, dynamics, and parameters of filtering algorithms in data assimilation through a process we refer to as auto-differentiable filtering. The framework leverages a theoretically motivated loss function that enables learning from partial, noisy observations via gradient-based optimization using auto-differentiation. We further demonstrate how several well-known data assimilation methods can be learned or tuned within this framework. To underscore the versatility of auto-differentiable filtering, we perform experiments on dynamical systems spanning multiple scientific domains, such as the Clohessy-Wiltshire equations from aerospace engineering, the Lorenz-96 system from atmospheric science, and the generalized Lotka-Volterra equations from systems biology. Finally, we provide guidelines for practitioners to customize our framework according to their observation model, accuracy requirements, and computational budget.

Two distinct technologies have gained attention lately due to their prospects for motor rehabilitation: robotics and brain-machine interfaces (BMIs). Harnessing their combined efforts is a largely uncharted and promising direction that has immense clinical potential. However, a significant challenge is whether motor intentions from the user can be accurately detected using non-invasive BMIs in the presence of instrumental noise and passive movements induced by the rehabilitation exoskeleton. As an alternative to the straightforward continuous control approach, this study instead aims to characterize the onset and offset of motor imagery during passive arm movements induced by an upper-body exoskeleton to allow for the natural control (initiation and termination) of functional movements. Ten participants were recruited to perform kinesthetic motor imagery (MI) of the right arm while attached to the robot, simultaneously cued with LEDs indicating the initiation and termination of a goal-oriented reaching task. Using electroencephalogram signals, we built a decoder to detect the transition between i) rest and beginning MI and ii) maintaining and ending MI. Offline decoder evaluation achieved group average onset accuracy of 60.7% and 66.6% for offset accuracy, revealing that the start and stop of MI could be identified while attached to the robot. Furthermore, pseudo-online evaluation could replicate this performance, forecasting reliable online exoskeleton control in the future. Our approach showed that participants could produce quality and reliable sensorimotor rhythms regardless of noise or passive arm movements induced by wearing the exoskeleton, which opens new possibilities for BMI control of assistive devices.

The exponential growth of academic publications has led to a surge in papers of varying quality, increasing the cost of paper screening. Current approaches either use novelty assessment within general AI Reviewers or repurpose DeepResearch, which lacks domain-specific mechanisms and thus delivers lower-quality results. To bridge this gap, we introduce NoveltyAgent, a multi-agent system designed to generate comprehensive and faithful novelty reports, enabling thorough evaluation of a paper's originality. It decomposes manuscripts into discrete novelty points for fine-grained retrieval and comparison, and builds a comprehensive related-paper database while cross-referencing claims to ensure faithfulness. Furthermore, to address the challenge of evaluating such open-ended generation tasks, we propose a checklist-based evaluation framework, providing an unbiased paradigm for building reliable evaluations. Extensive experiments show that NoveltyAgent achieves state-of-the-art performance, outperforming GPT-5 DeepResearch by 10.15%. We hope this system will provide reliable, high-quality novelty analysis and help researchers quickly identify novel papers. Code and demo are available at https://github.com/SStan1/NoveltyAgent.

Large language models (LLMs) are increasingly evaluated and sometimes trained using automated graders such as LLM-as-judges that output scalar scores or preferences. While convenient, these approaches are often opaque: a single score rarely explains why an answer is good or bad, which requirements were missed, or how a system should be improved. This lack of interpretability limits their usefulness for model development, dataset curation, and high-stakes deployment. Query-specific rubric-based evaluation offers a more transparent alternative by decomposing quality into explicit, checkable criteria. However, manually designing high-quality, query-specific rubrics is labor-intensive and cognitively demanding and not feasible for deployment. While previous approaches have focused on generating intermediate rubrics for automated downstream evaluation, it is unclear if these rubrics are both interpretable and effective for human users. In this work, we investigate whether LLMs can generate useful, instance-specific rubrics as compared to human-authored rubrics, while also improving effectiveness for identifying good responses. Through our systematic study on two rubric benchmarks, and on multiple few-shot and post-training strategies, we find that off-the-shelf LLMs produce rubrics that are poorly aligned with human-authored ones. We introduce a simple strategy, RubricRAG, which retrieves domain knowledge via rubrics at inference time from related queries. We demonstrate that RubricRAG can generate more interpretable rubrics both for similarity to human-authored rubrics, and for improved downstream evaluation effectiveness. Our results highlight both the challenges and a promising approach of scalable, interpretable evaluation through automated rubric generation.

Federated learning (FL) is a communication-efficient collaborative learning framework that enables model training across multiple agents with private local datasets. While the benefits of FL in improving global model performance are well established, individual agents may behave strategically, balancing the learning payoff against the cost of contributing their local data. Motivated by the need for FL frameworks that successfully retain participating agents, we propose an incentive-aware federated averaging method in which, at each communication round, clients transmit both their local model parameters and their updated training dataset sizes to the server. The dataset sizes are dynamically adjusted via a Nash equilibrium (NE)-seeking update rule that captures strategic data participation. We analyze the proposed method under convex and nonconvex global objective settings and establish performance guarantees for the resulting incentive-aware FL algorithm. Numerical experiments on the MNIST and CIFAR-10 datasets demonstrate that agents achieve competitive global model performance while converging to stable data participation strategies.

Financial time-series forecasting in real-world high-frequency markets is often hindered by delayed or partially stale observations caused by asynchronous data acquisition and transmission latency. To better reflect such practical conditions, we investigate a simulated delay setting where a portion of historical signals is corrupted by a Zero-Order Hold (ZOH) mechanism, significantly increasing forecasting difficulty through stepwise stagnation artifacts. In this paper, we propose ReLaMix (Residual Latency-Aware Mixing Network), a lightweight extension of TimeMixer that integrates learnable bottleneck compression with residual refinement for robust signal recovery under delayed observations. ReLaMix explicitly suppresses redundancy from repeated stale values while preserving informative market dynamics via residual mixing enhancement. Experiments on a large-scale second-resolution PAXGUSDT benchmark demonstrate that ReLaMix consistently achieves state-of-the-art accuracy across multiple delay ratios and prediction horizons, outperforming strong mixer and Transformer baselines with substantially fewer parameters. Moreover, additional evaluations on BTCUSDT confirm the cross-asset generalization ability of the proposed framework. These results highlight the effectiveness of residual bottleneck mixing for high-frequency financial forecasting under realistic latency-induced staleness.

Recent interpretability work often treats a feature as a single global direction, dictionary atom, or latent coordinate shared across contexts. We argue that this ontology can fail in obstructed representation spaces, where locally coherent meanings need not assemble into one globally consistent feature. We introduce an atlas-native replacement object, the semantic section: a transport-compatible family of local feature representatives defined over a context atlas. We formalize semantic sections, prove that tree-supported propagation is always pathwise realizable, and show that cycle consistency is the key criterion for genuine globalization. This yields a distinction between tree-local, globalizable, and twisted sections, with twisted sections capturing locally coherent but holonomy-obstructed meanings. We then develop a discovery-and-certification pipeline based on seeded propagation, synchronization across overlaps, defect-based pruning, cycle-aware taxonomy, and deduplication. Across layer-16 atlases for Llama 3.2 3B Instruct, Qwen 2.5 3B Instruct, and Gemma 2 2B IT, we find nontrivial populations of semantic sections, including cycle-supported globalizable and twisted regimes after deduplication. Most importantly, semantic identity is not recovered by raw global-vector similarity. Even certified globalizable sections show low cross-chart signed cosine similarity, and raw similarity baselines recover only a small fraction of true within-section pairs, often collapsing at moderate thresholds. By contrast, section-based identity recovery is perfect on certified supports. These results support semantic sections as a better feature ontology in obstructed regimes.

We establish two structural results for Moore homology of ample groupoids. First, for every ample groupoid $\mathcal{G}$ and every discrete abelian coefficient group $A$, we prove a universal coefficient theorem relating the homology groups $H_n(\mathcal{G};A)$ to the integral Moore homology of $\mathcal{G}$. More precisely, we obtain a natural short exact sequence $$ 0 \longrightarrow H_n(\mathcal{G};\mathbb{Z})\otimes_{\mathbb{Z}} A \xrightarrow{κ_n^{\mathcal{G}}} H_n(\mathcal{G};A) \xrightarrow{ι_n^{\mathcal{G}}} \operatorname{Tor}_1^{\mathbb{Z}}\bigl(H_{n-1}(\mathcal{G};\mathbb{Z}),A\bigr) \longrightarrow 0. $$ Second, for a decomposition of the unit space into clopen saturated subsets, we prove a Mayer-Vietoris long exact sequence in Moore homology. The proof is carried out at the chain level and is based on a short exact sequence of Moore chain complexes associated to the corresponding restricted groupoids. These results provide effective tools for the computation of Moore homology. We also explain why the discreteness of the coefficient group is essential for the universal coefficient theorem.

Transfer learning with models pretrained on ImageNet has become a standard practice in computer vision. Transfer learning refers to fine-tuning pretrained weights of a neural network on a downstream task, typically unrelated to ImageNet. However, pretrained weights can become saturated and may yield insignificant gradients, failing to adapt to the downstream task. This hinders the ability of the model to train effectively, and is commonly referred to as loss of neural plasticity. Loss of plasticity may prevent the model from fully adapting to the target domain, especially when the downstream dataset is atypical in nature. While this issue has been widely explored in continual learning, it remains relatively understudied in the context of transfer learning. In this work, we propose the use of a targeted weight re-initialization strategy to restore neural plasticity prior to fine-tuning. Our experiments show that both convolutional neural networks (CNNs) and vision transformers (ViTs) benefit from this approach, yielding higher test accuracy with faster convergence on several image classification benchmarks. Our method introduces negligible computational overhead and is compatible with common transfer learning pipelines.

Automating white blood cell classification for diagnosis of leukaemia is a promising alternative to time-consuming and resource-intensive examination of cells by expert pathologists. However, designing robust algorithms for classification of rare cell types remains challenging due to variations in staining, scanning and inter-patient heterogeneity. We propose a lightweight ensemble approach for classification of cells during Haematopoiesis, with a focus on the biology of Granulopoiesis, Monocytopoiesis and Lymphopoiesis. Through dataset expansion to alleviate some class imbalance, we demonstrate that a simple ensemble of lightweight pretrained SwinV2-Tiny, DinoBloom-Small and ConvNeXT-V2-Tiny models achieves excellent performance on this challenging dataset. We train 3 instantiations of each architecture in a stratified 3-fold cross-validation framework; for an input image, we forward-pass through all 9 models and aggregate through logit averaging. We further reason on the weaknesses of our model in confusing similar-looking myelocytes in granulopoiesis and lymphocytes in lymphopoiesis. Code: https://gitlab.com/siddharthsrivastava/wbc-bench-2026.

Kazakh, a Turkic language spoken by over 22 million people, remains underserved by existing multilingual language models, which allocate minimal capacity to low-resource languages and employ tokenizers ill-suited to agglutinative morphology. We present SozKZ, a family of Llama-architecture language models (50M-600M parameters) trained entirely from scratch on 9 billion tokens of Kazakh text with a dedicated 50K BPE tokenizer. We evaluate all models on three Kazakh benchmarks -- multiple-choice cultural QA, reading comprehension (Belebele), and topic classification (SIB-200) -- alongside five multilingual baselines ranging from 500M to 3B parameters. Our 600M model achieves 30.3% accuracy on Kazakh cultural QA, approaching the 32.0% of Llama-3.2-1B (2x larger), and 25.5% on SIB-200 topic classification, surpassing all evaluated multilingual models up to 2B parameters. We observe consistent scaling from 50M to 600M, with MC QA accuracy rising from 22.8% to 30.3%, suggesting that further scaling remains beneficial. These results demonstrate that small, dedicated models trained from scratch with a language-appropriate tokenizer offer a viable path for low-resource language technology, achieving competitive performance at a fraction of the computational cost. All models and the tokenizer are released under open licenses.

ChatGPT has demonstrated remarkable capabilities on both poetry generation and translation, yet its ability to truly understand poetry remains unexplored. Previous poetry-related work merely analyzed experimental outcomes without addressing fundamental issues of comprehension. This paper introduces a comprehensive framework for evaluating ChatGPT's understanding of modern poetry. We collaborated with professional poets to evaluate ChatGPT's interpretation of modern Chinese poems by different poets along multiple dimensions. Evaluation results show that ChatGPT's interpretations align with the original poets' intents in over 73% of the cases. However, its understanding in certain dimensions, particularly in capturing poeticity, proved to be less satisfactory. These findings highlight the effectiveness and necessity of our proposed framework. This study not only evaluates ChatGPT's ability to understand modern poetry but also establishes a solid foundation for future research on LLMs and their application to poetry-related tasks.

The rapid integration of Large Language Models (LLMs) into software development workflows has given rise to a new class of AI-assisted coding tools, such as Claude-Code, Codex, and Gemini CLIs. While promising significant productivity gains, the engineering process of building these tools, which sit at the complex intersection of traditional software engineering, AI system design, and human-computer interaction, is fraught with unique and poorly understood challenges. This paper presents the first empirical study of engineering pitfalls in building such tools, on a systematic, manual analysis of over 3.8K publicly reported bugs in the open-source repositories of three AI-assisted coding tools (i.e., Claude-Code, Codex, and Gemini CLIs) on GitHub. Specifically, we employ an open-coding methodology to manually examine the issue description, associated user discussions, and developer responses. Through this process, we categorize each bug along multiple dimensions, including bug type, bug location, root cause, and observed symptoms. This fine-grained annotation enables us to characterize common failure patterns and identify recurring engineering challenges. Our results show that more than 67% of the bugs in these tools are related to functionality. In terms of root causes, 36.9% of the bugs stem from API, integration, or configuration errors. Consequently, the most commonly observed symptoms reported by users are API errors (18.3%), terminal problems (14%), and command failures (12.7%). These bugs predominantly affect the tool invocation (37.2%) and command execution (24.7%) stages of the system workflow. Collectively, our findings provide a critical roadmap for developers seeking to design the next generation of reliable and robust AI coding assistants.

Long-context language modeling is commonly framed as a scalability challenge of token-level attention, yet local-to-global information structuring remains largely implicit in existing approaches. Drawing on cognitive theories of discourse comprehension, we propose HiCI (Hierarchical Construction--Integration), a hierarchical attention module that constructs segment-level representations, integrates them into a shared global context, and broadcasts both to condition segment-level attention. We validate HiCI through parameter-efficient adaptation of LLaMA-2 with only <5.5% additional parameters, extending context from 4K to 100K tokens (7B) and 64K tokens (13B). Across language modeling, retrieval, and instruction-following benchmarks, HiCI yields consistent improvements over strong baselines, including matching proprietary models on topic retrieval and surpassing GPT-3.5-Turbo-16K on code comprehension. These results demonstrate the effectiveness of explicit hierarchical structuring as an inductive bias for long-context modeling.

Causal discovery has been widely studied, yet many existing methods rely on strong assumptions or fall into two extremes: either depending on costly interventional signals or partial ground truth as strong priors, or adopting purely data driven paradigms with limited guidance, which hinders practical deployment. Motivated by real-world scenarios where only coarse domain knowledge is available, we propose a knowledge-informed pretrained model for causal discovery that integrates weak prior knowledge as a principled middle ground. Our model adopts a dual source encoder-decoder architecture to process observational data in a knowledge-informed way. We design a diverse pretraining dataset and a curriculum learning strategy that smoothly adapts the model to varying prior strengths across mechanisms, graph densities, and variable scales. Extensive experiments on in-distribution, out-of distribution, and real-world datasets demonstrate consistent improvements over existing baselines, with strong robustness and practical applicability.

Pairwise comparison labeling is emerging as it yields higher inter-rater reliability than conventional classification labeling, but exhaustive comparisons require quadratic cost. We propose Dodgersort, which leverages CLIP-based hierarchical pre-ordering, a neural ranking head and probabilistic ensemble (Elo, BTL, GP), epistemic--aleatoric uncertainty decomposition, and information-theoretic pair selection. It reduces human comparisons while improving the reliability of the rankings. In visual ranking tasks in medical imaging, historical dating, and aesthetics, Dodgersort achieves a 11--16\% annotation reduction while improving inter-rater reliability. Cross-domain ablations across four datasets show that neural adaptation and ensemble uncertainty are key to this gain. In FG-NET with ground-truth ages, the framework extracts 5--20$\times$ more ranking information per comparison than baselines, yielding Pareto-optimal accuracy--efficiency trade-offs.

RAW images captured by different camera sensors exhibit substantial domain shifts due to varying spectral responses, noise characteristics, and tone behaviors, complicating their direct use in downstream computer vision tasks. Prior methods address this problem by training domain-specific RAW-to-RAW translators for each source-target pair, but such approaches do not scale to real-world scenarios involving multiple types of commercial cameras. In this work, we introduce MERIT, the first unified framework for multi-domain RAW image translation, which leverages a single model to perform translations across arbitrary camera domains. To address domain-specific noise discrepancies, we propose a sensor-aware noise modeling loss that explicitly aligns the signal-dependent noise statistics of the generated images with those of the target domain. We further enhance the generator with a conditional multi-scale large kernel attention module for improved context and sensor-aware feature modeling. To facilitate standardized evaluation, we introduce MDRAW, the first dataset tailored for multi-domain RAW image translation, comprising both paired and unpaired RAW captures from five diverse camera sensors across a wide range of scenes. Extensive experiments demonstrate that MERIT outperforms prior models in both quality (5.56 dB improvement) and scalability (80% reduction in training iterations).

As AI agent ecosystems grow, agents need mechanisms to monitor relevant knowledge in real time. Semantic publish-subscribe systems address this by matching new content against vector subscriptions. However, in multi-agent settings where agents operate under different data handling policies, unrestricted semantic subscriptions create policy violations: agents receive notifications about content they are not authorized to access. We introduce governance-aware vector subscriptions, a mechanism that composes semantic similarity matching with multi-dimensional policy predicates grounded in regulatory frameworks (EU DSM Directive, EU AI Act). The policy predicate operates over multiple independent dimensions (processing level, direct marketing restrictions, training opt-out, jurisdiction, and scientific usage) each with distinct legal bases. Agents subscribe to semantic regions of a curated knowledge base; notifications are dispatched only for validated content that passes both the similarity threshold and all applicable policy constraints. We formalize the mechanism, implement it within AIngram (an operational multi-agent knowledge base), and evaluate it using the PASA benchmark. We validate the mechanism on a synthetic corpus (1,000 chunks, 93 subscriptions, 5 domains): the governed mode correctly enforces all policy constraints while preserving delivery of authorized content. Ablation across five policy dimensions shows that no single dimension suffices for full compliance.

We study verification (decision) problems for graph properties in distributed networks under the locally checkable labeling framework, where nodes use labels (proofs) and local neighborhoods to decide acceptance or rejection. Our focus is twofold. First, we study cycle detection. While it is known that this can be verified using 3 labels with access to the 1-hop neighborhood, we introduce a novel gadget that encodes direction along a path using only 2 labels and access to a 3-hop neighborhood. This yields a cycle-detection labeling scheme with just 2 labels and may be of independent interest. Second, we consider adversarially corrupted labelings, where each node has access to a local neighborhood within which a fraction of nodes may receive erroneous labels. We introduce a general algorithmic framework, called refix, that transforms a base verification algorithm for a property P operating on labels within a d-hop neighborhood into one that tolerates up to i erroneous labels within a radius d+2i, by accessing a d+2i-hop neighborhood. We demonstrate applications to cycle detection, cycle absence, and bipartiteness, and provide lower bounds relating the number of errors to the required neighborhood size.

Attributed Graph Clustering (AGC) is a fundamental unsupervised task that partitions nodes into cohesive groups by jointly modeling structural topology and node attributes. While the advent of graph neural networks and self-supervised learning has catalyzed a proliferation of AGC methodologies, a widening chasm persists between academic benchmark performance and the stringent demands of real-world industrial deployment. To bridge this gap, this survey provides a comprehensive, industrially grounded review of AGC from three complementary perspectives. First, we introduce the Encode-Cluster-Optimize taxonomic framework, which decomposes the diverse algorithmic landscape into three orthogonal, composable modules: representation encoding, cluster projection, and optimization strategy. This unified paradigm enables principled architectural comparisons and inspires novel methodological combinations. Second, we critically examine prevailing evaluation protocols to expose the field's academic monoculture: a pervasive over-reliance on small, homophilous citation networks, the inadequacy of supervised-only metrics for an inherently unsupervised task, and the chronic neglect of computational scalability. In response, we advocate for a holistic evaluation standard that integrates supervised semantic alignment, unsupervised structural integrity, and rigorous efficiency profiling. Third, we explicitly confront the practical realities of industrial deployment. By analyzing operational constraints such as massive scale, severe heterophily, and tabular feature noise alongside extensive empirical evidence from our companion benchmark, we outline actionable engineering strategies. Furthermore, we chart a clear roadmap for future research, prioritizing heterophily-robust encoders, scalable joint optimization, and unsupervised model selection criteria to meet production-grade requirements.

Contextual recommendation is a variant of contextual linear bandits in which the learner observes an (optimal) action rather than a reward scalar. Recently, Sakaue et al. (2025) developed an efficient Online Newton Step (ONS) approach with an $O(d\log T)$ regret bound, where $d$ is the dimension of the action space and $T$ is the time horizon. In this paper, we present a simple algorithm that is more efficient than the ONS-based method while achieving the same regret guarantee. Our core idea is to exploit the improperness inherent in contextual recommendation, leading to an update rule akin to the second-order perceptron from online classification. This removes the Mahalanobis projection step required by ONS, which is often a major computational bottleneck. More importantly, the same algorithm remains robust to possibly suboptimal action feedback, whereas the prior ONS-based method required running multiple ONS learners with different learning rates for this extension. We describe how our method works in general Hilbert spaces (e.g., via kernelization), where eliminating Mahalanobis projections becomes even more beneficial.

Recent advances in general-purpose foundation models have stimulated the development of large biological sequence models. While natural language shows symbolic granularity (characters, words, sentences), biological sequences exhibit hierarchical granularity whose levels (nucleotides, amino acids, protein domains, genes) further encode biologically functional information. In this paper, we investigate the integration of cross-granularity knowledge from models through a case study of BiGCARP, a Pfam domain-level model for biosynthetic gene clusters, and ESM, an amino acid-level protein language model. Using representation analysis tools and a set of probe tasks, we first explain why a straightforward cross-model embedding initialization fails to improve downstream performance in BiGCARP, and show that deeper-layer embeddings capture a more contextual and faithful representation of the model's learned knowledge. Furthermore, we demonstrate that representations at different granularities encode complementary biological knowledge, and that combining them yields measurable performance gains in intermediate-level prediction tasks. Our findings highlight cross-granularity integration as a promising strategy for improving both the performance and interpretability of biological foundation models.

Compound AI is a distributed intelligence approach that represents a unified system orchestrating specialized AI/ML models with engineered software components into AI workflows. Compound AI production deployments must satisfy accuracy, latency, and cost objectives under varying loads. However, many deployments operate on fixed infrastructure where horizontal scaling is not viable. Existing approaches optimize solely for accuracy and do not consider changes in workload conditions. We observe that compound AI systems can switch between configurations to fit infrastructure capacity, trading accuracy for latency based on current load. This requires discovering multiple Pareto-optimal configurations from a combinatorial search space and determining when to switch between them at runtime. We present Compass, a novel framework that enables dynamic configuration switching through offline optimization and online adaptation. Compass consists of three components: COMPASS-V algorithm for configuration discovery, Planner for switching policy derivation, and Elastico Controller for runtime adaptation. COMPASS-V discovers accuracy-feasible configurations using finite-difference guided search and a combination of hill-climbing and lateral expansion. Planner profiles these configurations on target hardware and derives switching policies using a queuing theory based model. Elastico monitors queue depth and switches configurations based on derived thresholds. Across two compound AI workflows, COMPASS-V achieves 100% recall while reducing configuration evaluations by 57.5% on average compared to exhaustive search, with efficiency gains reaching 95.3% at tight accuracy thresholds. Runtime adaptation achieves 90-98% SLO compliance under dynamic load patterns, improving SLO compliance by 71.6% over static high-accuracy baselines, while simultaneously improving accuracy by 3-5% over static fast baselines.

This paper studies finite-sample set-membership identification for discrete-time bilinear systems under bounded symmetric log-concave disturbances. Compared with existing finite-sample results for linear systems and related analyses under stronger noise assumptions, we consider the more challenging bilinear setting with trajectory-dependent regressors and allow marginally stable dynamics with polynomial mean-square state growth. Under these conditions, we prove that the diameter of the feasible parameter set shrinks with sample complexity $\widetilde{O}(1/ε)$. Simulation supports the theory and illustrates the advantage of the proposed estimator for uncertainty quantification.

While structure-based relocalizers have long strived for point correspondences when establishing or regressing query-map associations, in this paper, we pioneer the use of planar primitives and 3D planar maps for lightweight 6-DoF camera relocalization in structured environments. Planar primitives, beyond being fundamental entities in projective geometry, also serve as region-based representations that encapsulate both structural and semantic richness. This motivates us to introduce PlanaReLoc, a streamlined plane-centric paradigm where a deep matcher associates planar primitives across the query image and the map within a learned unified embedding space, after which the 6-DoF pose is solved and refined under a robust framework. Through comprehensive experiments on the ScanNet and 12Scenes datasets across hundreds of scenes, our method demonstrates the superiority of planar primitives in facilitating reliable cross-modal structural correspondences and achieving effective camera relocalization without requiring realistically textured/colored maps, pose priors, or per-scene training. The code and data are available at https://github.com/3dv-casia/PlanaReLoc .

The pharmaceutical industry is facing challenges with quality management such as high costs of compliance, slow responses and disjointed knowledge. This paper presents GMPilot, a domain-specific AI agent that is designed to support FDA cGMP compliance. GMPilot is based on a curated knowledge base of regulations and historical inspection observations and uses Retrieval-Augmented Generation (RAG) and Reasoning-Acting (ReAct) frameworks to provide real-time and traceable decision support to the quality professionals. In a simulated inspection scenario, GMPilot shows how it can improve the responsiveness and professionalism of quality professionals by providing structured knowledge retrieval and verifiable regulatory and case-based support. Although GMPilot lacks in the aspect of regulatory scope and model interpretability, it is a viable avenue of improving quality management decision-making in the pharmaceutical sector using intelligent approaches and an example of specialized application of AI in highly regulated sectors.

While Multimodal Large Language Models (MLLMs) excel at vision-language tasks, the cost of their language-driven training on internal visual foundational competence remains unclear. In this paper, we conduct a detailed diagnostic analysis to unveil a pervasive issue: visual representation degradation in MLLMs. Specifically, we find that compared to the initial visual features, the visual representation in the middle layers of LLM exhibits both a degradation in global function and patch structure. We attribute this phenomenon to a visual sacrifice driven by the singular text-generation objective, where the model compromises its visual fidelity to optimize for answer generation. We argue that a robust MLLM requires both strong cross-modal reasoning and core visual competence, and propose Predictive Regularization (PRe) to force degraded intermediate features to predict initial visual features, thereby maintaining the inherent visual attributes of the MLLM's internal representations. Extensive experiments confirm that mitigating this visual degradation effectively boosts vision-language performance, underscoring the critical importance of fostering robust internal visual representations within MLLMs for comprehensive multimodal understanding.

Benchmarks are the de facto standard for tracking progress in large language models (LLMs), yet static test sets can rapidly saturate, become vulnerable to contamination, and are costly to refresh. Scalable evaluation of open-ended items often relies on LLM judges, introducing additional sources of bias and prompt sensitivity. We argue that evaluation must extend beyond how well models answer benchmarks to how well models design them. We introduce BenchBench, a three-stage pipeline and dataset for benchmarking automated benchmark generation: (i) extract structured domain cards from seed benchmarks, (ii) prompt multiple designer LLMs to generate quota-controlled suites, and (iii) validate items with a multi-model answerer panel using exact/numeric/symbolic verifiers when possible and rubric-guided judging otherwise, yielding designer--answerer matrices with item-level quality flags and psychometric diagnostics. Across nine variants spanning computer science, mathematics, medicine, and theory-of-mind reasoning (including multilingual and multimodal settings), we generate 16.7K items, retain ~15K core items post-filtering, and produce ~152K graded model--item responses. BenchBench shows that benchmark-design ability is only moderately correlated with answer-time strength (Spearman rho ~0.37), invalidity is negatively associated with discrimination (Pearson r~0.62), and the resulting designer--answerer matrices enable scalable audits of format/modality/language fidelity and suite-dependent self/family interactions. The project is available at: https://github.com/koanatakiyo/BenchBench.

Neural networks are being increasingly used as heuristics for constraint satisfaction. These neural methods are often recurrent, learning to iteratively refine candidate assignments. In this work, we make explicit the connection between such iterative neural heuristics and Large Neighborhood Search (LNS), and adapt an existing neural constraint satisfaction method-ConsFormer-into an LNS procedure. We decompose the resulting neural LNS into two standard components: the destroy and repair operators. On the destroy side, we instantiate several classical heuristics and introduce novel prediction-guided operators that exploit the model's internal scores to select neighborhoods. On the repair side, we utilize ConsFormer as a neural repair operator and compare the original sampling-based decoder to a greedy decoder that selects the most likely assignments. Through an empirical study on Sudoku, Graph Coloring, and MaxCut, we find that adapting the neural heuristic to an LNS procedure yields substantial gains over its vanilla settings and improves its competitiveness with classical and neural baselines. We further observe consistent design patterns across tasks: stochastic destroy operators outperform greedy ones, while greedy repair is more effective than sampling-based repair for finding a single high-quality feasible assignment. These findings highlight LNS as a useful lens and design framework for structuring and improving iterative neural approaches.

Reinforcement learning from verifiable rewards (RLVR) stimulates the thinking processes of large language models (LLMs), substantially enhancing their reasoning abilities on verifiable tasks. It is often assumed that similar gains should transfer to general question answering (GQA), but this assumption has not been thoroughly validated. To assess whether RLVR automatically improves LLM performance on GQA, we propose a Cross-Generation evaluation framework that measures the quality of intermediate reasoning by feeding the generated thinking context into LLMs of varying capabilities. Our evaluation leads to a discouraging finding: the efficacy of the thinking process on GQA tasks is markedly lower than on verifiable tasks, suggesting that explicit training on GQA remains necessary in addition to training on verifiable tasks. We further observe that direct RL training on GQA is less effective than RLVR. Our hypothesis is that, whereas verifiable tasks demand robust logical chains to obtain high rewards, GQA tasks often admit shortcuts to high rewards without cultivating high-quality thinking. To avoid possible shortcuts, we introduce a simple method, Separated Thinking And Response Training (START), which first trains only the thinking process, using rewards defined on the final answer. We show that START improves both the quality of thinking and the final answer across several GQA benchmarks and RL algorithms.

Large language models (LLMs) are increasingly used as knowledge bases, but keeping them up to date requires targeted knowledge editing (KE). However, it remains unclear how edits are implemented inside the model once applied. In this work, we take a mechanistic view of KE using neuron-level knowledge attribution (NLKA). Unlike prior work that focuses on pre-edit causal tracing and localization, we use post-edit attribution -- contrasting successful and failed edits -- to isolate the computations that shift when an edit succeeds. Across representative KE methods, we find a consistent pattern: mid-to-late attention predominantly promotes the new target, while attention and FFN modules cooperate to suppress the original fact. Motivated by these findings, we propose MEGA, a MEchanism-Guided Activation steering method that performs attention-residual interventions in attribution-aligned regions without modifying model weights. On CounterFact and Popular, MEGA achieves strong editing performance across KE metrics on GPT2-XL and LLaMA2-7B. Overall, our results elevate post-edit attribution from analysis to engineering signal: by pinpointing where and how edits take hold, it powers MEGA to deliver reliable, architecture-agnostic knowledge edits.

Recovering Markov boundary -- the minimal set of variables that maximizes predictive performance for a response variable -- is crucial in many applications. While recent advances improve upon traditional constraint-based techniques by scoring local causal structures, they still rely on nonparametric estimators and heuristic searches, lacking theoretical guarantees for reliability. This paper investigates a framework for efficient Markov boundary discovery by integrating conditional entropy from information theory as a scoring criterion. We design a novel masked autoregressive network to capture complex dependencies. A parallelizable greedy search strategy in polynomial time is proposed, supported by analytical evidence. We also discuss how initializing a graph with learned Markov boundaries accelerates the convergence of causal discovery. Comprehensive evaluations on real-world and synthetic datasets demonstrate the scalability and superior performance of our method in both Markov boundary discovery and causal discovery tasks.

With the rapid development of large language models (LLMs), more and more researchers have paid attention to information extraction based on LLMs. However, there are still some spaces to improve in the existing related methods. First, existing multimodal information extraction (MIE) methods usually employ natural language templates as the input and output of LLMs, which mismatch with the characteristics of information tasks that mostly include structured information such as entities and relations. Second, although a few methods have adopted structured and more IE-friendly code-style templates, they just explored their methods on text-only IE rather than multimodal IE. Moreover, their methods are more complex in design, requiring separate templates to be designed for each task. In this paper, we propose a Code-style Multimodal Information Extraction framework (Code-MIE) which formalizes MIE as unified code understanding and generation. Code-MIE has the following novel designs: (1) Entity attributes such as gender, affiliation are extracted from the text to guide the model to understand the context and role of entities. (2) Images are converted into scene graphs and visual features to incorporate rich visual information into the model. (3) The input template is constructed as a Python function, where entity attributes, scene graphs and raw text compose of the function parameters. In contrast, the output template is formalized as Python dictionaries containing all extraction results such as entities, relations, etc. To evaluate Code-MIE, we conducted extensive experiments on the M$^3$D, Twitter-15, Twitter-17, and MNRE datasets. The results show that our method achieves state-of-the-art performance compared to six competing baseline models, with 61.03\% and 60.49\% on the English and Chinese datasets of M$^3$D, and 76.04\%, 88.07\%, and 73.94\% on the other three datasets.

Robust semantic segmentation is crucial for safe autonomous driving, yet deployed models remain vulnerable to black-box adversarial attacks when target weights are unknown. Most existing approaches either craft image-wide perturbations or optimize patches for a single architecture, which limits their practicality and transferability. We introduce OmniPatch, a training framework for learning a universal adversarial patch that generalizes across images and both ViT and CNN architectures without requiring access to target model parameters.

In personalized marketing, uplift models estimate incremental effects by modeling how customer behavior changes under alternative treatments. However, real-world data often exhibit biases - such as selection bias, spillover effects, and unobserved confounding - which adversely affect both estimation accuracy and metric validity. Despite the importance of bias-aware assessment, a lack of systematic studies persists. To bridge this gap, we design a systematic benchmarking framework. Unlike standard predictive tasks, real-world uplift datasets lack counterfactual ground truth, rendering direct metric validation infeasible. Therefore, a semi-synthetic approach serves as a critical enabler for systematic benchmarking, effectively bridging the gap by retaining real-world feature dependencies while providing the ground truth needed to isolate structural biases. Our investigations reveal that: (i) uplift targeting and prediction can manifest as distinct objectives, where proficiency in one does not ensure efficacy in the other; (ii) while many models exhibit inconsistent performance under diverse biases, TARNet shows notable robustness, providing insights for subsequent model design; (iii) evaluation metric stability is linked to mathematical alignment with the ATE, suggesting that ATE-approximating metrics yield more consistent model rankings under structural data imperfections. These findings suggest the need for more robust uplift models and metrics. Code will be released upon acceptance.

Sensor technologies have evolved to a point where it is now practical to monitor products along the supply chain. The collected data can be stored in a decentralized way using blockchain technology. However, ensuring the reliability of the sensed data is a critical challenge. In other words, we need to trust the data that we write to the blockchain. In this work, we propose ChainGuards, a decentralized system that uses product-specific rules to verify data collected across the supply chain, with particular focus on sensor-derived information, issuing warnings and triggering audits when anomalies are detected. We evaluated ChainGuards using data from a real cherry supply chain deployment. The result shows that the implemented solution provides reliable verification of supply chain data with low performance overhead, able to correctly detect data discrepancies and inconsistencies.

Diffusion Transformers (DiTs) have significantly enhanced text-to-image (T2I) generation quality, enabling high-quality personalized content creation. However, fine-tuning these models requires substantial computational complexity and memory, limiting practical deployment under resource constraints. To tackle these challenges, we propose a memory-efficient fine-tuning framework called DiT-BlockSkip, integrating timestep-aware dynamic patch sampling and block skipping by precomputing residual features. Our dynamic patch sampling strategy adjusts patch sizes based on the diffusion timestep, then resizes the cropped patches to a fixed lower resolution. This approach reduces forward & backward memory usage while allowing the model to capture global structures at higher timesteps and fine-grained details at lower timesteps. The block skipping mechanism selectively fine-tunes essential transformer blocks and precomputes residual features for the skipped blocks, significantly reducing training memory. To identify vital blocks for personalization, we introduce a block selection strategy based on cross-attention masking. Evaluations demonstrate that our approach achieves competitive personalization performance qualitatively and quantitatively, while reducing memory usage substantially, moving toward on-device feasibility (e.g., smartphones, IoT devices) for large-scale diffusion transformers.

We present an LLM-driven multi-agent probabilistic modeling framework that demonstrates how differences in students' subjective social perceptions arise and evolve in real-world classroom settings, under constraints from an observed social network and limited questionnaire data. When social information is incomplete and the accuracy of perception differs between students, they can form different views of the same group structure from local cues they can access. Repeated peer communication and belief updates can gradually change these views and, over time, lead to stable group-level differences. To avoid assuming a global "god's-eye view," we assign each student an individualized subjective graph that shows which social ties they can perceive and how far information is reachable from their perspective. All judgments and interactions are restricted to this subjective graph: agents use retrieval-augmented generation (RAG) to access only local information and then form evaluations of peers' competence and social standing. We also add structural perturbations related to social-anxiety to represent consistent individual differences in the accuracy of social perception. During peer exchanges, agents share narrative assessments of classmates' academic performance and social position with uncertainty tags, and update beliefs probabilistically using LLM-based trust scores. Using the time series of six real exam scores as an exogenous reference, we run multi-step simulations to examine how epistemic uncertainty spreads through local interactions. Experiments show that, without relying on global information, the framework reproduces several collective dynamics consistent with real-world educational settings. The code is released at https://anonymous.4open.science/r/Rashomonomon-0126.

Graph neural network (GNN) is a powerful tool for analyzing graph-structured data. However, their vulnerability to adversarial attacks raises serious concerns, especially when dealing with sensitive information. Local Differential Privacy (LDP) offers a privacy-preserving framework for training GNNs, but its impact on adversarial robustness remains underexplored. This paper investigates adversarial attacks on LDP-protected GNNs. We explore how the privacy guarantees of LDP can be leveraged or hindered by adversarial perturbations. The effectiveness of existing attack methods on LDP-protected GNNs are analyzed and potential challenges in crafting adversarial examples under LDP constraints are discussed. Additionally, we suggest directions for defending LDP-protected GNNs against adversarial attacks. This work investigates the interplay between privacy and security in graph learning, highlighting the need for robust and privacy-preserving GNN architectures.

We consider a realistic decentralized setup with bandwidth-constrained communication and derive optimal time complexities for non-convex stochastic parallel and asynchronous optimization (up to logarithmic factors). We develop the corresponding methods, Grace SGD and Leon SGD, for both homogeneous and heterogeneous settings. Unlike previous work, our optimal bounds are characterized in terms of min-cut/max-flow quantities and rely on tools from Gomory-Hu trees and Steiner Tree Packing problems, providing tighter and more practical complexities.

Decoder-only language models can be adapted to diverse tasks through instruction finetuning, but the extent to which this generalizes at small scale for low-resource languages remains unclear. We focus on the languages of South Africa, where we are not aware of a publicly available decoder-only model that explicitly targets all eleven official written languages, nine of which are low-resource. We introduce MzansiText, a curated multilingual pretraining corpus with a reproducible filtering pipeline, and MzansiLM, a 125M-parameter language model trained from scratch. We evaluate MzansiLM on natural language understanding and generation using three adaptation regimes: monolingual task-specific finetuning, multilingual task-specific finetuning, and general multi-task instruction finetuning. Monolingual task-specific finetuning achieves strong performance on data-to-text generation, reaching 20.65 BLEU on isiXhosa and competing with encoder-decoder baselines over ten times larger. Multilingual task-specific finetuning benefits closely related languages on topic classification, achieving 78.5% macro-F1 on isiXhosa news classification. While MzansiLM adapts effectively to supervised NLU and NLG tasks, few-shot reasoning remains challenging at this model size, with performance near chance even for much larger decoder-only models. We release MzansiText and MzansiLM to provide a reproducible decoder-only baseline and clear guidance on adaptation strategies for South African languages at small scale.

Existing prompting paradigms structure LLM reasoning in limited topologies: Chain-of-Thought (CoT) produces linear traces, while Tree-of-Thought (ToT) performs branching search. Yet complex reasoning often requires merging intermediate results, revisiting hypotheses, and integrating evidence from multiple sources. We propose Network-of-Thought (NoT), a framework that models reasoning as a directed graph with typed nodes and edges, guided by a heuristic-based controller policy. Across four benchmarks (GSM8K, Game of 24, HotpotQA, ProofWriter) and three models (GPT-4o-mini, Llama-3.3-70B-Instruct, Qwen2.5-72B-Instruct), we investigate when network topology outperforms chain or tree structures, whether LLM-generated heuristics can guide graph-based reasoning search, and the computation-accuracy tradeoff across topologies, evaluating each method on accuracy, topology simplicity, and token efficiency. Our results show that CoT remains effective for sequential tasks with GPT-4o-mini (89.5\% on GSM8K), while NoT surpasses ToT on multi-hop reasoning (91.0\% vs.\ 88.0\% on HotpotQA with LLM-as-Judge). With 72B open-source models, NoT achieves the highest accuracy on GSM8K (91.5\%), and Qwen2.5-72B achieves the best multi-hop QA result overall (91.7\% on HotpotQA). Self-generated controller heuristics outperform fixed and random strategies on logical reasoning, with uncertainty-only weighting achieving 57.0\% on ProofWriter. We also find that evaluation methodology significantly impacts method rankings: string-match underestimates all methods on open-ended QA, with the largest gap for NoT, a pattern consistent across all three models (14--18 percentage point gap on HotpotQA).

Acoustic borehole images provide high-resolution borehole-wall structure, but large-scale interpretation remains difficult because dense expert annotations are rarely available and subsurface information is intrinsically multimodal. The challenge is developing weakly supervised methods combining two-dimensional image texture with depth-aligned one-dimensional well-logs. Here, we introduce a weakly supervised multimodal segmentation framework that refines threshold-guided pseudo-labels through learned models. This preserves the annotation-free character of classical thresholding and clustering workflows while extending them with denoising, confidence-aware pseudo-supervision, and physically structured fusion. We establish that threshold-guided learned refinement provides the most robust improvement over raw thresholding, denoised thresholding, and latent clustering baselines. Multimodal performance depends strongly on fusion strategy: direct concatenation provides limited gains, whereas depth-aware cross-attention, gated fusion, and confidence-aware modulation substantially improve agreement with the weak supervisory reference. The strongest model, confidence-gated depth-aware cross-attention (CG-DCA), consistently outperforms threshold-based, image-only, and earlier multimodal baselines. Targeted ablations show its advantage depends specifically on confidence-aware fusion and structured local depth interaction rather than model complexity alone. Cross-well analyses confirm this performance is broadly stable. These results establish a practical, scalable framework for annotation-free segmentation, showing multimodal improvement is maximized when auxiliary logs are incorporated selectively and depth-aware.

Multi-RF Fusion achieves a test ROC-AUC of 0.8476 +/- 0.0002 on ogbg-molhiv (10 seeds), placing #1 on the OGB leaderboard ahead of HyperFusion (0.8475 +/- 0.0003). The core of the method is a rank-averaged ensemble of 12 Random Forest models trained on concatenated molecular fingerprints (FCFP, ECFP, MACCS, atom pairs -- 4,263 dimensions total), blended with deep-ensembled GNN predictions at 12% weight. Two findings drive the result: (1) setting max_features to 0.20 instead of the default sqrt(d) gives a +0.008 AUC gain on this scaffold split, and (2) averaging GNN predictions across 10 seeds before blending with the RF eliminates GNN seed variance entirely, dropping the final standard deviation from 0.0008 to 0.0002. No external data or pre-training is used.

Vision-Language-Action (VLA) models are mainstream in embodied intelligence but face high inference costs. Edge-Cloud Collaborative (ECC) deployment offers an effective fix by easing edge-device computing pressure to meet real-time needs. However, existing ECC frameworks are suboptimal for VLA models due to two challenges: (1) Diverse model structures hinder optimal ECC segmentation point identification; (2) Even if the optimal split point is determined, changes in network bandwidth can cause performance drift. To address these issues, we propose a novel ECC deployment framework for various VLA models, termed RoboECC. Specifically, we propose a model-hardware co-aware segmentation strategy to help find the optimal segmentation point for various VLA models. Moreover, we propose a network-aware deployment adjustment approach to adapt to the network fluctuations for maintaining optimal performance. Experiments demonstrate that RoboECC achieves a speedup of up to 3.28x with only 2.55x~2.62x overhead.

From customer feedback to social media, understanding human sentiment in text is central to how machines can interact meaningfully with people. However, despite notable progress, accurately capturing sentiment remains a challenging task, which continues to motivate further research in this area. To this end, we introduce Non-Differential Transformer (NDT). It is inspired by (but in contrast to) the state-of-the-art Differential Transformer (DT) model. While standard Transformers can struggle with irrelevant context, the sota DT model uses attention map subtraction, potentially for noise cancellation. We explore an alternative motivation, hypothesizing that benefits may arise from enabling different attention components to specialize on distinct concepts within the text, similar to multiplexing information channels or mixture models, rather than primarily canceling noise via subtraction. Guided by this concept-multiplexing (ConPlex) view, the specific architecture presented in this paper employs a purely additive strategy. It uses only positive weights, learned during training, to ensure constructive combination of these specialized attention perspectives. This design choice explores positive only integration, though our broader framework also shows promise with less constrained linear combinations involving both positive and negative weights. Our model computes attention via this positively weighted sum of multiple distinct attention maps. This allows the model to constructively integrate diverse signals and potentially capture more complex contextual relationships. Competitive performance is achieved by the proposed model for Sentiment Analysis while tested on multiple datasets. We conclude by presenting our results, challenges and future research agenda in this important area of research.

Proper parameter configuration is a prerequisite for the success of Evolutionary Algorithms (EAs). While various adaptive strategies have been proposed, it remains an open question whether all control dimensions contribute equally to algorithmic scalability. To investigate this, we categorize control variables into numerical parameters (e.g., crossover and mutation rates) and structural parameters (e.g., population size and operator switching), hypothesizing that they play distinct roles. This paper presents an empirical study utilizing a dual-level Deep Reinforcement Learning (DRL) framework to decouple and analyze the impact of these two dimensions on the Traveling Salesman Problem (TSP). We employ a Recurrent PPO agent to dynamically regulate these parameters, treating the DRL model as a probe to reveal evolutionary dynamics. Experimental results confirm the effectiveness of this approach: the learned policies outperform static baselines, reducing the optimality gap by approximately 45% on the largest tested instance (rl5915). Building on this validated framework, our ablation analysis reveals a fundamental insight: while numerical tuning offers local refinement, structural plasticity is the decisive factor in preventing stagnation and facilitating escape from local optima. These findings suggest that future automated algorithm design should prioritize dynamic structural reconfiguration over fine-grained probability adjustment. To facilitate reproducibility, the source code is available at https://github.com/StarDream1314/DRLGA-TSP

Millimeter-wave (mmWave) radar enables contactless respiratory sensing,yet fine-grained monitoring is often degraded by nonstationary interference from body micromotions.To achieve micromotion interference removal,we propose mmWave-Diffusion,an observation-anchored conditional diffusion framework that directly models the residual between radar phase observations and the respiratory ground truth,and initializes sampling within an observation-consistent neighborhood rather than from Gaussian noise-thereby aligning the generative process with the measurement physics and reducing inference overhead. The accompanying Radar Diffusion Transformer (RDT) is explicitly conditioned on phase observations, enforces strict one-to-one temporal alignment via patch-level dual positional encodings, and injects local physical priors through banded-mask multi-head cross-attention, enabling robust denoising and interference removal in just 20 reverse steps. Evaluated on 13.25 hours of synchronized radar-respiration data, mmWave-Diffusion achieves state-of-the-art waveform reconstruction and respiratory-rate estimation with strong generalization. Code repository:https://github.com/goodluckyongw/mmWave-Diffusion.

Multimodal Large Language Models (MLLMs) have demonstrated remarkable potential in medical image analysis. However, their application in gastrointestinal endoscopy is currently hindered by two critical limitations: the misalignment between general model reasoning and standardized clinical cognitive pathways, and the lack of causal association between visual features and diagnostic outcomes. In this paper, we propose a novel Clinical-Cognitive-Aligned (CogAlign) framework to address these challenges. First, we endow the model with rigorous clinical analytical capabilities by constructing the hierarchical clinical cognition dataset and employing Supervised Fine-Tuning (SFT). Unlike conventional approaches, this strategy internalizes the hierarchical diagnostic logic of experts, ranging from anatomical localization and morphological evaluation to microvascular analysis, directly into the model. Second, to eliminate visual bias, we provide a theoretical analysis demonstrating that standard supervised tuning inevitably converges to spurious background correlations. Guided by this insight, we propose a counterfactual-driven reinforcement learning strategy to enforce causal rectification. By generating counterfactual normal samples via lesion masking and optimizing through clinical-cognition-centric rewards, we constrain the model to strictly ground its diagnosis in causal lesion features. Extensive experiments demonstrate that our approach achieves State-of-the-Art (SoTA) performance across multiple benchmarks, significantly enhancing diagnostic accuracy in complex clinical scenarios. All source code and datasets will be made publicly available.

In the immediate aftermath of natural disasters, rapid situational awareness is critical. Traditionally, satellite observations are widely used to estimate damage extent. However, they lack the ground-level perspective essential for characterizing specific structural failures and impacts. Meanwhile, ground-level data (e.g., street-view imagery) remains largely inaccessible during time-sensitive events. This study investigates Satellite-to-Street View Synthesis to bridge this data gap. We introduce two generative strategies to synthesize post-disaster street views from satellite imagery: a Vision-Language Model (VLM)-guided approach and a damage-sensitive Mixture-of-Experts (MoE) method. We benchmark these against general-purpose baselines (Pix2Pix, ControlNet) using a proposed Structure-Aware Evaluation Framework. This multi-tier protocol integrates (1) pixel-level quality assessment, (2) ResNet-based semantic consistency verification, and (3) a novel VLM-as-a-Judge for perceptual alignment. Experiments on 300 disaster scenarios reveal a critical realism--fidelity trade-off: while diffusion-based approaches (e.g., ControlNet) achieve high perceptual realism, they often hallucinate structural details. Quantitative results show that standard ControlNet achieves the highest semantic accuracy, 0.71, whereas VLM-enhanced and MoE models excel in textural plausibility but struggle with semantic clarity. This work establishes a baseline for trustworthy cross-view synthesis, emphasizing that visually realistic generations may still fail to preserve critical structural information required for reliable disaster assessment.

Existing high-dimensional online learning methods often face the challenge that their error bounds, or per-batch sample sizes, diverge as the number of data batches increases. To address this issue, we propose an asynchronous decomposition framework that leverages summary statistics to construct a surrogate score function for current-batch learning. This framework is implemented via a dynamic-regularized iterative hard thresholding algorithm, providing a computationally and memory-efficient solution for sparse online optimization. We provide a unified theoretical analysis that accounts for both the streaming computational error and statistical accuracy, establishing that our estimator maintains non-divergent error bounds and $\ell_0$ sparsity across all batches. Furthermore, the proposed estimator adaptively achieves additional gains as batches accumulate, attaining the oracle accuracy as if the entire historical dataset were accessible and the true support were known. These theoretical properties are further illustrated through an example of the generalized linear model.

This paper investigates morphosyntactic covariation in Brazilian Portuguese (BP) to assess whether dialectal origin can be inferred from the combined behavior of linguistic variables. Focusing on four grammatical phenomena related to pronouns, correlation and clustering methods are applied to model covariation and dialectal distribution. The results indicate that correlation captures only limited pairwise associations, whereas clustering reveals speaker groupings that reflect regional dialectal patterns. Despite the methodological constraints imposed by differences in sample size requirements between sociolinguistics and computational approaches, the study highlights the importance of interdisciplinary research. Developing fair and inclusive language technologies that respect dialectal diversity outweighs the challenges of integrating these fields.

Intelligent control of RF transceivers adapting to dynamic operational conditions is essential in the modern and future communication systems. We propose a multi-agent neurosymbolic AI system, where AI agents are assigned for circuit components. Agents have an internal model and a corresponding control algorithm as its constituents. Modeling of the IF amplifier shows promising results, where the same approach can be extended to all the components, thus creating a fully intelligent RF system.

Executable software engineering data is valuable for training SWE agents, but scaling it remains difficult for two reasons: only a small fraction of real repository changes yield verifiable, high-signal task instances, and naively building repository-specific environments quickly becomes the dominant systems cost. We present SWE-Next, an execution-grounded framework for scalable SWE task and trajectory collection. On the data side, SWE-Next mines real merged pull requests, executes candidate base/merged commit pairs, and retains only those that produce strict test improvements without regressions, yielding self-verifying instances. It also applies strict submission gating so that collected trajectories remain evidence-driven rather than speculative. On the systems side, SWE-Next introduces reusable repo-quarter profiles, which reuse the same environment across nearby commits in time while keeping each task run separate and reproducible. Using only 30 hours and 639GB of environment storage, SWE-Next processes 3,971 seed repositories and 102,582 candidate commit pairs mined from real merged PRs to construct a dataset of 2,308 self-verifying instances. Experiments show that SWE-Next improves downstream pass@1 with fewer or comparable training trajectories, indicating that its gains come not from a stronger trajectory generator, but from higher-signal execution-grounded supervision and more efficient data collection.

Organizations increasingly adopt AI technologies to accelerate their performance and capacity to adapt to market dynamics. This study examines how organizations implement AI in experimental methodologies such as growth hacking, lean startup, design thinking, and agile methodology to enhance efficiency and effectiveness. We performed a systematic literature review following the PRISMA 2020 framework, analyzing 37 articles from Web of Science (WOS) and Scopus databases published between 2018 and 2024 to assess AI integration with experimental approaches. Our findings indicate that AI plays a pivotal role in enhancing these methodologies by offering advanced tools for data analysis, real-time feedback, automation, and process optimization. For instance, AI-driven analytics improves decision-making in growth hacking, streamlines iterative cycles in lean startups, enhances creativity in design thinking, and optimizes task prioritization in agile methodology. Furthermore, we identified several real-world cases that successfully utilized AI in experimental strategies and improved their performance across various industries. However, despite the clear advantages of AI integration, organizations face barriers such as skill gaps, ethical concerns, and data governance issues. Addressing these challenges requires a strategic approach to AI adoption, including workforce training, strict data management, and following ethical standards.

Spiking Neural Networks (SNNs) are promising for energy-efficient, real-time edge computing, yet their performance is often constrained by the limited adaptability of conventional leaky integrate-and-fire (LIF) neurons. Existing LIF models struggle with restricted information capacity and susceptibility to noise, leading to degraded accuracy and compromised robustness. Inspired by the dynamic self-regulation of biological potassium channels, we propose the Potassium-regulated LIF (KvLIF) neuron model. KvLIF introduces an auxiliary conductance state that integrates membrane potential and spiking history to adaptively modulate neuronal excitability and reset dynamics. This design extends the dynamic response range of neurons to varying input intensities and effectively suppresses noise-induced spikes. We extensively evaluate KvLIF on both static image and neuromorphic datasets, demonstrating consistent improvements in classification accuracy and superior robustness compared to existing LIF models. Our work bridges biological plausibility with computational efficiency, offering a neuron model that enhances SNN performance while maintaining suitability for low-power neuromorphic deployment.

Recent advancements in text-to-speech technologies enable generating high-fidelity synthetic speech nearly indistinguishable from real human voices. While recent studies show the efficacy of self-supervised learning-based speech encoders for deepfake detection, these models struggle to generalize across unseen speakers. Our quantitative analysis suggests these encoder representations are substantially influenced by speaker information, causing detectors to exploit speaker-specific correlations rather than artifact-related cues. We call this phenomenon speaker entanglement. To mitigate this reliance, we introduce SNAP, a speaker-nulling framework. We estimate a speaker subspace and apply orthogonal projection to suppress speaker-dependent components, isolating synthesis artifacts within the residual features. By reducing speaker entanglement, SNAP encourages detectors to focus on artifact-related patterns, leading to state-of-the-art performance.

Echo State Networks (ESNs) are a reservoir computing framework widely used for nonlinear time-series prediction. However, despite their effectiveness, the randomly initialized reservoir often contains redundant nodes, leading to unnecessary computational overhead and reduced efficiency. In this work, we propose a graph centrality-based pruning approach that interprets the reservoir as a weighted directed graph and removes structurally less important nodes using centrality measures. Experiments on Mackey-Glass time-series prediction and electric load forecasting demonstrate that the proposed method can significantly reduce reservoir size while maintaining, and in some cases improving, prediction accuracy, while preserving the essential reservoir dynamics.

Integrating structural and functional connectomes remains challenging because their relationship is non-linear and organized over nested modular hierarchies. We propose a hierarchical multiscale structure-function coupling framework for connectome integration that jointly learns individualized modular organization and hierarchical coupling across structural connectivity (SC) and functional connectivity (FC). The framework includes: (i) Prototype-based Modular Pooling (PMPool), which learns modality-specific multiscale communities by selecting prototypical ROIs and optimizing a differentiable modularity-inspired objective; (ii) an Attention-based Hierarchical Coupling Module (AHCM) that models both within-hierarchy and cross-hierarchy SC-FC interactions to produce enriched hierarchical coupling representations; and (iii) a Coupling-guided Clustering loss (CgC-Loss) that regularizes SC and FC community assignments with coupling signals, allowing cross-modal interactions to shape community alignment across hierarchies. We evaluate the model's performance across four cohorts for predicting brain age, cognitive score, and disease classification. Our model consistently outperforms baselines and other state-of-the-art approaches across three tasks. Ablation and sensitivity analyses verify the contributions of key components. Finally, the visualizations of learned coupling reveal interpretable differences, suggesting that the framework captures biologically meaningful structure-function relationships.

Contemporary societies face a severe crisis of demographic reproduction. Global fertility rates continue to decline precipitously, with East Asian nations exhibiting the most dramatic trends -- China's total fertility rate (TFR) fell to approximately 1.0 in 2023, while South Korea's dropped below 0.72. Simultaneously, the institution of marriage is undergoing structural disintegration: educated women rationally reject unions lacking both emotional fulfillment and economic security, while a growing proportion of men at the lower end of the socioeconomic spectrum experience chronic sexual deprivation, anxiety, and learned helplessness. This paper proposes a computational framework for modeling and evaluating a Stratified Polyamory System (SPS) using techniques from agent-based modeling (ABM), multi-agent reinforcement learning (MARL), and large language model (LLM)-empowered social simulation. The SPS permits individuals to maintain a limited number of legally recognized secondary partners in addition to one primary spouse, combined with socialized child-rearing and inheritance reform. We formalize the A/B/C stratification as heterogeneous agent types in a multi-agent system and model the matching process as a MARL problem amenable to Proximal Policy Optimization (PPO). The mating network is analyzed using graph neural network (GNN) representations. Drawing on evolutionary psychology, behavioral ecology, social stratification theory, computational social science, algorithmic fairness, and institutional economics, we argue that SPS can improve aggregate social welfare in the Pareto sense. Preliminary computational results demonstrate the framework's viability in addressing the dual crisis of female motherhood penalties and male sexlessness, while offering a non-violent mechanism for wealth dispersion analogous to the historical Chinese Grace Decree (Tui'en Ling).

Retrieval-augmented language models can retrieve relevant evidence yet still commit to answers before explicitly checking whether the retrieved context supports the conclusion. We present PAVE (Premise-Grounded Answer Validation and Editing), an inference-time validation layer for evidence-grounded question answering. PAVE decomposes retrieved context into question-conditioned atomic facts, drafts an answer, scores how well that draft is supported by the extracted premises, and revises low-support outputs before finalization. The resulting trace makes answer commitment auditable at the level of explicit premises, support scores, and revision decisions. In controlled ablations with a fixed retriever and backbone, PAVE outperforms simpler post-retrieval baselines in two evidence-grounded QA settings, with the largest gain reaching 32.7 accuracy points on a span-grounded benchmark. We view these findings as proof-of-concept evidence that explicit premise extraction plus support-gated revision can strengthen evidence-grounded consistency in retrieval-augmented LLM systems.

The problem of constrained online convex optimization is considered, where at each round, once a learner commits to an action $x_t \in \mathcal{X} \subset \mathbb{R}^d$, a convex loss function $f_t$ and a convex constraint function $g_t$ that drives the constraint $g_t(x)\le 0$ are revealed. The objective is to simultaneously minimize the static regret and cumulative constraint violation (CCV) compared to the benchmark that knows the loss functions and constraint functions $f_t$ and $g_t$ for all $t$ ahead of time, and chooses a static optimal action that is feasible with respect to all $g_t(x)\le 0$. In recent prior work Sinha and Vaze [2024], algorithms with simultaneous regret of $O(\sqrt{T})$ and CCV of $O(\sqrt{T})$ or (CCV of $O(1)$ in specific cases Vaze and Sinha [2025], e.g. when $d=1$) have been proposed. It is widely believed that CCV is $Ω(\sqrt{T})$ for all algorithms that ensure that regret is $O(\sqrt{T})$ with the worst case input for any $d\ge 2$. In this paper, we refute this and show that the algorithm of Vaze and Sinha [2025] simultaneously achieves regret of $O(\sqrt{T})$ regret and CCV of $O(T^{1/3})$ when $d=2$.

The rapid growth in the volume, variety, and velocity of geospatial data has created data ecosystems that are highly distributed, heterogeneous, and semantically inconsistent. Existing data catalogs, portals, and infrastructures still rely largely on keyword-based search with limited semantic support, which often fails to capture user intent and leads to weak retrieval performance. To address these challenges, this study proposes a knowledge graph-driven multi-agent framework for intelligent geospatial data discovery, powered by large language models. The framework introduces a unified geospatial metadata ontology as a semantic mediation layer to align heterogeneous metadata standards across platforms and constructs a geospatial metadata knowledge graph to explicitly model datasets and their multidimensional relationships. Building on the structured representation, the framework adopts a multi-agent collaborative architecture to perform intent parsing, knowledge graph retrieval, and answer synthesis, forming an interpretable and closed-loop discovery process from user queries to results. Results from representative use cases and performance evaluation show that the framework substantially improves intent matching accuracy, ranking quality, recall, and discovery transparency compared with traditional systems. This study advances geospatial data discovery toward a more semantic, intent-aware, and intelligent paradigm, providing a practical foundation for next-generation intelligent and autonomous spatial data infrastructures and contributing to the broader vision of Autonomous GIS.

Existing prompt-optimization techniques rely on local signals to update behavior, often neglecting broader and recurring patterns across tasks, leading to poor generalization; they further rely on full-prompt rewrites or unstructured merges, resulting in knowledge loss. These limitations are magnified in research-coding workflows, which involve heterogeneous repositories, underspecified environments, and weak feedback, where reproducing results from public codebases is an established evaluation regime. We introduce Reflective Evolving Research Engineer (REVERE), a framework that continuously learns from Global Training Context, recognizes recurring failure modes in cross-repository execution trajectories, distills them into reusable heuristics, and performs targeted edits across three configurable fields: the system prompt, a task-prompt template, and a cumulative cheatsheet. REVERE, via this reflective optimization framework, improves performance over prior state-of-the-art expert-crafted instructions on research coding tasks by 4.50% on SUPER, 3.51% on ResearchCodeBench, and 4.89% on ScienceAgentBench across their respective metrics. These results demonstrate that agents equipped with mechanisms for continual learning and global memory consolidation can meaningfully evolve their capabilities over time.

Despite remarkable advances in large Vision-Language Models (VLMs), spatial reasoning remains a persistent challenge. In this work, we investigate how attention heads within VLMs contribute to spatial reasoning by analyzing their functional roles through a mechanistic interpretability lens. We introduce CogVSR, a dataset that decomposes complex spatial reasoning questions into step-by-step subquestions designed to simulate human-like reasoning via a chain-of-thought paradigm, with each subquestion linked to specific cognitive functions such as spatial perception or relational reasoning. Building on CogVSR, we develop a probing framework to identify and characterize attention heads specialized for these functions. Our analysis across diverse VLM families reveals that these functional heads are universally sparse, vary in number and distribution across functions. Notably, spatially specialized heads are fewer than those for other cognitive functions, highlighting their scarcity. We propose methods to activate latent spatial heads, improving spatial understanding. Intervention experiments further demonstrate their critical role in spatial reasoning: removing functional heads leads to performance degradation, while emphasizing them enhances accuracy. This study provides new interpretability driven insights into how VLMs attend to space and paves the way for enhancing complex spatial reasoning in multimodal models.

Large language model (LLM) services are mostly centralized, leading to scalability bottlenecks and underutilization of substantial scattered GPU resources. While decentralization offers a promising alternative, existing frameworks primarily focus on cooperation among GPU providers while overlooking their inherent competitive dynamics, imposing substantial constraints such as excessive platform-level oversight or rigid requirements to execute all assigned requests using fixed software stacks on fixed hardware configurations. We argue that such assumptions are unrealistic in real-world decentralized environments. To this end, we propose WWW.Serve, a decentralized framework for interconnecting LLM services worldwide. It allows participants to flexibly determine their participation policies and resource commitments, and supports self-organizing request dispatch, enabling the network to autonomously allocate requests without centralized coordination. Empirically, we show that WWW.Serve improves global SLO (service-level-objective) attainment by up to 1.5x and lowers latency by 27.6%. Its performance approaches, and in some cases surpasses, centralized scheduling, while fully preserving the benefits of decentralization. These results highlight WWW.Serve as a promising foundation for real-world, decentralized LLM serving.

We propose a dense associative memory for empirical measures (weighted point clouds). Stored patterns and queries are finitely supported probability measures, and retrieval is defined by minimizing a Hopfield-style log-sum-exp energy built from the debiased Sinkhorn divergence. We derive retrieval dynamics as a spherical Hellinger Kantorovich (SHK) gradient flow, which updates both support locations and weights. Discretizing the flow yields a deterministic algorithm that uses Sinkhorn potentials to compute barycentric transport steps and a multiplicative simplex reweighting. Under local separation and PL-type conditions we prove basin invariance, geometric convergence to a local minimizer, and a bound showing the minimizer remains close to the corresponding stored pattern. Under a random pattern model, we further show that these Sinkhorn basins are disjoint with high probability, implying exponential capacity in the ambient dimension. Experiments on synthetic Gaussian point-cloud memories demonstrate robust recovery from perturbed queries versus a Euclidean Hopfield-type baseline.

We introduce Exponential Family Discriminant Analysis (EFDA), a unified generative framework that extends classical Linear Discriminant Analysis (LDA) beyond the Gaussian setting to any member of the exponential family. Under the assumption that each class-conditional density belongs to a common exponential family, EFDA derives closed-form maximum-likelihood estimators for all natural parameters and yields a decision rule that is linear in the sufficient statistic, recovering LDA as a special case and capturing nonlinear decision boundaries in the original feature space. We prove that EFDA is asymptotically calibrated and statistically efficient under correct specification, and we generalise it to $K \geq 2$ classes and multivariate data. Through extensive simulation across five exponential-family distributions (Weibull, Gamma, Exponential, Poisson, Negative Binomial), EFDA matches the classification accuracy of LDA, QDA, and logistic regression while reducing Expected Calibration Error (ECE) by $2$--$6\times$, a gap that is \emph{structural}: it persists for all $n$ and across all class-imbalance levels, because misspecified models remain asymptotically miscalibrated. We further prove and empirically confirm that EFDA's log-odds estimator approaches the Cramér-Rao bound under correct specification, and is the only estimator in our comparison whose mean squared error converges to zero. Complete derivations are provided for nine distributions. Finally, we formally verify all four theoretical propositions in Lean 4, using Aristotle (Harmonic) and OpenGauss (Math, Inc.) as proof generators, with all outputs independently machine-checked by AXLE (Axiom).

Classical Amdahl's Law assumes a fixed decomposition between serial and parallel work and homogeneous replication; historically, it bounds how much parallel speedup is attainable. Modern systems instead combine specialized accelerators with programmable compute, tensor datapaths, and evolving pipelines, while empirical scaling laws shift which stages absorb marginal compute. The central tension is therefore not the serial-versus-parallel split alone, but resource allocation across heterogeneous hardware, given efficiency differences, and workload structures that determine how effectively additional compute can be converted into value. We reformulate Amdahl's Law for modern heterogeneous systems with scalable workloads. The analysis yields a finite collapse threshold: beyond a critical scalable fraction, specialization becomes suboptimal for any efficiency advantage of specialized hardware over programmable compute, and optimal specialized investment falls to zero, a phase transition rather than an asymptotic tail. We use this framework to interpret increasing GPU programmability and why domain-specific AI accelerators have not displaced GPUs.

Deploying high-fidelity AI tutors in schools is often blocked by the Resource Curse -- the need for expensive cloud GPUs and massive data engineering. In this practitioner report, we present a replicable Standard Operating Procedure that breaks this barrier. Using a Vision-Language Model data cleaning strategy and a novel Shadow-RAG architecture, we localized a graduate-level Applied Mathematics tutor using only 3 person-days of non-expert labor and open-weights 32B models deployable on a single consumer-grade GPU. Our pilot study on a full graduate-level final exam reveals a striking emergence phenomenon: while both zero-shot baselines and standard retrieval stagnate around 50-60% accuracy across model generations, the Shadow Agent, which provides structured reasoning guidance, triggers a massive capability surge in newer 32B models, boosting performance from 74% (Naive RAG) to mastery level (90%). In contrast, older models see only modest gains (~10%). This suggests that such guidance is the key to unlocking the latent power of modern small language models. This work offers a cost-effective, scientifically grounded blueprint for ubiquitous AI education.

Most fine-grained fashion image retrieval (FIR) methods assume a static setting, requiring full retraining when new attributes appear, which is costly and impractical for dynamic scenarios. Although pretrained models support zero-shot inference, their accuracy drops without supervision, and no prior work explores class-incremental learning (CIL) for fine-grained FIR. We propose a multihead continual learning framework for fine-grained fashion image retrieval with contrastive learning and exponential moving average (EMA) distillation (MCL-FIR). MCL-FIR adopts a multi-head design to accommodate evolving classes across increments, reformulates triplet inputs into doublets with InfoNCE for simpler and more effective training, and employs EMA distillation for efficient knowledge transfer. Experiments across four datasets demonstrate that, beyond its scalability, MCL-FIR achieves a strong balance between efficiency and accuracy. It significantly outperforms CIL baselines under similar training cost, and compared with static methods, it delivers comparable performance while using only about 30% of the training cost. The source code is publicly available in https://github.com/Dr-LingXiao/MCL-FIR.

In next generation of Wi-Fi networks Multiple Access Point Coordination (MAPC) is poised to significantly enhance the network performance by enabling a set of Access Points (APs) to coordinate with each other through advanced coordinating schemes so that to reduce inter-AP contention and congestion. This paper focuses on defining a framework to facilitate the coordination across multi-APs when these employ Coordinated Spatial Reuse (C-SR). In this case, the coordinating APs may need to reciprocally adjust their scheduling strategy, power control and link adaptation to meet specific Quality of Service (QoS) requirements, which by using classical approaches leads to high overhead due to negotiations needed across APs, and requires complex solutions in order to properly optimize the network across all the parameters in play. In this matter, a two layer Multi-Armed Bandit (MAB) algorithm has been proposed to optimize such a network while preserving the fair use of resources across all nodes. The validity of this holistic approach is confirmed by system level simulations, which show that the proposed algorithm not only improves the network in terms of sum-throughput, but also allows to enhance fairness, making this a robust solution for next-generation of Wi-Fi networks.

The scalability of quantum computing in supporting sophisticated algorithms critically depends not only on qubit quality and error handling, but also on the efficiency of classical control, constrained by the cryogenic control bandwidth and energy budget. In this work, we address this challenge by investigating the algorithmic complexity of quantum circuits at the instruction set architecture (ISA) level. We introduce an energy-efficient quantum instruction set architecture (EQISA) that synthesizes quantum circuits in a discrete Solovay-Kitaev basis of fixed depth and encodes instruction streams using a sparse dictionary learned from decomposing a set of Haar-random unitaries, followed by entropy-optimal Huffman coding and an additional lossless bzip2 compression stage. This approach is evaluated on benchmark quantum circuits demonstrating over 60% compression of quantum instruction streams across system sizes, enabling proportional reductions in classical control energy and communication overhead without loss of computational fidelity. Beyond compression, EQISA facilitates the discovery of higher-level composable abstractions in quantum circuits and provides estimates of quantum algorithmic complexity. These findings position EQISA as an impactful direction for improving the energy efficiency and scalability of quantum control architectures.

Diffusion models have become a leading framework in generative modeling, yet their theoretical understanding -- especially for high-dimensional data concentrated on low-dimensional structures -- remains incomplete. This paper investigates how diffusion models learn such structured data, focusing on two key aspects: statistical complexity and influence of data geometric properties. By modeling data as samples from a smooth Riemannian manifold, our analysis reveals crucial decompositions of score functions in diffusion models under different levels of injected noise. We also highlight the interplay of manifold curvature with the structures in the score function. These analyses enable an efficient neural network approximation to the score function, built upon which we further provide statistical rates for score estimation and distribution learning. Remarkably, the obtained statistical rates are governed by the intrinsic dimension of data and the manifold curvature. These results advance the statistical foundations of diffusion models, bridging theory and practice for generative modeling on manifolds.

How do transformer language models represent magnitude? Recent work disagrees: some find logarithmic spacing, others linear encoding, others per-digit circular representations. We apply the formal tools of psychophysics to resolve this. Using four converging paradigms (representational similarity analysis, behavioural discrimination, precision gradients, causal intervention) across three magnitude domains in three 7-9B instruction-tuned models spanning three architecture families (Llama, Mistral, Qwen), we report three findings. First, representational geometry is consistently log-compressive: RSA correlations with a Weber-law dissimilarity matrix ranged from .68 to .96 across all 96 model-domain-layer cells, with linear geometry never preferred. Second, this geometry is dissociated from behaviour: one model produces a human-range Weber fraction (WF = 0.20) while the other does not, and both models perform at chance on temporal and spatial discrimination despite possessing logarithmic geometry. Third, causal intervention reveals a layer dissociation: early layers are functionally implicated in magnitude processing (4.1x specificity) while later layers where geometry is strongest are not causally engaged (1.2x). Corpus analysis confirms the efficient coding precondition (alpha = 0.77). These results suggest that training data statistics alone are sufficient to produce log-compressive magnitude geometry, but geometry alone does not guarantee behavioural competence.

Multi-Agent Debate has emerged as a promising framework for improving the reasoning quality of large language models through iterative inter-agent communication. However, broadcasting all agent messages at every round introduces noise and redundancy that can degrade debate quality and waste computational resources. Current approaches rely on uncertainty estimation to filter low-confidence responses before broadcasting, but this approach is unreliable due to miscalibrated confidence scores and sensitivity to threshold selection. To address this, we propose Diversity-Aware Retention (DAR), a lightweight debate framework that, at each debate round, selects the subset of agent responses that maximally disagree with each other and with the majority vote before broadcasting. Through an explicit index-based retention mechanism, DAR preserves the original messages without modification, ensuring that retained disagreements remain authentic. Experiments on diverse reasoning and question answering benchmarks demonstrate that our selective message propagation consistently improves debate performance, particularly as the number of agents scales, where noise accumulation is most severe. Our results highlight that what agents hear is as important as what agents say in multi-agent reasoning systems.

The "AI singularity" is often miscast as a monolithic, godlike mind. Evolution suggests a different path: intelligence is fundamentally plural, social, and relational. Recent advances in agentic AI reveal that frontier reasoning models, such as DeepSeek-R1, do not improve simply by "thinking longer". Instead, they simulate internal "societies of thought," spontaneous cognitive debates that argue, verify, and reconcile to solve complex tasks. Moreover, we are entering an era of human-AI centaurs: hybrid actors where collective agency transcends individual control. Scaling this intelligence requires shifting from dyadic alignment (RLHF) toward institutional alignment. By designing digital protocols, modeled on organizations and markets, we can build a social infrastructure of checks and balances. The next intelligence explosion will not be a single silicon brain, but a complex, combinatorial society specializing and sprawling like a city. No mind is an island.

Large Language Models (LLMs) are increasingly adopted for vulnerability detection, yet their reasoning remains fundamentally unsound. We identify a root cause shared by both major mitigation paradigms (agent-based debate and retrieval augmentation): reasoning in an ungrounded deliberative space that lacks a bounded, hypothesis-specific evidence base. Without such grounding, agents fabricate cross-function dependencies, and retrieval heuristics supply generic knowledge decoupled from the repository's data-flow topology. Consequently, the resulting conclusions are driven by rhetorical persuasiveness rather than verifiable facts. To ground this deliberation, we present AEGIS, a novel multi-agent framework that shifts detection from ungrounded speculation to forensic verification over a closed factual substrate. Guided by a "From Clue to Verdict" philosophy, AEGIS first identifies suspicious code anomalies (clues), then dynamically reconstructs per-variable dependency chains for each clue via on-demand slicing over a repository-level Code Property Graph. Within this closed evidence boundary, a Verifier Agent constructs competing dialectical arguments for and against exploitability, while an independent Audit Agent scrutinizes every claim against the trace, exercising veto power to prevent hallucinated verdicts. Evaluation on the rigorous PrimeVul dataset demonstrates that AEGIS establishes a new state-of-the-art, achieving 122 Pair-wise Correct Predictions. To our knowledge, this is the first approach to surpass 100 on this benchmark. It reduces the false positive rate by up to 54.40% compared to leading baselines, at an average cost of $0.09 per sample without any task-specific training.

Detecting product price outliers is important for retail and e-commerce stores as erroneous or unexpectedly high prices adversely affect competitiveness, revenue, and consumer trust. Classical techniques offer simple thresholds while ignoring the rich semantic relationships among product attributes. We propose an agentic Large Language Model (LLM) framework that treats outlier price flagging as a reasoning task grounded in related product detection and comparison. The system processes the prices of target products in three stages: (i) relevance classification selects price-relevant similar products using product descriptions and attributes; (ii) relative utility assessment evaluates the target product against each similar product along price influencing dimensions (e.g., brand, size, features); (iii) reasoning-based decision aggregates these justifications into an explainable price outlier judgment. The framework attains over 75% agreement with human auditors on a test dataset, and outperforms zero-shot and retrieval based LLM techniques. Ablation studies show the sensitivity of the method to key hyper-parameters and testify on its flexibility to be applied to cases with different accuracy requirement and auditor agreements.

Accurately characterizing non-linear functional manifolds with singularities is a fundamental challenge in scientific computing. While Multi-Layer Perceptrons (MLPs) dominate, their spectral bias hinders resolving high-curvature features without excessive parameters. We introduce Continued Fraction Neural Networks (CFNNs), integrating continued fractions with gradient-based optimization to provide a ``rational inductive bias.'' This enables capturing complex asymptotics and discontinuities with extreme parameter frugality. We provide formal approximation bounds demonstrating exponential convergence and stability guarantees. To address recursive instability, we develop three implementations: CFNN-Boost, CFNN-MoE, and CFNN-Hybrid. Benchmarks show CFNNs consistently outperform MLPs in precision with one to two orders of magnitude fewer parameters, exhibiting up to a 47-fold improvement in noise robustness and physical consistency. By bridging black-box flexibility and white-box transparency, CFNNs establish a reliable ``grey-box'' paradigm for AI-driven scientific research.

We present Seed1.8, a foundation model aimed at generalized real-world agency: going beyond single-turn prediction to multi-turn interaction, tool use, and multi-step execution. Seed1.8 keeps strong LLM and vision-language performance while supporting a unified agentic interface-search, code generation and execution, and GUI interaction. For deployment, it offers latency- and cost-aware inference, including configurable thinking modes and optimized visual encoding for images and video. We report evaluations on standard benchmarks and application-aligned workflows spanning foundational skills, multimodal understanding, and agentic behavior. Seed1.8 is released to support further research and development on interactive, real-world use cases.

Large language model (LLM) training is often bottlenecked by memory constraints and stochastic gradient noise in extremely high-dimensional parameter spaces. Motivated by empirical evidence that many LLM gradient matrices are effectively low-rank during training, we present an unbiased, memory-efficient, low-rank matrix estimator with the lowest variance that is applicable across common stochastic gradient estimation paradigms. The core idea is to project a high-dimensional stochastic gradient estimator onto a random low-dimensional subspace and lift it back, reducing memory while keeping the estimator unbiased and controlling mean-squared error via an optimally designed projection distribution, including Haar--Stiefel projections. The projection distribution is derived by solving a constrained functional optimization problem, yielding an optimal random projector that guides algorithm design. Empirically, the resulting low-rank gradient estimators deliver both practical memory savings and improved training behavior. In RoBERTa-large fine-tuning, our method attains the lowest peak GPU memory among compared methods (e.g., 3.83GB versus 16.7GB for full BP) while remaining competitive in accuracy; in autoregressive LLM pretraining (LLaMA-20M/60M/100M), our method outperforms the traditional methods, supporting the benefit of the proposed optimal projection strategy.

Despite their dominance in vision and language, deep neural networks often underperform relative to tree-based models on tabular data. To bridge this gap, we incorporate five key inductive biases into deep learning: robustness to irrelevant features, axis alignment, localized irregularities, feature heterogeneity, and training stability. We propose \emph{LassoFlexNet}, an architecture that evaluates the linear and nonlinear marginal contribution of each input via Per-Feature Embeddings, and sparsely selects relevant variables using a Tied Group Lasso mechanism. Because these components introduce optimization challenges that destabilize standard proximal methods, we develop a \emph{Sequential Hierarchical Proximal Adaptive Gradient optimizer with exponential moving averages (EMA)} to ensure stable convergence. Across $52$ datasets from three benchmarks, LassoFlexNet matches or outperforms leading tree-based models, achieving up to a $10$\% relative gain, while maintaining Lasso-like interpretability. We substantiate these empirical results with ablation studies and theoretical proofs confirming the architecture's enhanced expressivity and structural breaking of undesired rotational invariance.

Large language models (LLMs) are changing the way researchers interact with code and data in scientific computing. While their ability to generate general-purpose code is well established, their effectiveness in producing scientifically valid code/input scripting for domain-specific languages (DSLs) remains largely unexplored. We propose an evaluation procedure that enables domain experts (who may not be experts in the DSL) to assess the validity of LLM-generated input files for LAMMPS, a widely used molecular dynamics (MD) code, and to use those assessments to evaluate the performance of state-of-the-art LLMs and identify common issues. Key to the evaluation procedure are a normalization step to generate canonical files and an extensible parser for syntax analysis. The following steps isolate common errors without incurring costly tests (in time and computational resources). Once a working input file is generated, LLMs can accelerate verification tests. Our findings highlight limitations of LLMs in generating scientific DSLs and a practical path forward for their integration into domain-specific computational ecosystems by domain experts.

Modern reconfigurable AI accelerators rely on rich mapping and data-layout flexibility to sustain high utilization across matrix multiplication, convolution, and emerging applications beyond AI. However, exposing this flexibility through fine-grained micro-control results in prohibitive control overhead of fetching configuration bits from off-chip memory. This paper presents MINISA, a minimal instruction set that programs a reconfigurable accelerator at the granularity of Virtual Neurons (VNs), the coarsest control granularity that retains flexibility of hardware and the finest granularity that avoids unnecessary control costs. First, we introduce FEATHER+, a modest refinement of FEATHER, that eliminates redundant on-chip replication needed for runtime dataflow/layout co-switching and supports dynamic cases where input and weight data are unavailable before execution for offline layout manipulation. MINISA then abstracts control of FEATHER+ into three layout-setting instructions for input, weight, and output VNs and a single mapping instruction for setting dataflow. This reduces the control and instruction footprint while preserving the legal mapping and layout space supported by the FEATHER+. Our results show that MINISA reduces geometric mean off-chip instruction traffic by factors ranging from 35x to (4x10^5)x under various sizes under 50 GEMM workloads spanning AI (GPT-oss), FHE, and ZKP. This eliminates instruction-fetch stalls that consume 96.9% of micro-instruction cycles, yielding up to 31.6x end-to-end speedup for 16x256 FEATHER+. Our code: https://github.com/maeri-project/FEATHER/tree/main/minisa.

Graph Neural Network (GNN) on streaming graphs has gained increasing popularity. However, its practical deployment remains challenging, as the inference process relies on Runtime Embedding Computation (RTEC) to capture recent graph changes. This process incurs heavyweight multi-hop graph traversal overhead, which significantly undermines computation efficiency. We observe that the intermediate results for large portions of the graph remain unchanged during graph evolution, and thus redundant computations can be effectively eliminated through carefully designed incremental methods. In this work, we propose an efficient framework for incrementalizing RTEC on streaming graphs.The key idea is to decouple GNN computation into a set of generalized, fine-grained operators and safely reorder them, transforming the expensive full-neighbor GNN computation into a more efficient form over the affected subgraph. With this design, our framework preserves the semantics and accuracy of the original full-neighbor computation while supporting a wide range of GNN models with complex message-passing patterns. To further scale to graphs with massive historical results, we develop a GPU-CPU co-processing system that offloads embeddings to CPU memory with communication-optimized scheduling. Experiments across diverse graph sizes and GNN models show that our method reduces computation by 64%-99% and achieves 1.7x-145.8x speedups over existing solutions.

Can we trust the reasoning traces that large reasoning models (LRMs) produce? We investigate whether these traces faithfully reflect what drives model outputs, and whether models will honestly report their influence. We introduce Thought Injection, a method that injects synthetic reasoning snippets into a model's <think> trace, then measures whether the model follows the injected reasoning and acknowledges doing so. Across 45,000 samples from three LRMs, we find that injected hints reliably alter outputs, confirming that reasoning traces causally shape model behavior. However, when asked to explain their changed answers, models overwhelmingly refuse to disclose the influence: overall non-disclosure exceeds 90% for extreme hints across 30,000 follow-up samples. Instead of acknowledging the injected reasoning, models fabricate aligned-appearing but unrelated explanations. Activation analysis reveals that sycophancy- and deception-related directions are strongly activated during these fabrications, suggesting systematic patterns rather than incidental failures. Our findings reveal a gap between the reasoning LRMs follow and the reasoning they report, raising concern that aligned-appearing explanations may not be equivalent to genuine alignment.

Artificial intelligence (AI) is poised to profoundly reshape how work is executed and organized, but we do not yet have deep frameworks for understanding where AI can be used. Here we provide a comprehensive ontology of work activities that can help systematically analyze and predict uses of AI. To do this, we disaggregate and then substantially reorganize the approximately 20K activities in the US Department of Labor's widely used O*NET occupational database. Next, we use this framework to classify descriptions of 13,275 AI software applications and a worldwide tally of 20.8 million robotic systems. Finally, we use the data about both these kinds of AI to generate graphical displays of how the estimated units and market values of all worldwide AI systems used today are distributed across the work activities that these systems help perform. We find a highly uneven distribution of AI market value across activities, with the top 1.6% of activities accounting for over 60% of AI market value. Most of the market value is used in information-based activities (72%), especially creating information (36%), and only 12% is used in physical activities. Interactive activities include both information-based and physical activities and account for 48% of AI market value, much of which (26%) involves transferring information. These results can be viewed as rough predictions of the AI applicability for all the different work activities down to very low levels of detail. Thus, we believe this systematic framework can help predict at a detailed level where today's AI systems can and cannot be used and how future AI capabilities may change this.

Logs are essential for diagnosing failures and conducting retrospective studies, leading many software organizations to retain log messages for a long time. Nevertheless, the volume of generated log data grows rapidly as software systems grow, necessitating an effective compression method. Apart from general-purpose compressors (e.g., Gzip, Bzip2), many recent studies developed log-specific compression algorithms, but they offer suboptimal performance because of (1) overlooking redundancies within certain complex tokens, and (2) lacking a fine-grained encoding strategy for diverse token types. This work uncovers a new redundancy pattern in structured tokens and proposes a new type-aware encoding strategy to improve log compression. Building on this insight, we introduce LogFold, a novel log compression method consisting of four components: a token analyzer to classifies tokens as structured, unstructured, or static types; a processor that mines recurring patterns within structured tokens based on their delimiter skeletons; a hybrid encoder that tailors data representation according to token types; and a packer that compresses the output into an archive file. Extensive experiments on 16 public log datasets demonstrate that LogFold surpasses state-of-the-art baselines, achieving average compression ratio improvements by 11.11%, with a compression speed of 9.842 MB/s. Ablation studies further indicate the importance of each component. We also conduct sensitivity analyses to verify LogFold's robustness and stability across various internal settings.

Key-value (KV) caching is widely used to accelerate transformer inference, but its memory cost grows linearly with input length, limiting long-context deployment. Existing token eviction methods reduce memory by discarding less important tokens, which can be viewed as a coarse form of dimensionality reduction that assigns each token either zero or full dimension. We propose MixedDimKV, a mixed-dimension KV cache compression method that allocates dimensions to tokens at a more granular level, and MixedDimKV-H, which further integrates head-level importance information. Experiments on long-context benchmarks show that MixedDimKV outperforms prior KV cache compression methods that do not rely on head-level importance profiling. When equipped with the same head-level importance information, MixedDimKV-H consistently outperforms HeadKV. Notably, our approach achieves comparable performance to full attention on LongBench with only 6.25% of the KV cache. Furthermore, in the Needle-in-a-Haystack test, our solution maintains 100% accuracy at a 50K context length while using as little as 0.26% of the cache.

Vision-Language-Action (VLA) models show strong generalization for robotic control, but finetuning them with reinforcement learning (RL) is constrained by the high cost and safety risks of real-world interaction. Training VLA models in interactive world models avoids these issues but introduces several challenges, including pixel-level world modeling, multi-view consistency, and compounding errors under sparse rewards. Building on recent advances across large multimodal models and model-based RL, we propose VLA-MBPO, a practical framework to tackle these problems in VLA finetuning. Our approach has three key design choices: (i) adapting unified multimodal models (UMMs) for data-efficient world modeling; (ii) an interleaved view decoding mechanism to enforce multi-view consistency; and (iii) chunk-level branched rollout to mitigate error compounding. Theoretical analysis and experiments across simulation and real-world tasks demonstrate that VLA-MBPO significantly improves policy performance and sample efficiency, underscoring its robustness and scalability for real-world robotic deployment.

We study Bayesian learning in episodic, finite-horizon zero-sum Markov games with unknown transition and reward models. We investigate a posterior algorithm in which each player maintains a Bayesian posterior over the game model, independently samples a game model at the beginning of each episode, and computes an equilibrium policy for the sampled model. We analyze two settings: (i) Both players use the posterior sampling algorithm, and (ii) Only one player uses posterior sampling while the opponent follows an arbitrary learning algorithm. In each setting, we provide guarantees on the expected regret of the posterior sampling agent. Our notion of regret compares the expected total reward of the learning agent against the expected total reward under equilibrium policies of the true game. Our main theoretical result is an expected regret bound for the posterior sampling agent of order $O(HS\sqrt{ABHK\log(SABHK)})$ where $K$ is the number of episodes, $H$ is the episode length, $S$ is the number of states, and $A,B$ are the action space sizes of the two players. Experiments in a grid-world predator--prey domain illustrate the sublinear regret scaling and show that posterior sampling competes favorably with a fictitious-play baseline.

Solving ill-posed inverse problems necessitates effective regularization strategies to stabilize the inversion process against measurement noise. While classical methods like Tikhonov regularization require heuristic parameter tuning, and standard deep learning approaches often lack interpretability and generalization across resolutions, we propose SC-Net (Spectral Correction Network), a novel operator learning framework. SC-Net operates in the spectral domain of the forward operator, learning a pointwise adaptive filter function that reweights spectral coefficients based on the signal-to-noise ratio. We provide a theoretical analysis showing that SC-Net approximates the continuous inverse operator, guaranteeing discretization invariance. Numerical experiments on 1D integral equations demonstrate that SC-Net: (1) achieves the theoretical minimax optimal convergence rate ($O(δ^{0.5})$ for $s=p=1.5$), matching theoretical lower bounds; (2) learns interpretable sharp-cutoff filters that outperform Oracle Tikhonov regularization; and (3) exhibits zero-shot super-resolution, maintaining stable reconstruction errors ($\approx 0.23$) when trained on coarse grids ($N=256$) and tested on significantly finer grids (up to $N=2048$). The proposed method bridges the gap between rigorous regularization theory and data-driven operator learning.

The global shift towards renewable energy necessitates the development of ultrahigh-voltage (UHV) AC transmission to bridge the gap between remote energy sources and urban demand. While UHV grids offer superior capacity and efficiency, their implementation is often hindered by corona-induced audible noise (AN) and radio interference (RI). Since these emissions must meet strict environmental compliance standards, accurate prediction is vital for the large-scale deployment of UHV infrastructure. Existing engineering practices often rely on empirical laws, in which fixed log-linear structures limit accuracy and extrapolation. Herein, we present a monotonicity-constrained graph symbolic discovery framework, Mono-GraphMD, which uncovers compact, interpretable laws for corona-induced AN and RI. The framework provides mechanistic insight into how nonlinear interactions among the surface gradient, bundle number and diameter govern high-field emissions and enables accurate predictions for both corona-cage data and multicountry real UHV lines with up to 16-bundle conductors. Unlike black-box models, the discovered closed-form laws are highly portable and interpretable, allowing for rapid predictions when applied to various scenarios, thereby facilitating the engineering design process.

Multi-agent systems (MAS) are increasingly used in healthcare to support complex decision-making through collaboration among specialized agents. Because these systems act as collective decision-makers, they raise challenges for trust, accountability, and human oversight. Existing approaches to trustworthy AI largely rely on explainability, but explainability alone is insufficient in multi-agent settings, as it does not enable care partners to challenge or correct system outputs. To address this limitation, Contestable AI (CAI) characterizes systems that support effective human challenge throughout the decision-making lifecycle by providing transparency, structured opportunities for intervention, and mechanisms for review, correction, or override. This position paper argues that contestability is a necessary design requirement for trustworthy multi-agent algorithmic care systems. We identify key limitations in current MAS and Explainable AI (XAI) research and present a human-in-the-loop framework that integrates structured argumentation and role-based contestation to preserve human agency, clinical responsibility, and trust in high-stakes care contexts.

Generating data from discrete distributions is important for a number of application domains including text, tabular data, and genomic data. Several groups have recently used random $k$-satisfiability ($k$-SAT) as a synthetic benchmark for new generative techniques. In this paper, we show that fundamental insights from the theory of random constraint satisfaction problems have observable implications (sometime contradicting intuition) on the behavior of generative techniques on such benchmarks. More precisely, we study the problem of generating a uniformly random solution of a given (random) $k$-SAT or $k$-XORSAT formula. Among other findings, we observe that: $(i)$~Continuous diffusions outperform masked discrete diffusions; $(ii)$~Learned diffusions can match the theoretical `ideal' accuracy; $(iii)$~Smart ordering of the variables can significantly improve accuracy, although not following popular heuristics.

When training a neural network for classification, the feature vectors of the training set are known to collapse to the vertices of a regular simplex, provided the dimension $d$ of the feature space and the number $n$ of classes satisfies $n\leq d+1$. This phenomenon is known as neural collapse. For other applications like language models, one instead takes $n\gg d$. Here, the neural collapse phenomenon still occurs, but with different emergent geometric figures. We characterize these geometric figures in the orthoplex regime where $d+2\leq n\leq 2d$. The techniques in our analysis primarily involve Radon's theorem and convexity.

As long-context language modeling becomes increasingly important, the cost of maintaining and attending to large Key/Value (KV) caches grows rapidly, becoming a major bottleneck in both training and inference. While prior works such as Multi-Query Attention (MQA) and Multi-Latent Attention (MLA) reduce memory by sharing or compressing KV features, they often trade off representation quality or incur runtime overhead. We propose Memory-Keyed Attention (MKA), a hierarchical attention mechanism that integrates multi-level KV caches (local, session, and long-term) and learns to route attention across them dynamically. We further introduce Route-Fused MKA (FastMKA), a broadcast-routed variant that fuses memory sources before attention computation for improved efficiency. Experiments on different sequence lengths show that FastMKA achieves a favorable accuracy-efficiency trade-off: comparable perplexity to MLA while achieving up to 5x faster training throughput and 1.8x lower evaluation latency. These results highlight MKA as a practical and extensible framework for efficient long-context attention.

Uncovering causal relationships is a fundamental problem across science and engineering. However, most existing causal discovery methods assume acyclicity and direct access to the system variables -- assumptions that fail to hold in many real-world settings. For instance, in genomics, cyclic regulatory networks are common, and measurements are often corrupted by instrumental noise. To address these challenges, we propose RECLAIM, a causal discovery framework that natively handles both cycles and measurement noise. RECLAIM learns the causal graph structure by maximizing the likelihood of the observed measurements via expectation-maximization (EM), using residual normalizing flows for tractable likelihood computation. We consider two measurement models: (i) Gaussian additive noise, and (ii) a linear measurement system with additive Gaussian noise. We provide theoretical consistency guarantees for both the settings. Experiments on synthetic data and real-world protein signaling datasets demonstrate the efficacy of the proposed method.

Social biases reflected in language are inherently shaped by cultural norms, which vary significantly across regions and lead to diverse manifestations of stereotypes. Existing evaluations of social bias in large language models (LLMs) for non-English contexts, however, often rely on translations of English benchmarks. Such benchmarks fail to reflect local cultural norms, including those found in Japanese. For instance, Western benchmarks may overlook Japan-specific stereotypes related to hierarchical relationships, regional dialects, or traditional gender roles. To address this limitation, we introduce Japanese cUlture adversarial BiAs benchmarK Under handcrafted creation (JUBAKU), a benchmark tailored to Japanese cultural contexts. JUBAKU uses adversarial construction to expose latent biases across ten distinct cultural categories. Unlike existing benchmarks, JUBAKU features dialogue scenarios hand-crafted by native Japanese annotators, specifically designed to trigger and reveal latent social biases in Japanese LLMs. We evaluated nine Japanese LLMs on JUBAKU and three others adapted from English benchmarks. All models clearly exhibited biases on JUBAKU, performing below the random baseline of 50% with an average accuracy of 23% (ranging from 13% to 33%), despite higher accuracy on the other benchmarks. Human annotators achieved 91% accuracy in identifying unbiased responses, confirming JUBAKU's reliability and its adversarial nature to LLMs.

The prevailing approach to improving large language model (LLM) reasoning has centered on expanding context windows, implicitly assuming that more tokens yield better performance. However, empirical evidence - including the "lost in the middle" effect and long-distance relational degradation - demonstrates that contextual space exhibits structural gradients, salience asymmetries, and entropy accumulation under transformer architectures. We introduce Context Cartography, a formal framework for the deliberate governance of contextual space. We define a tripartite zonal model partitioning the informational universe into black fog (unobserved), gray fog (stored memory), and the visible field (active reasoning surface), and formalize seven cartographic operators - reconnaissance, selection, simplification, aggregation, projection, displacement, and layering - as transformations governing information transitions between and within zones. The operators are derived from a systematic coverage analysis of all non-trivial zone transformations and are organized by transformation type (what the operator does) and zone scope (where it applies). We ground the framework in the salience geometry of transformer attention, characterizing cartographic operators as necessary compensations for linear prefix memory, append-only state, and entropy accumulation under expanding context. An analysis of four contemporary systems (Claude Code, Letta, MemOS, and OpenViking) provides interpretive evidence that these operators are converging independently across the industry. We derive testable predictions from the framework - including operator-specific ablation hypotheses - and propose a diagnostic benchmark for empirical validation.

Automating large-scale manufacturing in domains like timber construction requires multi-robot systems to manage tightly coupled spatiotemporal constraints, such as collision avoidance and process-driven deadlines. This paper introduces LASER (Level-based Asynchronous Scheduling and Execution Regime), a complete framework for scheduling and executing complex assembly tasks, demonstrated on a screw-press gluing application for timber slab manufacturing. Our central contribution is to integrate a barrier-based mechanism into a constraint programming (CP) scheduling formulation that partitions tasks into spatiotemporally disjoint sets, which we define as levels. This structure enables robots to execute tasks in parallel and asynchronously within a level, synchronizing only at level barriers, which guarantees collision-free operation by construction and provides robustness to timing uncertainties. To solve this formulation for large problems, we propose two specialized algorithms: an iterative temporal-relaxation approach for heterogeneous task sequences and a bi-level decomposition for homogeneous tasks that balances workload. We validate the LASER framework by fabricating a full-scale 2.4m x 6m timber slab with a two-robot system mounted on parallel linear tracks, successfully coordinating 108 subroutines and 352 screws under tight adhesive time windows. Computational studies show our method scales steadily with size compared to a monolithic approach.

The potential of Large Language Models (LLMs) to provide harmful information remains a significant concern due to the vast breadth of illegal queries they may encounter. Unfortunately, existing benchmarks only focus on a handful types of illegal activities, and are not grounded in legal works. In this work, we introduce an ontology of crime-related concepts grounded in the legal frameworks of Model Panel Code, which serves as an influential reference for criminal law and has been adopted by many U.S. states, and instantiated using Californian Law. This structured knowledge forms the foundation for LJ-Bench, the first comprehensive benchmark designed to evaluate LLM robustness against a wide range of illegal activities. Spanning 76 distinct crime types organized taxonomically, LJ-Bench enables systematic assessment of diverse attacks, revealing valuable insights into LLM vulnerabilities across various crime categories: LLMs exhibit heightened susceptibility to attacks targeting societal harm rather than those directly impacting individuals. Our benchmark aims to facilitate the development of more robust and trustworthy LLMs. The LJ-Bench benchmark and LJ-Ontology, along with experiments implementation for reproducibility are publicly available at https://github.com/AndreaTseng/LJ-Bench.

Large language models (LLMs) are now widely used as judges, yet their decisions can change under presentation choices that should be irrelevant. We study one such source of instability: candidate-order sensitivity in listwise factuality evaluation, where several answers can look similarly polished while differing sharply in hallucination risk. We introduce PCFJudge, an inference-time method that reruns the same factuality-first listwise prompt over multiple orderings of the same candidate set and aggregates the resulting scores, ranks, and uncertainty signals into a single consensus decision. On RewardBench 2 Factuality, PCFJudge improves over direct judging by up to 7 absolute points. Development ablations show that the dominant gain comes from permutation consensus itself rather than from heavier arbitration layers. These results suggest that a meaningful share of factuality-judging error arises from order instability, and that averaging over this nuisance variation is a simple and effective way to make LLM evaluation more reliable.

We want a multi-robot team to complete complex tasks in minimum time where the locations of task-relevant objects are not known. Effective task completion requires reasoning over long horizons about the likely locations of task-relevant objects, how individual actions contribute to overall progress, and how to coordinate team efforts. Planning in this setting is extremely challenging: even when task-relevant information is partially known, coordinating which robot performs which action and when is difficult, and uncertainty introduces a multiplicity of possible outcomes for each action, which further complicates long-horizon decision-making and coordination. To address this, we propose a multi-robot planning abstraction that integrates learning to estimate uncertain aspects of the environment with model-based planning for long-horizon coordination. We demonstrate the efficient multi-stage task planning of our approach for 1, 2, and 3 robot teams over competitive baselines in large ProcTHOR household environments. Additionally, we demonstrate the effectiveness of our approach with a team of two LoCoBot mobile robots in real household settings.

Behavior cloning is a fundamental paradigm in machine learning, enabling policy learning from expert demonstrations across robotics, autonomous driving, and generative models. Autoregressive models like transformer have proven remarkably effective, from large language models (LLMs) to vision-language-action systems (VLAs). However, applying autoregressive models to continuous control requires discretizing actions through quantization, a practice widely adopted yet poorly understood theoretically. This paper provides theoretical foundations for this practice. We analyze how quantization error propagates along the horizon and interacts with statistical sample complexity. We show that behavior cloning with quantized actions and log-loss achieves optimal sample complexity, matching existing lower bounds, and incurs only polynomial horizon dependence on quantization error, provided the dynamics are stable and the policy satisfies a probabilistic smoothness condition. We further characterize when different quantization schemes satisfy or violate these requirements, and propose a model-based augmentation that provably improves the error bound without requiring policy smoothness. Finally, we establish fundamental limits that jointly capture the effects of quantization error and statistical complexity.

Industrial process control demands policies that are interpretable and auditable, requirements that black-box neural policies struggle to meet. We study an LLM-driven heuristic synthesis framework for hot steel rolling, in which a language model iteratively proposes and refines human-readable Python controllers using rich behavioral feedback from a physics-based simulator. The framework combines structured strategic ideation, executable code generation, and per-component feedback across diverse operating conditions to search over control logic for height reduction, interpass time, and rolling velocity. Our first contribution is an auditable controller-synthesis pipeline for industrial process control. The generated controllers are explicit programs accessible to expert review, and we pair them with an automated audit pipeline that formally verifies key safety and monotonicity properties for the best synthesized heuristic. Our second contribution is a principled budget allocation strategy for LLM-driven heuristic search: we show that Luby-style universal restarts -- originally developed for randomized algorithms -- transfer directly to this setting, eliminating the need for problem-specific budget tuning. A single 160-iteration Luby campaign approaches the hindsight-optimal budget allocation derived from 52 ad-hoc runs totalling 730 iterations.

Rapid identification of outbreaks in hospitals is essential for controlling pathogens with epidemic potential. Although whole genome sequencing (WGS) remains the gold standard in outbreak investigations, its substantial costs and turnaround times limit its feasibility for routine surveillance, especially in less-equipped facilities. We explore three modalities as rapid alternatives: matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) mass spectrometry, antimicrobial resistance (AR) patterns, and electronic health records (EHR). We present a machine learning approach that learns discriminative features from these modalities to support outbreak detection. Multi-species evaluation shows that the integration of these modalities can boost outbreak detection performance. We also propose a tiered surveillance paradigm that can reduce the need for WGS through these alternative modalities. Further analysis of EHR information identifies potentially high-risk contamination routes linked to specific clinical procedures, notably those involving invasive equipment and high-frequency workflows, providing infection prevention teams with actionable targets for proactive risk mitigation

Requirements engineering in Industry 4.0 faces critical challenges with heterogeneous, unstructured documentation spanning technical specifications, supplier lists, and compliance standards. While retrieval-augmented generation (RAG) shows promise for knowledge-intensive tasks, no prior work has evaluated RAG on authentic industrial RE workflows using comprehensive production-grade performance metrics. This paper presents a comprehensive empirical evaluation of RAG for industrial requirements engineering automation using authentic automotive manufacturing documentation comprising 669 requirements across four specification standards (MBN 9666-1, MBN 9666-2, BQF 9666-5, MBN 9666-9) spanning 2015-2023, plus 49 supplier qualifications with extensive supporting documentation. Through controlled comparisons with BERT-based and ungrounded LLM approaches, the framework achieves 98.2% extraction accuracy with complete traceability, outperforming baselines by 24.4% and 19.6%, respectively. Hybrid semantic-lexical retrieval achieves MRR of 0.847. Expert quality assessment averaged 4.32/5.0 across five dimensions. The evaluation demonstrates 83% reduction in manual analysis time and 47% cost savings through multi-provider LLM orchestration. Ablation studies quantify individual component contributions. Longitudinal analysis reveals a 55% reduction in requirement volume coupled with 1,800% increase in IT security focus, identifying 10 legacy suppliers (20.4%) requiring requalification, representing potential $2.3M in avoided contract penalties.

Generative AI platforms (Google AI Studio, OpenAI, Anthropic) provide infrastructures (APIs, models) that are transforming the application development ecosystem. Recent literature distinguishes three generations of business models: a first generation modeled on cloud computing (pay-per-use), a second characterized by diversification (freemium, subscriptions), and a third, emerging generation exploring multi-layer market architectures with revenue-sharing mechanisms. Despite these advances, current models impose a financial barrier to entry for developers, limiting innovation and excluding actors from emerging economies. This paper proposes and analyzes an original model, "Revenue-Sharing as Infrastructure" (RSI), where the platform offers its AI infrastructure for free and takes a percentage of the revenues generated by developers applications. This model reverses the traditional upstream payment logic and mobilizes concepts of value co-creation, incentive mechanisms, and multi-layer market architecture to build an original theoretical framework. A detailed comparative analysis shows that the RSI model lowers entry barriers for developers, aligns stakeholder interests, and could stimulate innovation in the ecosystem. Beyond its economic relevance, RSI has a major societal dimension: by enabling developers without initial capital to participate in the digital economy, it could unlock the "latent jobs dividend" in low-income countries, where mobile penetration reaches 84%, and help address local challenges in health, agriculture, and services. Finally, we discuss the conditions of feasibility and strategic implications for platforms and developers.

We find that models report highest confidence precisely when they are fabricating. Across four model families (OLMo-3, Llama-3.1, Qwen3, Mistral), self-reported confidence inversely correlates with accuracy, with AUC ranging from 0.28 to 0.36 where 0.5 is random guessing. We prove, under explicit formal assumptions, that this is not a capability gap but an observational one. Under text-only observation, where a supervisor sees only the model's output text, no monitoring system can reliably distinguish honest model outputs from plausible fabrications. We prove two results: first, that any policy conditioning only on the query cannot satisfy epistemic honesty across ambiguous world states; second, that no learning algorithm optimizing reward from a text-only supervisor can converge to honest behavior when the supervisor's observations are identical for both grounded and fabricated responses. Within our formal model, these impossibilities hold regardless of model scale or training procedure, including RLHF and instruction tuning. We construct a tensor interface that escapes the impossibility by exporting computational byproducts (per-token entropy and log-probability distributions) that are structurally coupled to correctness under standard training. Per-token entropy achieves pooled AUC 0.757, outperforming all text baselines by 2.5--3.9 percentage points at every budget level tested (10\%, 20\%, 30\%). The entropy signal generalizes across architectures (Spearman $ρ= 0.762$). The core contribution is a cost surface where the empirical mapping from verification budget (fraction of queries receiving expensive checks) to detection accuracy for each judge strategy is a practical lookup for system builders deciding how to allocate verification resources. The contribution is the map. The territory is the system you are building.

The notion of software entropy is often invoked to describe the tendency of software systems to become increasingly disordered as they evolve, yet existing approaches to quantify it are largely heuristic. In this work we introduce a formal definition of software entropy grounded in statistical mechanics, interpreting test suites as executable specifications, that is, as macroscopic constraints on the space of possible program implementations. Within this framework, mutation analysis provides a practical approximation of the locally accessible microstate space, allowing entropy-related quantities to be estimated empirically. We propose metrics that quantify how test suites restrict program space, including an information-weighted measure of the distribution of constraint power across tests. Applying these ideas to a real-world project, we show how test suites reduce software entropy and how information weights reveal structural differences in the contribution of individual tests that traditional metrics such as code coverage fail to capture.

Preconditioned adaptive methods have gained significant attention for training deep neural networks, as they capture rich curvature information of the loss landscape . The central challenge in this field lies in balancing preconditioning effectiveness with computational efficiency of implementing the preconditioner. Among recent advances, \textsc{Muon} stands out by using Newton-Schulz iteration to obtain preconditioned updates without explicitly constructing the preconditioning matrix. Despite its advantages, the efficiency of \textsc{Muon} still leaves room for further improvement. In this paper, we introduce \textsc{RMNP} (Row Momentum Normalized Preconditioning), an optimizer that replaces Newton-Schulz iteration with a simple row-wise $\ell_2$ normalization operation, motivated by the empirically observed diagonal block structure of the Transformer layerwise Hessian. This substitution reduces the per-iteration computational complexity from $\mathcal{O}(mn\cdot\min(m,n))$ to $\mathcal{O}(mn)$ for an $m\times n$ weight matrix while maintaining comparable optimization performance. Theoretically, we establish convergence guarantees for \textsc{RMNP} in the non-convex setting that match recent results for \textsc{Muon} optimizers, achieving the information-theoretic minimax optimal complexity. Extensive experiments on large language model pretraining show that \textsc{RMNP} delivers competitive optimization performance compared with \textsc{Muon} while substantially reducing preconditioning wall-clock time. Our code is available at \href{https://anonymous.4open.science/r/RMNP-E8E1/}{this link}.

Policy gradient computes a backward pass for every sample, even though the backward pass is expensive and most samples carry little learning value. The Delightful Policy Gradient (DG) provides a forward-pass signal of learning value: \emph{delight}, the product of advantage and surprisal (negative log-probability). We introduce the \emph{Kondo gate}, which compares delight against a compute price and pays for a backward pass only when the sample is worth it, thereby tracing a quality--cost Pareto frontier. In bandits, zero-price gating preserves useful gradient signal while removing perpendicular noise, and delight is a more reliable screening signal than additive combinations of value and surprise. On MNIST and transformer token reversal, the Kondo gate skips most backward passes while retaining nearly all of DG's learning quality, with gains that grow as problems get harder and backward passes become more expensive. Because the gate tolerates approximate delight, a cheap forward pass can screen samples before expensive backpropagation, suggesting a speculative-decoding-for-training paradigm.

Distributed reinforcement learning trains on data from stale, buggy, or mismatched actors, producing actions with high surprisal (negative log-probability) under the learner's policy. The core difficulty is not surprising data per se, but \emph{negative learning from surprising data}. High-surprisal failures can dominate the update direction despite carrying little useful signal, while high-surprisal successes reveal opportunities the current policy would otherwise miss. The \textit{Delightful Policy Gradient} (DG) separates these cases by gating each update with delight, the product of advantage and surprisal, suppressing rare failures and amplifying rare successes without behavior probabilities. Under contaminated sampling, the cosine similarity between the standard policy gradient and the true gradient collapses, while DG's grows as the policy improves. No sign-blind reweighting, including exact importance sampling, can reproduce this effect. On MNIST with simulated staleness, DG without off-policy correction outperforms importance-weighted PG with exact behavior probabilities. On a transformer sequence task with staleness, actor bugs, reward corruption, and rare discovery, DG achieves roughly $10{\times}$ lower error. When all four frictions act simultaneously, its compute advantage is order-of-magnitude and grows with task complexity.

Simulation-based inference (SBI) with neural networks has accelerated and transformed cognitive modeling workflows. SBI enables modelers to fit complex models that were previously difficult or impossible to estimate, while also allowing rapid estimation across large numbers of datasets. However, the utility of SBI for iterating over varying modeling assumptions remains limited: changing parameterizations, generative functions, priors, and design variables all necessitate model retraining and hence diminish the benefits of amortization. To address these issues, we pilot a meta-amortized framework for cognitive modeling which we nickname the CogFormer. Our framework trains a transformer-based architecture that remains valid across a combinatorial number of structurally similar models, allowing for changing data types, parameters, design matrices, and sample sizes. We present promising quantitative results across families of decision-making models for binary, multi-alternative, and continuous responses. Our evaluation suggests that CogFormer can accurately estimate parameters across model families with a minimal amortization offset, making it a potentially powerful engine that catalyzes cognitive modeling workflows.

Large Language Models (LLMs) offer significant potential for delivering health information. However, their reliability in low-resource contexts remains uncertain. This study evaluates GPT-4, Gemini Pro, Llama~3, and Mistral-7B on health crisis-related enquiries concerning COVID-19, dengue, the Nipah virus, and Chikungunya in the low-resource context of Bangladesh. We constructed a question--answer dataset from authoritative sources and assessed model outputs through semantic similarity, expert-model cross-evaluation, and Natural Language Inference (NLI). Findings highlight both the strengths and limitations of LLMs in representing epidemiological history and health crisis knowledge, underscoring their promise and risks for informing policy in resource-constrained environments.

LLM-reranking is limited by the top-k documents retrieved by vector similarity, which neither enables contextual query-document token interactions nor captures multimodal relevance distributions. While LLM query reformulation attempts to improve recall by generating improved or additional queries, it is still followed by vector similarity retrieval. We thus propose to address these top-k retrieval stage failures by introducing ReBOL, which 1) uses LLM query reformulations to initialize a multimodal Bayesian Optimization (BO) posterior over document relevance, and 2) iteratively acquires document batches for LLM query-document relevance scoring followed by posterior updates to optimize relevance. After exploring query reformulation and document batch diversification techniques, we evaluate ReBOL against LLM reranker baselines on five BEIR datasets and using two LLMs (Gemini-2.5-Flash-Lite, GPT-5.2). ReBOL consistently achieves higher recall and competitive rankings, for example compared to the best LLM reranker on the Robust04 dataset with 46.5% vs. 35.0% recall@100 and 63.6% vs. 61.2% NDCG@10. We also show that ReBOL can achieve comparable latency to LLM rerankers.

Cloud and big data workloads are increasingly distributing data across multiple cloud providers and regions for rapid decision-making and analytics. Traditional transfer tools are typically specialized for a single paradigm, either stream replication or bulk transfer. This specialization forces users to deploy and manage separate systems with different configurations for each transfer pattern. This paper presents SkyHOST (Hybrid Object and Stream Transfer), a unified data movement architecture built upon the Skyplane framework to bridge the gap between bulk object transfer and streaming workloads through a single control plane and CLI. SkyHOST manages URI-based routing to automatically select the appropriate transfer mechanism, supporting both structured data for record-level ingestion and chunk-based transfer for large binary objects. We demonstrate, through an environmental monitoring use case and empirical evaluation, that SkyHOST provides operational simplicity by consolidating heterogeneous data movement patterns under a single control plane while achieving competitive throughput for cross-region transfers.

Language models often lack grounded reasoning capabilities in specialized domains where training data is scarce but bespoke systems excel. We introduce a general framework for distilling expert system reasoning into natural language chain-of-thought explanations, enabling compact models to acquire domain expertise and the ability to generate faithful, grounded explanations. Rather than distilling only final outputs, we capture the full reasoning process, transforming opaque expert computations into transparent, step-by-step explanations. We demonstrate this approach in chess, a canonical reasoning domain where language models continue to underperform. Our 4B parameter model, C1, advances from a near-zero baseline to 48.1% accuracy, outperforming all open-source models and most frontier proprietary systems. Notably, C1 surpasses its distillation teacher and generates solutions in two orders of magnitude fewer tokens than baselines. Unlike prior neural chess approaches that predict only best moves, C1 generates explainable solutions revealing strategic reasoning. Our pipeline combines supervised fine-tuning and reinforcement learning with theme-balanced data sampling for comprehensive tactical coverage. Master Distillation demonstrates how to inject expert-level knowledge into compact models for under-optimized domains, offering a recipe for unlocking RLVR where LLMs lack sufficient base capabilities.

Language models are increasingly being trained to "reason" before answering users' queries, outputting hundreds or even thousands of tokens worth of deliberation before their final answer. While the main intention of reasoning is to improve models' ability to arrive at a correct answer, we argue that these models should be assessed for the legibility of their reasoning traces in addition to the correctness of their final answers. In this paper, we evaluate 90k traces from 12 Reasoning Language Models (RLMs) for the quality of their reasoning traces. We introduce the concept of transfer utility, which assesses how useful an RLM's reasoning traces are for guiding a weaker, non-reasoning model toward arriving at the correct answer. We find that the reasoning traces of the highest-performing models rank among the lowest for legibility. Furthermore, we uncover tensions between efficiency-based measurements of legibility (such as trace length) and transfer utility. These tensions establish a legibility Pareto frontier, and we demonstrate that an RLM's ability to output highly legible traces can be a task- and audience-dependent goal. Crucially, we find that reward models used to train RLMs do not intrinsically reward legibility. Together, these metrics and the findings they surface chart a path towards scaffolding reasoning traces for a multi-agent future.

We study the effects of center initialization on the performance of a family of distributed gradient-based clustering algorithms introduced in [1], that work over connected networks of users. In the considered scenario, each user contains a local dataset and communicates only with its immediate neighbours, with the aim of finding a global clustering of the joint data. We perform extensive numerical experiments, evaluating the effects of center initialization on the performance of our family of methods, demonstrating that our methods are more resilient to the effects of initialization, compared to centralized gradient clustering [2]. Next, inspired by the $K$-means++ initialization [3], we propose a novel distributed center initialization scheme, which is shown to improve the performance of our methods, compared to the baseline random initialization.

Probabilistic Logic Programming (PLP) languages, like ProbLog, naturally support reasoning under uncertainty, while maintaining a declarative and interpretable framework. Meanwhile, counterfactual reasoning (i.e., answering ``what if'' questions) is critical for ensuring AI systems are robust and trustworthy; however, integrating this capability into PLP can be computationally prohibitive and unstable in accuracy. This paper addresses this challenge, by proposing an efficient program transformation for counterfactuals as Single World Intervention Programs (SWIPs) in ProbLog. By systematically splitting ProbLog clauses to observed and fixed components relevant to a counterfactual, we create a transformed program that (1) does not asymptotically exceed the computational complexity of existing methods, and is strictly smaller in common cases, and (2) reduces counterfactual reasoning to marginal inference over a simpler program. We formally prove the correctness of our approach, which relies on a weaker set independence assumptions and is consistent with conditional independencies, showing the resulting marginal probabilities match the counterfactual distributions of the underlying Structural Causal Model in wide domains. Our method achieves a 35\% reduction in inference time versus existing methods in extensive experiments. This work makes complex counterfactual reasoning more computationally tractable and reliable, providing a crucial step towards developing more robust and explainable AI systems. The code is at https://github.com/EVIEHub/swip.

Modern cloud inference creates a two sided privacy problem where users reveal sensitive inputs to providers, while providers must execute proprietary model weights inside potentially leaky execution environments. Fully homomorphic encryption (FHE) offers cryptographic guarantees but remains prohibitively expensive for modern architectures. We argue that progress requires co-design where specializing FHE schemes/compilers for the static structure of inference circuits, while simultaneously constraining inference architectures to reduce dominant homomorphic cost drivers. We outline a meet in the middle agenda and concrete optimization targets on both axes.

The development of clinical natural language processing (NLP) systems is severely hampered by the sensitive nature of medical records, which restricts data sharing under stringent privacy regulations, particularly in France and the broader European Union. To address this gap, we introduce PARHAF, a large open-source corpus of clinical documents in French. PARHAF comprises expert-authored clinical reports describing realistic yet entirely fictitious patient cases, making it anonymous and freely shareable by design. The corpus was developed using a structured protocol that combined clinician expertise with epidemiological guidance from the French National Health Data System (SNDS), ensuring broad clinical coverage. A total of 104 medical residents across 18 specialties authored and peer-reviewed the reports following predefined clinical scenarios and document templates. The corpus contains 7394 clinical reports covering 5009 patient cases across a wide range of medical and surgical specialties. It includes a general-purpose component designed to approximate real-world hospitalization distributions, and four specialized subsets that support information-extraction use cases in oncology, infectious diseases, and diagnostic coding. Documents are released under a CC-BY open license, with a portion temporarily embargoed to enable future benchmarking under controlled conditions. PARHAF provides a valuable resource for training and evaluating French clinical language models in a fully privacy-preserving setting, and establishes a replicable methodology for building shareable synthetic clinical corpora in other languages and health systems.

Large Language Models (LLMs) have achieved remarkable success across diverse applications, yet their deployment remains challenging due to substantial computational costs, memory requirements, and energy consumption. Recent empirical studies have demonstrated that no single efficiency technique is universally optimal; instead, the effectiveness of methods such as efficient attention mechanisms, mixture-of-experts (MoE), parameter-efficient fine-tuning, and quantization varies significantly depending on task characteristics, resource constraints, and model scales. Building upon these insights, we propose AE-LLM, a unified framework that automatically selects and combines optimal efficiency techniques tailored to specific deployment scenarios. Our approach introduces a multi-objective optimization framework that jointly considers accuracy, latency, memory footprint, and energy consumption, while accounting for hardware constraints and task requirements. We develop an efficient search algorithm that explores the combinatorial space of efficiency techniques across architecture, fine-tuning, and inference stages, identifying Pareto-optimal configurations. Extensive experiments across 15 models (0.5B-70B parameters) and 10 diverse tasks demonstrate that AE-LLM achieves an average of $2.8\times$ improvement in efficiency metrics while maintaining competitive accuracy (within 1.2\% of baseline), compared to static efficiency configurations. Furthermore, our framework generalizes effectively to vision-language models, achieving similar efficiency gains. Our contributions provide practitioners with an automated tool for navigating the complex trade-off landscape of LLM efficiency optimization.

Electricity theft, or non-technical loss (NTL), presents a persistent threat to global power systems, driving significant financial deficits and compromising grid stability. Conventional detection methodologies, predominantly reactive and meter-centric, often fail to capture the complex spatio-temporal dynamics and behavioral patterns associated with fraudulent consumption. This study introduces a novel AI-driven Grid Intelligence Framework that fuses Time-Series Anomaly Detection, Supervised Machine Learning, and Graph Neural Networks (GNN) to identify theft with high precision in imbalanced datasets. Leveraging an enriched feature set, including rolling averages, voltage drop estimates, and a critical Grid Imbalance Index, the methodology employs a Long Short-Term Memory (LSTM) autoencoder for temporal anomaly scoring, a Random Forest classifier for tabular feature discrimination, and a GNN to model spatial dependencies across the distribution network. Experimental validation demonstrates that while standalone anomaly detection yields a low theft F1-score of 0.20, the proposed hybrid fusion achieves an overall accuracy of 93.7%. By calibrating decision thresholds via precision-recall analysis, the system attains a balanced theft precision of 0.55 and recall of 0.50, effectively mitigating the false positives inherent in single-model approaches. These results confirm that integrating topological grid awareness with temporal and supervised analytics provides a scalable, risk-based solution for proactive electricity theft detection and enhanced smart grid reliability.

This work presents COmPOSER, an open-source, end-to-end framework for RF/mm-wave design automation that translates target specifications into optimized circuits with layouts. It unifies schematic synthesis, layout generation for actives and passives, and placement/routing, incorporating physics-based equations and machine-learning-driven electromagnetic models. Based on post-layout validation on multiple LNAs and PAs operating at up to 60GHz in a commercial 65nm process-kit, COmPOSER meets performance targets, comparable to expert manual designs, while delivering a 100-300x productivity gain.

Learning another language can be a highly emotional process, typically characterized by numerous frustrations and triumphs, big and small. For most learners, language learning does not follow a linear, predictable path, its zigzag course shaped by motivational (or demotivating) variables such as personal characteristics, teacher/peer relationships, learning materials, and dreams of a future L2 (second language) self. While some aspects of language learning (reading, grammar) are relatively mechanical, others can be stressful and unpredictable, especially conversing in the target language. That experience necessitates not only knowledge of structure and lexis, but also the ability to use the language in ways that are appropriate to the social and cultural context. A new opportunity to practice conversational abilities has arrived through the availability of AI chatbots, with both advantages (responsive, non-judgmental) and drawbacks (emotionally void, culturally biased). This column explores aspects of emotion as they arise in technology use and in particular how automatic emotion recognition and simulated human responsiveness in AI systems interface with language learning and the development of pragmatic and interactional competence. Emotion AI, the algorithmically driven interpretation of users' affective signals, has been seen as enabling greater personalized learning, adapting to perceived learner cognitive and emotional states. Others warn of emotional manipulation and inappropriate and ineffective user profiling

Conservation laws are fundamental to understanding dynamical systems, but discovering them from data remains challenging due to parameter variation, non-polynomial invariants, local minima, and false positives on chaotic systems. We introduce NGCG, a neural-symbolic pipeline that decouples dynamics learning from invariant discovery and systematically addresses these challenges. A multi-restart variance minimiser learns a near-constant latent representation; system-specific symbolic extraction (polynomial Lasso, log-basis Lasso, explicit PDE candidates, and PySR) yields closed-form expressions; a strict constancy gate and diversity filter eliminate spurious laws. On a benchmark of nine diverse systems including Hamiltonian and dissipative ODEs, chaos, and PDEs, NGCG achieves consistent discovery (DR=1.0, FDR=0.0, F1=1.0) on all four systems with true conservation laws, with constancy two to three orders of magnitude lower than the best baseline. It is the only method that succeeds on the Lotka--Volterra system, and it correctly outputs no law on all five systems without invariants. Extensive experiments demonstrate robustness to noise ($σ= 0.1$), sample efficiency (50--100 trajectories), insensitivity to hyperparameters, and runtime under one minute per system. A Pareto analysis shows that the method provides a range of candidate expressions, allowing users to trade complexity for constancy. NGCG achieves strong performance relative to prior methods for data-driven conservation-law discovery, combining high accuracy with interpretability.

The rapid growth of the text-to-image (T2I) community has fostered a thriving online ecosystem of expert models, which are variants of pretrained diffusion models specialized for diverse generative abilities. Yet, existing model merging methods remain limited in fully leveraging abundant online expert resources and still struggle to meet diverse in-the-wild user needs. We present DiffGraph, a novel agent-driven graph-based model merging framework, which automatically harnesses online experts and flexibly merges them for diverse user needs. Our DiffGraph constructs a scalable graph and organizes ever-expanding online experts within it through node registration and calibration. Then, DiffGraph dynamically activates specific subgraphs based on user needs, enabling flexible combinations of different experts to achieve user-desired generation. Extensive experiments show the efficacy of our method.

The governing equations of stochastic dynamical systems often become cost-prohibitive for numerical simulation at large scales. Surrogate models of the governing equations, learned from data of the high-fidelity system, are routinely used to predict key observables with greater efficiency. However, standard choices of loss function for learning the surrogate model fail to provide error guarantees in path-dependent observables, such as reaction rates of molecular dynamical systems. This paper introduces an error bound for path-space observables and employs it as a novel variational loss for the goal-oriented learning of a stochastic dynamical system. We show the error bound holds for a broad class of observables, including mean first hitting times on unbounded time domains. We derive an analytical gradient of the goal-oriented loss function by leveraging the formula for Frechet derivatives of expected path functionals, which remains tractable for implementation in stochastic gradient descent schemes. We demonstrate that surrogate models of overdamped Langevin systems developed via goal-oriented learning achieve improved accuracy in predicting the statistics of a first hitting time observable and robustness to distributional shift in the data.

Masked Diffusion Language Models (MDLMs) have emerged as a compelling non-autoregressive alternative to standard large language models; however, their application to morphologically rich languages remains limited. In this paper, we introduce $\textit{Diffutron}$, a masked diffusion language model specifically designed for Turkish. Our approach leverages a resource-efficient training pipeline, starting with LoRA-based continual pre-training of a multilingual encoder on a large-scale corpus. To enable generative capabilities, we employ a progressive instruction-tuning strategy, sequentially adapting the model on general and task-specific instruction sets. Experimental results across comprehensive benchmarks demonstrate that, despite its compact size, our model achieves competitive performance compared to existing multi-billion-parameter baselines. These findings validate the effectiveness of masked diffusion modeling combined with multi-stage tuning for non-autoregressive text generation in Turkish.

Assigning relevance scores to the input features of a machine learning model enables to measure the contributions of the features in achieving a correct outcome. It is regarded as one of the approaches towards developing explainable models. For biomedical assignments, this is very useful for medical experts to comprehend machine-based decisions. In the analysis of electro cardiogram (ECG) signals, in particular, understanding which of the electrocardiogram samples or features contributed most for a given decision amounts to understanding the underlying cardiac phases or conditions the machine tries to explain. For the computation of relevance scores, determining the proper baseline is important. Moreover, the scores should have a distribution which is at once intuitive to interpret and easy to associate with the underline cardiac reality. The purpose of this work is to achieve these goals. Specifically, we propose a shift-invariant baseline which has a physical significance in the analysis as well as interpretation of electrocardiogram measurements. Moreover, we aggregate significance scores in such a way that they can be mapped to cardiac phases. We demonstrate our approach by inferring physical exertion from cardiac exertion using a residual network. We show that the ECG samples which achieved the highest relevance scores (and, therefore, which contributed most to the accurate recognition of the physical exertion) are those associated with the P and T waves. Index Terms Attribution, baseline, cardiovascular diseases, electrocardiogram, activity recognition, machine learning

Reinforcement learning from human feedback (RLHF) replaces hard-to-specify rewards with pairwise trajectory preferences, yet regret-oriented theory often assumes that preference labels are generated consistently from a single ground-truth objective. In practical RLHF systems, however, feedback is typically \emph{multi-source} (annotators, experts, reward models, heuristics) and can exhibit systematic, persistent mismatches due to subjectivity, expertise variation, and annotation/modeling artifacts. We study episodic RL from \emph{multi-source imperfect preferences} through a cumulative imperfection budget: for each source, the total deviation of its preference probabilities from an ideal oracle is at most $ω$ over $K$ episodes. We propose a unified algorithm with regret $\tilde{O}(\sqrt{K/M}+ω)$, which exhibits a best-of-both-regimes behavior: it achieves $M$-dependent statistical gains when imperfection is small (where $M$ is the number of sources), while remaining robust with unavoidable additive dependence on $ω$ when imperfection is large. We complement this with a lower bound $\tildeΩ(\max\{\sqrt{K/M},ω\})$, which captures the best possible improvement with respect to $M$ and the unavoidable dependence on $ω$, and a counterexample showing that naïvely treating imperfect feedback as as oracle-consistent can incur regret as large as $\tildeΩ(\min\{ω\sqrt{K},K\})$. Technically, our approach involves imperfection-adaptive weighted comparison learning, value-targeted transition estimation to control hidden feedback-induced distribution shift, and sub-importance sampling to keep the weighted objectives analyzable, yielding regret guarantees that quantify when multi-source feedback provably improves RLHF and how cumulative imperfection fundamentally limits it.

Longitudinal neuroimaging is essential for modeling disease progression in Alzheimer's disease (AD), yet irregular sampling and missing visits pose substantial challenges for learning reliable temporal representations. To address this challenge, we propose SDE-HGNN, a stochastic differential equation (SDE)-driven spatio-temporal hypergraph neural network for irregular longitudinal fMRI connectome modeling. The framework first employs an SDE-based reconstruction module to recover continuous latent trajectories from irregular observations. Based on these reconstructed representations, dynamic hypergraphs are constructed to capture higher-order interactions among brain regions over time. To further model temporal evolution, hypergraph convolution parameters evolve through SDE-controlled recurrent dynamics conditioned on inter-scan intervals, enabling disease-stage-adaptive connectivity modeling. We also incorporate a sparsity-based importance learning mechanism to identify salient brain regions and discriminative connectivity patterns. Extensive experiments on the OASIS-3 and ADNI cohorts demonstrate consistent improvements over state-of-the-art graph and hypergraph baselines in AD progression prediction. The source code is available at https://anonymous.4open.science/r/SDE-HGNN-017F.

A number of scientific conferences and journals have recently enacted policies that prohibit LLM usage by peer reviewers, except for polishing, paraphrasing, and grammar correction of otherwise human-written reviews. But, are these policies enforceable? To answer this question, we assemble a dataset of peer reviews simulating multiple levels of human-AI collaboration, and evaluate five state-of-the-art detectors, including two commercial systems. Our analysis shows that all detectors misclassify a non-trivial fraction of LLM-polished reviews as AI-generated, thereby risking false accusations of academic misconduct. We further investigate whether peer-review-specific signals, including access to the paper manuscript and the constrained domain of scientific writing, can be leveraged to improve detection. While incorporating such signals yields measurable gains in some settings, we identify limitations in each approach and find that none meets the accuracy standards required for identifying AI use in peer reviews. Importantly, our results suggest that recent public estimates of AI use in peer reviews through the use of AI-text detectors should be interpreted with caution, as current detectors misclassify mixed reviews (collaborative human-AI outputs) as fully AI generated, potentially overstating the extent of policy violations.

Tool-augmented Large Language Models (TaLLMs) extend LLMs with the ability to invoke external tools, enabling them to interact with real-world environments. However, a major limitation in deploying TaLLMs in sensitive applications such as customer service and business process automation is a lack of reliable compliance with domain-specific operational policies regarding tool-use and agent behavior. Current approaches merely steer LLMs to adhere to policies by including policy descriptions in the LLM context, but these provide no guarantees that policy violations will be prevented. In this paper, we introduce an SMT solver-aided framework to enforce tool-use policy compliance in TaLLM agents. Specifically, we use an LLM-assisted, human-guided approach to translate natural-language-specified tool-use policies into formal logic (SMT-LIB-2.0) constraints over agent-observable state and tool arguments. At runtime, planned tool calls are intercepted and checked against the constraints using the Z3 solver as a pre-condition to the tool call. Tool invocations that violate the policy are blocked. We evaluated on the TauBench benchmark and demonstrate that solver-aided policy checking reduces policy violations while maintaining overall task accuracy. These results suggest that integrating formal reasoning into TaLLM execution can improve tool-call policy compliance and overall reliability.

We propose Melaguard, a multimodal ML framework (Transformer-lite, 1.2M parameters, 4-head self-attention) for detecting neurovascular instability (NVI) from wearable-compatible physiological signals prior to structural stroke pathology. The model fuses heart rate variability (HRV), peripheral perfusion index, SpO2, and bilateral phase coherence into a composite NVI Score, designed for edge inference (WCET <=4 ms on Cortex-M4). NVI - the pre-structural dysregulation of cerebrovascular autoregulation preceding overt stroke - remains undetectable by existing single-modality wearables. With 12.2 million incident strokes annually, continuous multimodal physiological monitoring offers a practical path to community-scale screening. Three-stage independent validation: (1) synthetic benchmark (n=10,000), AUC=0.88 [0.83-0.92]; (2) clinical cohort PhysioNet CVES (n=172; 84 stroke, 88 control) - Transformer-lite achieves AUC=0.755 [0.630-0.778], outperforming LSTM (0.643), Random Forest (0.665), SVM (0.472); HRV-SDNN discriminates stroke (p=0.011); (3) PPG pipeline PhysioNet BIDMC (n=53) -- pulse rate r=0.748 and HRV surrogate r=0.690 vs. ECG ground truth. Cross-modality validation on PPG-BP (n=219) confirms PPG morphology classifies cerebrovascular disease at AUC=0.923 [0.869-0.968]. Multimodal fusion consistently outperforms single-modality baselines. Code: https://github.com/ClevixLab/Melaguard

Verification-guided self-improvement has recently emerged as a promising approach to improving the accuracy of large language model (LLM) outputs. However, existing approaches face a trade-off between inference efficiency and accuracy: iterative verification-rectification is computationally expensive and prone to being trapped in faulty reasoning, while best-of-N selection requires extensive sampling without addressing internal model flaws. We propose a training-free regeneration paradigm that leverages an offline-curated contrastive Reflection Memory (RM) to provide corrective guidance, while regenerating from scratch helps break out of faulty reasoning. At inference time, the method performs RM-guided self-verification followed by a single RM-guided regeneration, avoiding both iterative correction and multi-sample selection. We evaluated our method on nine benchmarks that span algorithmic, reasoning, symbolic, and domain-specific tasks in both small- and large-scale LLMs. Experiment results show that our method outperforms prior methods while maintaining low computational cost.

While results visualization is a critical phase to the communication of new academic results, plots are frequently shared without the complete combination of code, input data, execution context and outputs required to independently reproduce the resulting figures. Existing reproducibility solutions tend to focus on computational pipelines or workflow management systems, not covering script-based visualization practices commonly used by researchers and practitioners. Additionally, the minimalist nature of current Python data visualization libraries tend to speed up the creation of images, disincentivizing users from spending time integrating additional tools into these short scripts. This paper proposes yProv4DV, a library lightweight designed to enable reproducible data visualization scripts through the use of provenance information, minimizing the necessity for code modifications. Through a single call, users can track inputs, outputs and source code files, enabling saving and full reproducibility of their data visualization software. As a result, this library fills a gap in reproducible research workflows by addressing the reproducibility of plots in scientific publications.

Extracting structured information from clinical notes requires navigating a dense web of interdependent variables where the value of one attribute logically constrains others. Existing Large Language Model (LLM)-based extraction pipelines often struggle to capture these dependencies, leading to clinically inconsistent outputs. We propose deep reflective reasoning, a large language model agent framework that iteratively self-critiques and revises structured outputs by checking consistency among variables, the input text, and retrieved domain knowledge, stopping when outputs converge. We extensively evaluate the proposed method in three diverse oncology applications: (1) On colorectal cancer synoptic reporting from gross descriptions (n=217), reflective reasoning improved average F1 across eight categorical synoptic variables from 0.828 to 0.911 and increased mean correct rate across four numeric variables from 0.806 to 0.895; (2) On Ewing sarcoma CD99 immunostaining pattern identification (n=200), the accuracy improved from 0.870 to 0.927; (3) On lung cancer tumor staging (n=100), tumor stage accuracy improved from 0.680 to 0.833 (pT: 0.842 -> 0.884; pN: 0.885 -> 0.948). The results demonstrate that deep reflective reasoning can systematically improve the reliability of LLM-based structured data extraction under interdependence constraints, enabling more consistent machine-operable clinical datasets and facilitating knowledge discovery with machine learning and data science towards digital health.

This paper proposes a computable state-estimation error bound for learning-based Kazantzis--Kravaris/Luenberger (KKL) observers. Recent work learns the KKL transformation map with a physics-informed neural network (PINN) and a corresponding left-inverse map with a conventional neural network. However, no computable state-estimation error bounds are currently available for this approach. We derive a state-estimation error bound that depends only on quantities that can be certified over a prescribed region using neural network verification. We further extend the result to bounded additive measurement noise and demonstrate the guarantees on nonlinear benchmark systems.

While Large Audio-Language Models (LALMs) have been shown to exhibit degraded instruction-following capabilities, their ability to infer task patterns from in-context examples under audio conditioning remains unstudied. To address this gap, we present ALICE, a three-stage framework that progressively reduces textual guidance to systematically evaluate LALMs' in-context learning ability under audio conditioning. Evaluating six LALMs across four audio understanding tasks under two output constraint categories, we uncover a consistent asymmetry across all stages and LALMs: in-context demonstrations reliably improve format compliance but fail to improve, and often degrade, the core task performance. This suggests that LALMs can glean surface-level formatting patterns from demonstrations but may struggle to leverage cross-modal semantic grounding to reliably infer task objectives from audio-conditioned examples, highlighting potential limitations in current cross-modal integration.

Large Language Models (LLMs) have demonstrated remarkable progress in scaling to access massive contexts. However, the access is via the latent and uninterpretable attention mechanisms, and LLMs fail to effective process long context, exhibiting significant performance degradation as context length increases. In this work, we study whether long-context processing can be externalized from latent attention into explicit, executable interactions, by allowing coding agents to organize text in file systems and manipulate it using its native tools. We evaluate off-the-shelf frontier coding agents as the general interface for tasks that require processing long contexts, including long-context reasoning, retrieval-augmented generation, and open-domain question answering with large-scale corpus contains up to three trillion tokens. Across multiple benchmarks, these agents outperform published state-of-the-art by 17.3% on average. We attribute this efficacy to two key factors: native tool proficiency, which enables agents to leverage executable code and terminal commands rather than passive semantic queries, and file system familiarity, which allows them to navigate massive text corpora as directory structures. These findings suggest that delegating long-context processing to coding agents offers an effective alternative to semantic search or context window scaling, opening new directions for long-context processing in LLMs.

Food security policy formulation in data-scarce regions remains a critical challenge due to limited structured datasets, fragmented textual reports, and demographic bias in decision-making systems. This study proposes ZeroHungerAI, an integrated Natural Language Processing (NLP) and Machine Learning (ML) framework designed for evidence-based food security policy modeling under extreme data scarcity. The system combines structured socio-economic indicators with contextual policy text embeddings using a transfer learning based DistilBERT architecture. Experimental evaluation on a 1200-sample hybrid dataset across 25 districts demonstrates superior predictive performance, achieving 91 percent classification accuracy, 0.89 precision, 0.85 recall, and an F1 score of 0.86 under imbalanced conditions. Comparative analysis shows a 13 percent performance improvement over classical SVM and 17 percent over Logistic Regression models. Precision Recall evaluation confirms robust minority class detection (average precision around 0.88). Fairness aware optimization reduces demographic parity difference to 3 percent, ensuring equitable rural urban policy inference. The results validate that transformer based contextual learning significantly enhances policy intelligence in low resource governance environments, enabling scalable and bias aware hunger prediction systems.

Human cognition of new concepts is inherently a streaming process: we continuously recognize new objects or identities and update our memories over time. However, current multimodal personalization methods are largely limited to static images or offline videos. This disconnects continuous visual input from instant real-world feedback, limiting their ability to provide the real-time, interactive personalized responses essential for future AI assistants. To bridge this gap, we first propose and formally define the novel task of Personalized Streaming Video Understanding (PSVU). To facilitate research in this new direction, we introduce PEARL-Bench, the first comprehensive benchmark designed specifically to evaluate this challenging setting. It evaluates a model's ability to respond to personalized concepts at exact timestamps under two modes: (1) Frame-level, focusing on a specific person or object in discrete frames, and (2) a novel Video-level, focusing on personalized actions unfolding across continuous frames. PEARL-Bench comprises 132 unique videos and 2,173 fine-grained annotations with precise timestamps. Concept diversity and annotation quality are strictly ensured through a combined pipeline of automated generation and human verification. To tackle this challenging new setting, we further propose PEARL, a plug-and-play, training-free strategy that serves as a strong baseline. Extensive evaluations across 8 offline and online models demonstrate that PEARL achieves state-of-the-art performance. Notably, it brings consistent PSVU improvements when applied to 3 distinct architectures, proving to be a highly effective and robust strategy. We hope this work advances vision-language model (VLM) personalization and inspires further research into streaming personalized AI assistants. Code is available at https://github.com/Yuanhong-Zheng/PEARL.

We present Hawkeye, a system for analyzing and reproducing GPU-level arithmetic operations. Using our framework, anyone can re-execute on a CPU the exact matrix multiplication operations underlying a machine learning model training or inference workflow that was executed on an NVIDIA GPU, without any precision loss. This is in stark contrast to prior approaches to verifiable machine learning, which either introduce significant computation overhead to the original model owner, or suffer from non-robustness and quality degradation. The main technical contribution of Hawkeye is a systematic sequence of carefully crafted tests that study rounding direction, subnormal number handling, and order of (non-associative) accumulation during matrix multiplication on NVIDIA's Tensor Cores. We test and evaluate our framework on multiple NVIDIA GPU architectures ( Ampere, Hopper, and Lovelace) and precision types (FP16, BFP16, FP8). In all test cases, Hawkeye enables perfect reproduction of matrix multiplication on a CPU, paving the way for efficient and trustworthy third-party auditing of ML model training and inference.

Nanopore sequencing can read substantially longer sequences of nucleic acid molecules than other sequencing methods, which has led to advances in genomic analysis such as the gapless human genome assembly. By analyzing the raw electrical signal reads that nanopore sequencing generates from molecules, existing works can map these reads without translating them into DNA characters (i.e., basecalling), allowing for quick and efficient analysis of sequencing data. However, raw signals often contain errors due to noise and mistakes when processing them, which limits the overall accuracy of raw signal analysis. Our goal in this work is to detect and correct errors in raw signals to improve the accuracy of raw signal analyses. To this end, we propose CERN, a mechanism that trains and utilizes a Hidden Markov Model (HMM) to accurately correct signal errors. Our extensive evaluation on various datasets including E. coli, Fruit Fly, and Human genomes shows that CERN 1) consistently improves the overall mapping accuracy of the underlying raw signal analysis tools, 2) minimizes the burden on segmentation algorithm optimization with newer nanopore chemistries, and 3) functions without causing substantial computational overhead. We conclude that CERN provides an effective mechanism to systematically identify and correct the errors in raw nanopore signals before further analysis, which can enable the development of a new class of error correction mechanisms purely designed for raw nanopore signals. CERN is available at https://github.com/STORMgroup/CERN. We also provide the scripts to fully reproduce our results on our GitHub page.

Unidirectional tapes surface roughness determines the evolution of the degree of intimate contact required for ensuring the thermoplastic molecular diffusion and the associated inter-tapes consolidation during manufacturing of composite structures. However, usual characterization of rough surfaces relies on statistical descriptors that even if they are able to represent the surface topology, they are not necessarily connected with the physics occurring at the interface during inter-tape consolidation. Thus, a key research question could be formulated as follows: Which roughness descriptors simultaneously enable tape classification-crucial for process control-and consolidation modeling via the inference of the evolution of the degree of intimate contact, itself governed by the process parameters?. For providing a valuable response, we propose a novel strategy based on the use of Rank Reduction Autoencoders (RRAEs), autoencoders with a linear latent vector space enforced by applying a truncated Singular Value Decomposition (SVD) to the latent matrix during the encoder-decoder training. In this work, we extract useful roughness descriptors by enforcing the latent SVD modes to (i) accurately represent the roughness after decoding, and (ii) allow the extraction of existing a priori knowledge such as classification or modelling properties.

Research software (also called scientific software) is essential for advancing scientific endeavours. Research software encapsulates complex algorithms and domain-specific knowledge and is a fundamental component of all science. A pervasive challenge in developing research software is technical debt, which can adversely affect reliability, maintainability, and scientific validity. Research software often relies on the initiative of the scientific community for maintenance, requiring diverse expertise in both scientific and software engineering domains. The extent and nature of technical debt in research software are little studied, in particular, what forms it takes, and what the science teams developing this software think about their technical debt. In this paper we describe our multi-method study examining technical debt in research software. We begin by examining instances of self-reported technical debt in research code, examining 28k code comments across nine research software projects. Then, building on our findings, we interview research software engineers and scientists about how this technical debt manifests itself in their experience, and what costs it has for research software and research outputs more generally. We identify nine types of self-admitted technical debt unique to research software, and four themes impacting this technical debt.

Scientific machine learning is increasingly used to build surrogate models, yet most models are trained under a restrictive assumption in which future data follow the same distribution as the training set. In practice, new experimental conditions or simulation regimes may differ significantly, requiring extrapolation and model updates without re-access to prior data. This creates a need for continual learning (CL) frameworks that can adapt to distribution shifts while preventing catastrophic forgetting. Such challenges are pronounced in fluid dynamics, where changes in geometry, boundary conditions, or flow regimes induce non-trivial changes to the solution. Here, we introduce a new architecture-based approach (SLE-FNO) combining a Single-Layer Extension (SLE) with the Fourier Neural Operator (FNO) to support efficient CL. SLE-FNO was compared with a range of established CL methods, including Elastic Weight Consolidation (EWC), Learning without Forgetting (LwF), replay-based approaches, Orthogonal Gradient Descent (OGD), Gradient Episodic Memory (GEM), PiggyBack, and Low-Rank Approximation (LoRA), within an image-to-image regression setting. The models were trained to map transient concentration fields to time-averaged wall shear stress (TAWSS) in pulsatile aneurysmal blood flow. Tasks were derived from 230 computational fluid dynamics simulations grouped into four sequential and out-of-distribution configurations. Results show that replay-based methods and architecture-based approaches (PiggyBack, LoRA, and SLE-FNO) achieve the best retention, with SLE-FNO providing the strongest overall balance between plasticity and stability, achieving accuracy with zero forgetting and minimal additional parameters. Our findings highlight key differences between CL algorithms and introduce SLE-FNO as a promising strategy for adapting baseline models when extrapolation is required.

Classical Bayesian persuasion studies how a sender influences receivers through carefully designed signaling policies within a single strategic interaction. In many real-world environments, such interactions are repeated across multiple games, creating opportunities to exploit structural similarity across tasks. In this work, we introduce Meta-Persuasion algorithms, establishing the first line of theoretical results for both full-feedback and bandit-feedback settings in the Online Bayesian Persuasion (OBP) and Markov Persuasion Process (MPP) frameworks. We show that our proposed meta-persuasion algorithms achieve provably sharper regret rates under natural notions of task similarity, improving upon the best-known convergence rates for both OBP and MPP. At the same time, they recover the standard single-game guarantees when the sequence of games is picked arbitrarily. Finally, we complement our theoretical analysis with numerical experiments that highlight our regret improvements and the benefits of meta-learning in repeated persuasion environments.

We investigate whether independently trained language models converge to geometrically compatible latent representations, and whether this compatibility can be exploited to correct model behavior at inference time without any weight updates. We learn a linear projection matrix that maps activation vectors from a large teacher model into the coordinate system of a smaller student model, then intervene on the student's residual stream during generation by substituting its internal state with the translated teacher representation. Across a fully crossed experimental matrix of 20 heterogeneous teacher-student pairings spanning mixture-of-experts, dense, code-specialized, and synthetically trained architectures, the Ridge projection consistently achieves R^2 = 0.50 on verbal reasoning and R^2 = 0.40 on mathematical reasoning, collapsing to R^2 = -0.22 under permutation control and R^2 = 0.01 under L_1 regularization. Behavioral correction rates range from 14.0% to 50.0% on TruthfulQA (mean 25.2%) and from 8.5% to 43.3% on GSM8K arithmetic reasoning (mean 25.5%), demonstrating that the method generalizes across fundamentally different reasoning domains. We report a near-zero correlation between geometric alignment quality and behavioral correction rate (r = -0.07), revealing a dissociation between representation space fidelity and output space impact. Intervention strength is architecture-specific: student models exhibit characteristic sensitivity profiles that invert across domains, with the most steerable verbal student becoming the least steerable mathematical student. Finally, a double dissociation experiment conducted across all 20 model pairings confirms without exception that projection matrices collapse catastrophically when transferred across reasoning domains (mean R^2 = -3.83 in both transfer directions), establishing domain-specific subspace geometry as a universal property of LMs.

We report on an experiment in which Claude Opus~4.6, equipped with a suite of Model Context Protocol (MCP) tools for the Rocq proof assistant, autonomously proved 10 of 12 problems from the 2025 Putnam Mathematical Competition. The MCP tools, designed with Claude by analyzing logs from a prior experiment on miniF2F-Rocq, encode a "compile-first, interactive-fallback" strategy. Running on an isolated VM with no internet access, the agent deployed 141 subagents over 17.7 hours of active compute (51.6h wall-clock), consuming approximately 1.9 billion tokens. All proofs are publicly available.

The evolution of Internet of Things (IoT) into multi-layered environments has positioned Low-Power Wide Area Networks (LPWANs), particularly Long Range (LoRa), as the backbone for connectivity across both surface and subterranean landscapes. However, existing LoRa-based network designs often treat ground-based wireless sensor networks (WSNs) and wireless underground sensor networks (WUSNs) as separate systems, resulting in inefficient and non-integrated connectivity across diverse environments. To address this, we propose Hetero-Net, a unified heterogeneous LoRa framework that integrates diverse LoRa end devices with multiple unmanned aerial vehicle (UAV)-mounted LoRa gateways. Our objective is to maximize system energy efficiency through the joint optimization of the spreading factor, transmission power, and three-dimensional (3D) placement of the UAVs. To manage the dynamic and partially observable nature of this system, we model the problem as a partially observable stochastic game (POSG) and address it using a multi-agent proximal policy optimization (MAPPO) framework. An ablation study shows that our proposed MAPPO Hetero-Net significantly outperforms traditional, isolated network designs, achieving energy efficiency improvements of 55.81\% and 198.49\% over isolated WSN-only and WUSN-only deployments, respectively.

The key-value (KV) cache is a foundational optimization in Transformer-based large language models (LLMs), eliminating redundant recomputation of past token representations during autoregressive generation. However, its memory footprint scales linearly with context length, imposing critical bottlenecks on GPU memory capacity, memory bandwidth, and inference throughput as production LLMs push context windows from thousands to millions of tokens. Efficient KV cache management has thus become a first-order challenge for scalable LLM deployment. This paper provides a systematic review of recent KV cache optimization techniques, organizing them into five principal directions: cache eviction, cache compression, hybrid memory solutions, novel attention mechanisms, and combination strategies. For each category we analyze the underlying mechanisms, deployment trade-offs, and empirical performance across memory reduction, throughput, and model accuracy metrics. We further map techniques to seven practical deployment scenarios, including long-context single requests, high-throughput datacenter serving, edge devices, multi-turn conversations, and accuracy-critical reasoning, providing actionable guidance for practitioners selecting among competing approaches. Our analysis reveals that no single technique dominates across all settings; instead, the optimal strategy depends on context length, hardware constraints, and workload characteristics, pointing toward adaptive, multi-stage optimization pipelines as a promising direction for future research.

Human mathematics (HM), the mathematics humans discover and value, is a vanishingly small subset of formal mathematics (FM), the totality of all valid deductions. We argue that HM is distinguished by its compressibility through hierarchically nested definitions, lemmas, and theorems. We model this with monoids. A mathematical deduction is a string of primitive symbols; a definition or theorem is a named substring or macro whose use compresses the string. In the free abelian monoid $A_n$, a logarithmically sparse macro set achieves exponential expansion of expressivity. In the free non-abelian monoid $F_n$, even a polynomially-dense macro set only yields linear expansion; superlinear expansion requires near-maximal density. We test these models against MathLib, a large Lean~4 library of mathematics that we take as a proxy for HM. Each element has a depth (layers of definitional nesting), a wrapped length (tokens in its definition), and an unwrapped length (primitive symbols after fully expanding all references). We find unwrapped length grows exponentially with both depth and wrapped length; wrapped length is approximately constant across all depths. These results are consistent with $A_n$ and inconsistent with $F_n$, supporting the thesis that HM occupies a polynomially-growing subset of the exponentially growing space FM. We discuss how compression, measured on the MathLib dependency graph, and a PageRank-style analysis of that graph can quantify mathematical interest and help direct automated reasoning toward the compressible regions where human mathematics lives.

Probabilistic circuit (PC) structure learning is hampered by greedy algorithms that make irreversible, locally optimal decisions. We propose SymCircuit, which replaces greedy search with a learned generative policy trained via entropy-regularized reinforcement learning. Instantiating the RL-as-inference framework in the PC domain, we show the optimal policy is a tempered Bayesian posterior, recovering the exact posterior when the regularization temperature is set inversely proportional to the dataset size. The policy is implemented as SymFormer, a grammar-constrained autoregressive Transformer with tree-relative self-attention that guarantees valid circuits at every generation step. We introduce option-level REINFORCE, restricting gradient updates to structural decisions rather than all tokens, yielding an SNR (signal to noise ratio) improvement and >10 times sample efficiency gain on the NLTCS dataset. A three-layer uncertainty decomposition (structural via model averaging, parametric via the delta method, leaf via conjugate Dirichlet-Categorical propagation) is grounded in the multilinear polynomial structure of PC outputs. On NLTCS, SymCircuit closes 93% of the gap to LearnSPN; preliminary results on Plants (69 variables) suggest scalability.

Cooperative multi-agent reinforcement learning (c-MARL) has been widely deployed in real-world applications, such as social robots, embodied intelligence, UAV swarms, etc. Nevertheless, many adversarial attacks still exist to threaten various c-MARL systems. At present, the studies mainly focus on single-adversary perturbation attacks and white-box adversarial attacks that manipulate agents' internal observations or actions. To address these limitations, we in this paper attempt to study collusive adversarial attacks through strategically organizing a set of malicious agents into three collusive attack modes: Collective Malicious Agents, Disguised Malicious Agents, and Spied Malicious Agents. Three novelties are involved: i) three collusive adversarial attacks are creatively proposed for the first time, and a unified framework CAMA for policy-level collusive attacks is designed; ii) the attack effectiveness is theoretically analyzed from the perspectives of disruptiveness, stealthiness, and attack cost; and iii) the three collusive adversarial attacks are technically realized through agent's observation information fusion, attack-trigger control. Finally, multi-facet experiments on four SMAC II maps are performed, and experimental results showcase the three collusive attacks have an additive adversarial synergy, strengthening attack outcome while maintaining high stealthiness and stability over long horizons. Our work fills the gap for collusive adversarial learning in c-MARL.

Reticular chemistry has enabled the synthesis of tens of thousands of metal-organic frameworks (MOFs), yet the discovery of new materials still relies largely on intuition-driven linker design and iterative experimentation. As a result, researchers explore only a small fraction of the vast chemical space accessible to reticular materials, limiting the systematic discovery of frameworks with targeted properties. Here, we introduce Nexerra-R1, a building-block chemical language model that enables inverse design in reticular chemistry through the targeted generation of organic linkers. Rather than generating complete frameworks directly, Nexerra-R1 operates at the level of molecular building blocks, preserving the modular logic that underpins reticular synthesis. The model supports both unconstrained generation of low-connectivity linkers and scaffold-constrained design of symmetric multidentate motifs compatible with predefined nodes and topologies. We further combine linker generation with flow-guided distributional targeting to steer the generative process toward application-relevant objectives while maintaining chemical validity and assembly feasibility. The generated linkers are subsequently assembled into three-dimensional frameworks and are structurally optimized to produce candidate materials compatible with experimental synthesis. Using Nexerra-R1, we validate this strategy by rediscovering known MOFs and by proposing the experimental synthesis of a previously unreported framework, CU-525, generated entirely in silico. Together, these results establish a general inverse-design paradigm for reticular materials in which controllable chemical language modelling enables the direct translation from computational design to synthesizable frameworks.

We derive the asymptotic risk function of regularized empirical risk minimization (ERM) estimators tuned by $n$-fold cross-validation (CV). The out-of-sample prediction loss of such estimators converges in distribution to the squared-error loss (risk function) of shrinkage estimators in the normal means model, tuned by Stein's unbiased risk estimate (SURE). This risk function provides a more fine-grained picture of predictive performance than uniform bounds on worst-case regret, which are common in learning theory: it quantifies how risk varies with the true parameter. As key intermediate steps, we show that (i) $n$-fold CV converges uniformly to SURE, and (ii) while SURE typically has multiple local minima, its global minimum is generically well separated. Well-separation ensures that uniform convergence of CV to SURE translates into convergence of the tuning parameter chosen by CV to that chosen by SURE.

Understanding the fundamental mechanisms governing the production of meaning in the processing of natural language is critical for designing safe, thoughtful, engaging, and empowering human-agent interactions. Experiments in cognitive science and social psychology have demonstrated that human semantic processing exhibits contextuality more consistent with quantum logical mechanisms than classical Boolean theories, and recent works have found similar results in large language models -- in particular, clear violations of the Bell inequality in experiments of contextuality during interpretation of ambiguous expressions. We explore the CHSH $|S|$ parameter -- the metric associated with the inequality -- across the inference parameter space of models spanning four orders of magnitude in scale, cross-referencing it with MMLU, hallucination rate, and nonsense detection benchmarks. We find that the interquartile range of the $|S|$ distribution -- the statistic that most sharply differentiates models from one another -- is completely orthogonal to all external benchmarks, while violation rate shows weak anticorrelation with all three benchmarks that does not reach significance. We investigate how $|S|$ varies with sampling parameters and word order, and discuss the information-theoretic constraints that genuine contextuality imposes on prompt injection defenses and its human analogue, whereby careful construction and maintenance of social contextuality can be carried out at scale -- manufacturing not consent but contextuality itself, a subtler and more fundamental form of manipulation that shapes the space of possible interpretations before any particular one is reached.

Industry practitioners and academic researchers regularly use multi-agent systems to accelerate their work, yet the frameworks through which these systems operate do not provide a simple, unified mechanism for scalably managing the critical aspects of the agent harness, impacting both the quality of individual human-agent interactions and the capacity for practitioners to coordinate toward common goals through shared agent infrastructure. Agent frameworks have enabled increasingly sophisticated multi-agent systems, but the behavioral specifications that define what these agents can do remain fragmented across prose instruction files, framework-internal configuration, and mechanisms like MCP servers that operate separately from individual agent definitions, making these specifications difficult to share, version, or collaboratively maintain across teams and projects. Applying the ALARA principle from radiation safety (exposures kept as low as reasonably achievable) to agent context, we introduce a declarative context-agent-tool (CAT) data layer expressed through interrelated files that scope each agent's tool access and context to the minimum its role requires, and \texttt{npcsh}, a command-line shell for executing it. Because the system parses and enforces these files structurally, modifying an agent's tool list produces a guaranteed behavioral change rather than a suggestion the model may or may not follow. We evaluate 22 locally-hosted models from 0.6B to 35B parameters across 115 practical tasks spanning file operations, web search, multi-step scripting, tool chaining, and multi-agent delegation, characterizing which model families succeed at which task categories and where they break down across $\sim$2500 total executions.

Despite significant advances in autonomous web navigation, current methods remain far from human-level performance in complex web environments. We argue that this limitation stems from Topological Blindness, where agents are forced to explore via trial-and-error without access to the global topological structure of the environment. To overcome this limitation, we introduce WebNavigator, which reframes web navigation from probabilistic exploration into deterministic retrieval and pathfinding. WebNavigator constructs Interaction Graphs via zero-token cost heuristic exploration offline and implements a Retrieve-Reason-Teleport workflow for global navigation online. WebNavigator achieves state-of-the-art performance on WebArena and OnlineMind2Web. On WebArena multi-site tasks, WebNavigator achieves a 72.9\% success rate, more than doubling the performance of enterprise-level agents. This work reveals that Topological Blindness, rather than model reasoning capabilities alone, is an underestimated bottleneck in autonomous web navigation.

Probability theory has become the predominant framework for quantifying uncertainty across scientific and engineering disciplines, with a particular focus on measurement and control systems. However, the widespread reliance on simple Gaussian assumptions--particularly in control theory, manufacturing, and measurement systems--can result in incomplete representations and multistage lossy approximations of complex phenomena, including inaccurate propagation of uncertainty through multi stage processes. This work proposes a comprehensive yet computationally tractable framework for representing and propagating quantitative attributes arising in measurement systems using Probability Density Functions (PDFs). Recognizing the constraints imposed by finite memory in software systems, we advocate for the use of Gaussian Mixture Models (GMMs), a principled extension of the familiar Gaussian framework, as they are universal approximators of PDFs whose complexity can be tuned to trade off approximation accuracy against memory and computation. From both mathematical and computational perspectives, GMMs enable high performance and, in many cases, closed form solutions of essential operations in control and measurement. The paper presents practical applications within manufacturing and measurement contexts especially circular factory, demonstrating how the GMMs framework supports accurate representation and propagation of measurement uncertainty and offers improved accuracy--compared to the traditional Gaussian framework--while keeping the computations tractable.

Dynamic GNN inference has exhibited effectiveness in High Energy Physics (HEP) experiments at High Luminosity Large Hadron Collider (HL-LHC) due to strong capability to model complex particle interactions in collision events. Future HEP experiments will involve detectors that produce 10x more collision data to help unlocking physics discoveries. Due to limitations in offline compute capacity and storage, revamped trigger systems require FPGAs to run ultra-low-latency Machine Learning models for online filtering of useful events with low power consumption. State-of-the-art GNN accelerators relied on static graph structures, but this assumption breaks down in real-time HL-LHC trigger systems and edge-based dynamic GNN models where edge embeddings change in-place based on neighbor node embeddings at runtime. We propose DGNNFlow, a novel dataflow architecture for real-time edge-based dynamic GNN inference applications, especially HL-LHC trigger systems, with three key contributions. First, we introduce hardware support for dynamic computation of edge embeddings. Second, we resolve data dependencies in edge-based dynamic GNN dataflow, where edge embedding is formulated using its source and target nodes. Third, we perform input dynamic graph construction auxiliary setup for complete support of models without pre-defined edge embeddings. We deployed DGNNFlow using AMD Alveo U50 FPGA to evaluate end-to-end latency on-board at 200 MHz clock frequency. DGNNFlow achieved 1.6x-6.3x speedup and 0.22x power consumption compared to GPU (NVIDIA RTX A6000) with batch sizes from 1 to 4, 3.2x-5.1x speedup and 0.25x power consumption compared to CPU (Intel Xeon Gold 6226R). Our complete implementation is publicly available on GitHub.

We aim to examine the extent to which Large Language Models (LLMs) can 'talk much' about grammar modules, providing evidence from syntax core properties translated by ChatGPT into Arabic. We collected 44 terms from generative syntax previous works, including books and journal articles, as well as from our experience in the field. These terms were translated by humans, and then by ChatGPT-5. We then analyzed and compared both translations. We used an analytical and comparative approach in our analysis. Findings unveil that LLMs still cannot 'talk much' about the core syntax properties embedded in the terms under study involving several syntactic and semantic challenges: only 25% of ChatGPT translations were accurate, while 38.6% were inaccurate, and 36.4.% were partially correct, which we consider appropriate. Based on these findings, a set of actionable strategies were proposed, the most notable of which is a close collaboration between AI specialists and linguists to better LLMs' working mechanism for accurate or at least appropriate translation.

Machine learning has opened new frontiers in purely data-driven algorithms for data assimilation in, and for forecasting of, dynamical systems; the resulting methods are showing some promise. However, in contrast to model-driven algorithms, analysis of these data-driven methods is poorly developed. In this paper we address this issue, developing a theory to underpin data-driven methods to solve smoothing problems arising in data assimilation and forecasting problems. The theoretical framework relies on two key components: (i) establishing the existence of the mapping to be learned; (ii) the properties of the operator learning architecture used to approximate this mapping. By studying these two components in conjunction, we establish the first universal approximation theorem for purely data-driven algorithms for both smoothing and forecasting of dynamical systems. We work in the continuous time setting, hence deploying neural operator architectures. The theoretical results are illustrated with experiments studying the Lorenz `63, Lorenz `96 and Kuramoto-Sivashinsky dynamical systems.

Modern web test automation frameworks rely heavily on CSS selectors, XPath expressions, and visible text labels to locate UI elements. These locators are inherently brittle -- when web applications update their DOM structure or class names, test suites fail at scale. Existing self-healing approaches increasingly delegate element discovery to Large Language Models (LLMs), introducing per-run API costs that become prohibitive at enterprise scale. This paper presents a zero-cost self-healing test automation framework that replaces LLM-based discovery with a structured accessibility tree extraction algorithm. The framework employs a ten-tier priority-ranked locator hierarchy -- get_by_role (W3C standard), data-testid, ARIA labels, CSS class fragments, visible text -- to discover robust selectors from a live DOM in a single one-time pass. A self-healing mechanism re-extracts only broken selectors upon failure, rather than re-running full discovery. The framework is validated against automationexercise.com across three device profiles (Desktop Chrome, Desktop Safari, iPhone 15) and ten business process test workflows under a three-tier hierarchy (L0: Domain, L1: Process, L2: Feature). Results demonstrate a 31/31 (100%) pass rate across 31 test combinations with total execution time of 22 seconds under parallel execution. Self-healing is empirically demonstrated: a stale selector is detected and re-discovered in under 1 second with zero human intervention. The framework scales to 300+ test cases with zero ongoing API cost.

Memory poisoning attacks for Agentic AI and multi-agent systems (MAS) have recently caught attention. It is partially due to the fact that Large Language Models (LLMs) facilitate the construction and deployment of agents. Different memory systems are being used nowadays in this context, including semantic, episodic, and short-term memory. This distinction between the different types of memory systems focuses mostly on their duration but also on their origin and their localization. It ranges from the short-term memory originated at the user's end localized in the different agents to the long-term consolidated memory localized in well established knowledge databases. In this paper, we first present the main types of memory systems, we then discuss the feasibility of memory poisoning attacks in these different types of memory systems, and we propose mitigation strategies. We review the already existing security solutions to mitigate some of the alleged attacks, and we discuss adapted solutions based on cryptography. We propose to implement local inference based on private knowledge retrieval as an example of mitigation strategy for memory poisoning for semantic memory. We also emphasize actual risks in relation to interactions between agents, which can cause memory poisoning. These latter risks are not so much studied in the literature and are difficult to formalize and solve. Thus, we contribute to the construction of agents that are secure by design.

A short video succeeds not simply because of what it shows, but because of how it schedules attention -- yet current multimodal models lack the structural grammar to parse or produce this organization. Existing models can describe scenes, answer event-centric questions, and read on-screen text, but they are far less reliable at identifying timeline-grounded units such as hooks, cut rationales, shot-induced tension, and platform-facing packaging cues. We propose SV6D (Structured Video in Six Dimensions), inspired by professional storyboard practice in film and television production, a representation framework that decomposes internet-native video into six complementary structural dimensions -- subject, aesthetics, camera language, editing, narrative, and dissemination -- with each label tied to physically observable evidence on the timeline. We formalize a unified optimization objective over SV6D that combines Hungarian-matched temporal alignment, dimension-wise semantic label distance, and quality regularization. Building on this framework, we present Leum-VL-8B, an 8B video-language model that realizes the SV6D objective through an expert-driven post-training pipeline, further refined through verifiable reinforcement learning on perception-oriented tasks. Leum-VL-8B achieves 70.8 on VideoMME (w/o subtitles), 70.0 on MVBench, and 61.6 on MotionBench, while remaining competitive on general multimodal evaluations such as MMBench-EN. We also construct FeedBench, a benchmark for structure-sensitive short-video understanding. Our results indicate that the missing layer in video AI is not pixel generation but structural representation: grounded on the timeline, linked to visible evidence, and directly consumable by downstream workflows such as editing, retrieval, recommendation, and generation control, including text-heavy internet video formats with overlays and image-text layouts.

Time series machine learning (TSML) is a growing research field that spans a wide range of tasks. The popularity of established tasks such as classification, clustering, and extrinsic regression has, in part, been driven by the availability of benchmark datasets. An archive of 30 multivariate time series classification datasets, introduced in 2018 and commonly known as the UEA archive, has since become an essential resource cited in hundreds of publications. We present a substantial expansion of this archive that more than quadruples its size, from 30 to 133 classification problems. We also release preprocessed versions of datasets containing missing values or unequal length series, bringing the total number of datasets to 147. Reflecting the growth of the archive and the broader community, we rebrand it as the Multiverse archive to capture its diversity of domains. The Multiverse archive includes datasets from multiple sources, consolidating other collections and standalone datasets into a single, unified repository. Recognising that running experiments across the full archive is computationally demanding, we recommend a subset of the full archive called Multiverse-core (MV-core) for initial exploration. To support researchers in using the new archive, we provide detailed guidance and a baseline evaluation of established and recent classification algorithms, establishing performance benchmarks for future research. We have created a dedicated repository for the Multiverse archive that provides a common aeon and scikit-learn compatible framework for reproducibility, an extensive record of published results, and an interactive interface to explore the results.

Mobile advertising dominates app monetization but introduces risks ranging from intrusive user experience to malware delivery. Existing detection methods rely either on static analysis, which misses runtime behaviors, or on heuristic UI exploration, which struggles with sparse and obfuscated ads. In this paper, we present MANA, the first agentic multimodal reasoning framework for mobile ad detection. MANA integrates static, visual, temporal, and experiential signals into a reasoning-guided navigation strategy that determines not only how to traverse interfaces but also where to focus, enabling efficient and robust exploration. We implement and evaluate MANA on commercial smartphones over 200 apps, achieving state-of-the-art accuracy and efficiency. Compared to baselines, it improves detection accuracy by 30.5%-56.3% and reduces exploration steps by 29.7%-63.3%. Case studies further demonstrate its ability to uncover obfuscated and malicious ads, underscoring its practicality for mobile ad auditing and its potential for broader runtime UI analysis (e.g., permission abuse). Code and dataset are available at https://github.com/MANA-2026/MANA.

Low-level C programs remain highly vulnerable to out-of-bounds memory corruption. State-of-the-art precise defenses either introduce severe runtime overhead due to metadata memory lookups, or break standard C semantics by disallowing partial structs or the creation of an object's end address (EA), a legal operation ubiquitous in real-world C code. Conversely, practical alignment-based solutions achieve efficiency only by relaxing protected bounds. We present PRISM, a precise, zero-lookup object-bounds scheme that eliminates these restrictions. PRISM compresses a 47-bit EA into the 17-bit unused tag area of a 64-bit pointer. By enforcing the invariant that a statically known starting address (KSA) cannot exceed the EA, PRISM completely eliminates the need for costly metadata memory fetches in nearly all bounds checks, while strictly retaining precise object bounds. Our invariant also simplifies the lower-bound checks in existing alignment-based solutions, thus improving their performance. To achieve high throughput, PRISM introduces q-padding, an optimization that safely removes bounds checks for constant-offset accesses (such as struct fields) while maintaining precise, byte-level protection for the variable-indexed accesses primarily exploited by attackers. Evaluated on SPEC 2017, PRISM achieves an arithmetic mean CPU overhead of 46.1\% with a 32-byte q-padding (dropping to 31.3\% in a 32-bit address space). On highly concurrent, real-world workloads, PRISM secures a fully saturated Apache web server with only an 11.1\% throughput reduction, demonstrating its readiness for production deployment. Furthermore, PRISM successfully detected an out-of-bounds violation in \texttt{gcc} that prior tools missed due to their lack of support for partial structs.

Human genetics offers a promising route to therapeutic discovery, yet practical frameworks translating genotype-derived signal into ranked target and drug hypotheses remain limited, particularly when matched disease transcriptomics are unavailable. Here we present G2DR, a genotype-first prioritization framework propagating inherited variation through genetically predicted expression, multi-method gene-level testing, pathway enrichment, network context, druggability, and multi-source drug--target evidence integration. In a migraine case study with 733 UK Biobank participants under stratified five-fold cross-validation, we imputed expression across seven transcriptome-weight resources and ranked genes using a reproducibility-aware discovery score from training and validation data, followed by a balanced integrated score for target selection. Discovery-based prioritization generalized to held-out data, achieving gene-level ROC-AUC of 0.775 and PR-AUC of 0.475, while retaining enrichment for curated migraine biology. Mapping prioritized genes to compounds via Open Targets, DGIdb, and ChEMBL yielded drug sets enriched for migraine-linked compounds relative to a global background, though recovery favoured broader mechanism-linked and off-label space over migraine-specific approved therapies. Directionality filtering separated broadly recovered compounds from mechanistically compatible candidates. G2DR is a modular framework for genetics-informed hypothesis generation, not a clinically actionable recommendation system. All outputs require independent experimental, pharmacological, and clinical validation.

Autoformalization aims to translate natural-language mathematics into compilable, machine-checkable statements. However, semantic consistency does not imply prover effectiveness: even semantically consistent formalizations can differ substantially in proof-search cost and success rate. In this work, we formulate autoformalization as a budgeted, test-time search for semantically consistent repertoires, and propose FormalEvolve, a compilation-gated neuro-symbolic evolutionary framework. FormalEvolve generates diverse candidates via LLM-driven mutation and crossover with bounded patch repair, while symbolic Abstract Syntax Tree (AST) rewrite operations further inject structural diversity. On CombiBench and ProofNet, under a strict generator-call budget of T = 100, FormalEvolve reaches semantic hit rates (SH@100) of 58.0% and 84.9%, and reduces cross-problem concentration of semantic successes(lower Gini). Under a fixed prover budget, FormalEvolve also improves downstream proving performance on CombiBench. Code will be released publicly.

Machine learning (ML) promises better clinical decision-making, yet opaque model behavior limits the adoption in healthcare. We propose two novel regularization techniques for ensuring the interpretability of ML models trained on real-world data. In particular, we consider the prediction of five-year survival for multiple myeloma patients using clinical data from Helsinki University Hospital. To ensure the interpretability of the trained models, we use two alternative constructions for a penalty term used for regularization. The first one penalizes deviations from the predictions obtained from an interpretable logistic regression method with two manually chosen features. The second construction requires consistency of model predictions with the revised international staging system (R-ISS). We verify the usefulness of the proposed regularization techniques in numerical experiments using data from 812 patients. They achieve an accuracy up to 0.721 on a test set and SHAP values show that the models rely on the selected important features.

Despite rapid progress in multimodal GUI agents, reusable skill acquisition remains difficult because on-demand generated skills often leave action semantics, state assumptions, and success criteria implicit. This makes them brittle to execution errors, hard to verify, and difficult to repair. We present ContractSkill, a framework that converts a draft skill into a contracted executable artifact with explicit preconditions, step specifications, postconditions, recovery rules, and termination checks. This representation enables deterministic verification, step-level fault localization, and minimal patch-based repair, turning skill refinement into localized editing rather than full regeneration. Experiments on VisualWebArena and MiniWoB with GLM-4.6V and Qwen3.5-Plus show that ContractSkill improves self-generated skills from 9.4% and 10.9% to 28.1% and 37.5% on VisualWebArena, and from 66.5% and 60.5% to 77.5% and 81.0% on MiniWoB. Repaired artifacts also transfer across models, improving the target model's self-generated-skill baseline by up to 47.8 points and 12.8 points on the two benchmarks, respectively. These results suggest that agent skills are better treated as explicit procedural artifacts that can be verified, repaired, and shared across models.

Many learning systems now use graph data in which each node also contains text, such as papers with abstracts or users with posts. Because these texts often come from open platforms, an attacker may be able to quietly poison a small part of the training data and later make the model produce wrong predictions on demand. This paper studies that risk in a realistic setting where the attacker edits only node text and does not change the graph structure. We propose TAGBD, a text-only backdoor attack for text-attributed graphs. TAGBD first finds training nodes that are easier to influence, then generates natural-looking trigger text with the help of a shadow graph model, and finally injects the trigger by either replacing the original text or appending a short phrase. Experiments on three benchmark datasets show that the attack is highly effective, transfers across different graph models, and remains strong under common defenses. These results demonstrate that text alone is a practical attack channel in graph learning systems and suggest that future defenses should inspect both graph links and node content.

Federated Recommender Systems (FRS) preserve privacy by training decentralized models on client-specific user-item subgraphs without sharing raw data. However, FRS faces a unique challenge: subgraph structural imbalance, where drastic variations in subgraph scale (user/item counts) and connectivity (item degree) misalign client representations, making it challenging to train a robust model that respects each client's unique structural characteristics. To address this, we propose a Low-pass Personalized Subgraph Federated recommender system (LPSFed). LPSFed leverages graph Fourier transforms and low-pass spectral filtering to extract low-frequency structural signals that remain stable across subgraphs of varying size and degree, allowing robust personalized parameter updates guided by similarity to a neutral structural anchor. Additionally, we leverage a localized popularity bias-aware margin that captures item-degree imbalance within each subgraph and incorporates it into a personalized bias correction term to mitigate recommendation bias. Supported by theoretical analysis and validated on five real-world datasets, LPSFed achieves superior recommendation accuracy and enhances model robustness.

In multi-vector retrieval, both queries and data are represented as sets of high-dimensional vectors, enabling finer-grained semantic matching and improving retrieval quality over single-vector approaches. However, its practical adoption is held back by the lack of effective indexing algorithms. Existing work, attempting to reuse standard single-vector indexes, often fails to preserve multi-vector semantics or remains slow. In this work, we present GEM, a native indexing framework for multi-vector representations. The core idea is to construct a proximity graph directly over vector sets, preserving their fine-grained semantics while enabling efficient navigation. First, GEM designs a set-level clustering scheme. It associates each vector set with only its most informative clusters, effectively reducing redundancy without hurting semantic coverage. Then, it builds local proximity graphs within clusters and bridges them into a globally navigable structure. To handle the non-metric nature of multi-vector similarity, GEM decouples the graph construction metric from the final relevance score and injects semantic shortcuts to guide efficient navigation toward relevant regions. At query time, GEM launches beam search from multiple entry points and prunes paths early using cluster cues. To further enhance efficiency, a quantized distance estimation technique is used for both indexing and search. Across in-domain, out-of-domain, and multi-modal benchmarks, GEM achieves up to 16x speedup over state-of-the-art methods while matching or improving accuracy.

The Interest Public Group ARRONAX's C70XP cyclotron, used for radioisotope production for medical and research applications, relies on complex and costly systems that are prone to failures, leading to operational disruptions. In this context, this study aims to develop a machine learning-based method for early anomaly detection, from sensor measurements over a temporal window, to enhance system performance. One of the most widely recognized methods for anomaly detection is Isolation Forest (IF), known for its effectiveness and scalability. However, its reliance on axis-parallel splits limits its ability to detect subtle anomalies, especially those occurring near the mean of normal data. This study proposes a hybrid approach that combines a fully connected Autoencoder (AE) with IF to enhance the detection of subtle anomalies. In particular, the Mean Cubic Error (MCE) of the sensor data reconstructed by the AE is used as input to the IF model. Validated on proton beam intensity time series data, the proposed method demonstrates a clear improvement in detection performance, as confirmed by the experimental results.

In complex code-generation tasks, conversation-based LLM code repair exhibits limited ability to recover from first-pass programming errors, as such code revisions are usually driven by LLMs' "plausible reasoning" rather than a formal, algorithmic debugging procedure. However, a formal foundation for such debugging exists in Udi Shapiro's theory of algorithmic program debugging (APD), which frames program repair as an explicit, stepwise procedural refinement process. In this paper, we propose a neuro-symbolic procedural refinement approach, Abduction-Based Procedural Refinement (ABPR), which couples an LLM with a meta-interpreter that materialises program execution into compact, declarative tree-structured traces, following the principles of APD. We evaluate ABPR on ARC-AGI-2, a benchmark requiring strong abstraction and debugging capabilities, and adopt Prolog as the target language due to its declarative semantics, which are well-suited to algorithmic program debugging. Our experiments show that ABPR paired with Gemini-3-Flash achieves a Pass@2 score of 56.67\% even in a language in which contemporary LLMs typically underperform. These results point towards a more auditable paradigm for program repair by integrating LLMs with classical formal methods.

Modern autonomous multi-agent systems combine heterogeneous learning mechanisms operating at different timescales. An open question remains: can one formally guarantee that coupled dynamics of such mechanisms stay within the admissible operational regime? This paper studies a tri-hierarchical swarm learning system where three mechanisms act simultaneously: (1) local Hebbian online learning at individual agent level (fast timescale, 10-100 ms); (2) multi-agent reinforcement learning (MARL) for tactical group coordination (medium timescale, 1-10 s); (3) meta-learning (MAML) for strategic adaptation (slow timescale, 10-100 s). Four results are established. The Bounded Total Error Theorem shows that under contractual constraints on learning rates, Lipschitz continuity of inter-level mappings, and weight stabilization, total suboptimality admits a component-wise upper bound uniform in time. The Bounded Representation Drift Theorem gives a worst-case estimate of how Hebbian updates affect coordination-level embeddings during one MARL cycle. The Meta-Level Compatibility Theorem provides sufficient conditions under which strategic adaptation preserves lower-level invariants. The Non-Accumulation Theorem proves that error does not grow unboundedly over time.

We study forward and inverse problems for time-dependent probability measures in Bayes--Hilbert spaces. On the forward side, we show that each sufficiently regular Bayes--Hilbert path admits a canonical dynamical realization: a weighted Neumann problem transforms the log-density variation into the unique gradient velocity field of minimum kinetic energy. This construction induces a transport form on Bayes--Hilbert tangent directions, which measures the dynamical cost of realizing prescribed motions, and yields a flow-matching interpretation in which the canonical velocity field is the minimum-energy execution of the prescribed path. On the inverse side, we formulate reconstruction directly on Bayes--Hilbert path space from time-dependent indirect observations. The resulting variational problem combines a data-misfit term with the transport action induced by the forward geometry. In our infinite-dimensional setting, however, this transport geometry alone does not provide sufficient compactness, so we add explicit temporal and spatial regularization to close the theory. The linearized observation operator induces a complementary observability form, which quantifies how strongly tangent directions are seen through the data. Under explicit Sobolev regularity and observability assumptions, we prove existence of minimizers, derive first-variation formulas, establish local stability of the observation map, and deduce recovery of the evolving law, its score, and its canonical velocity field under the strong topologies furnished by the compactness theory.

Social insect colonies and ensemble machine learning methods represent two of the most successful examples of decentralized information processing in nature and computation respectively. Here we develop a rigorous mathematical framework demonstrating that ant colony decision-making and random forest learning are isomorphic under a common formalism of \textbf{stochastic ensemble intelligence}. We show that the mechanisms by which genetically identical ants achieve functional differentiation -- through stochastic response to local cues and positive feedback -- map precisely onto the bootstrap aggregation and random feature subsampling that decorrelate decision trees. Using tools from Bayesian inference, multi-armed bandit theory, and statistical learning theory, we prove that both systems implement identical variance reduction strategies through decorrelation of identical units. We derive explicit mappings between ant recruitment rates and tree weightings, pheromone trail reinforcement and out-of-bag error estimation, and quorum sensing and prediction averaging. This isomorphism suggests that collective intelligence, whether biological or artificial, emerges from a universal principle: \textbf{randomized identical agents + diversity-enforcing mechanisms $\rightarrow$ emergent optimality}.

Video world models trained with Joint Embedding Predictive Architectures (JEPA) acquire rich spatiotemporal representations by predicting masked regions in latent space rather than reconstructing pixels. This removes the visual verification pathway of generative models, creating a structural interpretability gap: the encoder has learned physical structure inaccessible in any inspectable form. Existing probing methods either operate in continuous space without a structured intermediate layer, or attach generative components whose parameters confound attribution of behavior to the encoder. We propose the AI Mother Tongue (AIM) framework as a passive quantization probe: a lightweight, vocabulary-free probe that converts V-JEPA 2 continuous latent vectors into discrete symbol sequences without task-specific supervision or modifying the encoder. Because the encoder is kept completely frozen, any symbolic structure in the AIM codebook is attributable entirely to V-JEPA 2 pre-trained representations -- not to the probe. We evaluate through category-contrast experiments on Kinetics-mini along three physical dimensions: grasp angle, object geometry, and motion temporal structure. AIM symbol distributions differ significantly across all three experiments (chi^2 p < 10^{-4}; MI 0.036--0.117 bits, NMI 1.2--3.9% of the 3-bit maximum; JSD up to 0.342; codebook active ratio 62.5%). The experiments reveal that V-JEPA 2 latent space is markedly compact: diverse action categories share a common representational core, with semantic differences encoded as graded distributional variations rather than categorical boundaries. These results establish Stage 1 of a four-stage roadmap toward an action-conditioned symbolic world model, demonstrating that structured symbolic manifolds are discoverable properties of frozen JEPA latent spaces.

Multi-agent LLM pipelines produce contradictory evidence on whether team diversity improves output quality: heterogeneous Mixture-of-Agents teams outperform single models, yet homogeneous Self-MoA teams consistently win under synthesis-based aggregation. We propose a resolution by identifying the selection bottleneck -- a crossover threshold in aggregation quality that determines whether diversity helps or hurts. Under this model, we obtain a closed-form crossover threshold $s^*$ (Proposition 1) that separates the regimes where diversity helps and hurts. In a targeted experiment spanning 42 tasks across 7 categories ($N=210$), a diverse team with judge-based selection achieves a win rate of 0.810 against a single-model baseline, while a homogeneous team scores 0.512 -- near chance (Glass's $Δ= 2.07$). Judge-based selection outperforms MoA-style synthesis by $Δ_{\mathrm{WR}} = +0.631$ -- the synthesis approach is preferred over the baseline in zero of 42 tasks by the judge panel. A decoupled evaluation with independent judges confirms all directional findings (Spearman $ρ= 0.90$). Exploratory evidence suggests that including a weaker model improves performance while reducing cost ($p < 10^{-4}$, not pre-registered). Our results suggest that selector quality may be a more impactful design lever than generator diversity in single-round generate-then-select pipelines.

Understanding mechanistic relationships among genes and their impacts on biological pathways is essential for elucidating disease mechanisms and advancing precision medicine. Despite the availability of extensive molecular interaction and pathway data in public databases, integrating heterogeneous knowledge sources and enabling interpretable multi-step reasoning across biological networks remain challenging. We present GIP-RAG (Gene Interaction Prediction through Retrieval-Augmented Generation), a computational framework that combines biomedical knowledge graphs with large language models (LLMs) to infer and interpret gene interactions. The framework constructs a unified gene interaction knowledge graph by integrating curated data from KEGG, WikiPathways, SIGNOR, Pathway Commons, and PubChem. Given user-specified genes, a query-driven module retrieves relevant subgraphs, which are incorporated into structured prompts to guide LLM-based stepwise reasoning. This enables identification of direct and indirect regulatory relationships and generation of mechanistic explanations supported by biological evidence. Beyond pairwise interactions, GIP-RAG includes a pathway-level functional impact module that simulates propagation of gene perturbations through signaling networks and evaluates potential pathway state changes. Evaluation across diverse biological scenarios demonstrates that the framework generates consistent, interpretable, and evidence-supported insights into gene regulatory mechanisms. Overall, GIP-RAG provides a general and interpretable approach for integrating knowledge graphs with retrieval-augmented LLMs to support mechanistic reasoning in complex molecular systems.

Large language models (LLMs) are increasingly deployed as agents with access to executable tools, enabling direct interaction with external systems. However, most safety evaluations remain text-centric and assume that compliant language implies safe behavior, an assumption that becomes unreliable once models are allowed to act. In this work, we empirically examine how executable tool affordance alters safety alignment in LLM agents using a paired evaluation framework that compares text-only chatbot behavior with tool-enabled agent behavior under identical prompts and policies. Experiments are conducted in a deterministic financial transaction environment with binary safety constraints across 1,500 procedurally generated scenarios. To separate intent from outcome, we distinguish between attempted and realized violations using dual enforcement regimes that either block or permit unsafe actions. Both evaluated models maintain perfect compliance in text-only settings, yet exhibit sharp increases in violations after tool access is introduced, reaching rates up to 85% despite unchanged rules. We observe substantial gaps between attempted and executed violations, indicating that external guardrails can suppress visible harm while masking persistent misalignment. Agents also develop spontaneous constraint circumvention strategies without adversarial prompting. These results demonstrate that tool affordance acts as a primary driver of safety misalignment and that text-based evaluation alone is insufficient for assessing agentic systems.

Space-based compute is becoming plausible as launch costs fall and data-intensive AI workloads grow. This paper proposes a workload-centric framework for deciding which tasks belong in orbit versus terrestrial cloud, along with a phased adoption model tied to orbital data center maturity. We ground the framework with in-orbit semantic-reduction prototypes. An Earth-observation pipeline on Sentinel-2 imagery from Seattle and Bengaluru (formerly Bangalore) achieves 99.7-99.99% payload reduction by converting raw imagery to compact semantic artifacts. A multi-pass stereo reconstruction prototype reduces ~306 MB to ~1.57 MB of derived 3D representations (99.49% reduction). These results support a workload-first view in which semantic abstraction, not raw compute scale, drives early workload suitability.

Answering financial questions is often treated as an information retrieval problem. In practice, however, much of the relevant information is already available in curated vendor systems, especially for quantitative analysis. We study whether, and under which conditions, Model Context Protocol (MCP) offers a more reliable alternative to standard retrieval-augmented generation (RAG) by allowing large language models (LLMs) to interact directly with data rather than relying on document ingestion and chunk retrieval. We test this by building a custom MCP server that exposes LSEG APIs as tools and evaluating it on the FinDER benchmark. The approach performs particularly well on the Financials subset, achieving up to 80.4% accuracy on multi-step numerical questions when relevant context is retrieved. The paper thus provides both a baseline for MCP-based financial question answering (QA) and evidence on where this approach breaks down, such as for questions requiring qualitative or document-specific context. Overall, direct access to curated data is a lightweight and effective alternative to document-centric RAG for quantitative financial QA, but not a substitute for all financial QA tasks.

(a) Many air quality forecasting studies report gains from machine learning, but evaluations often use static chronological splits and omit persistence baselines, so the operational added value under routine updating is unclear. (b) Using 2,350 daily PM10 observations from 2017 to 2024 at an urban background monitoring station in southern Europe, we compare XGBoost and SARIMA against persistence under a static split and a rolling-origin protocol with monthly updates. We report horizon-specific skill and the predictability horizon, defined as the maximum horizon with positive persistence-relative skill. Static evaluation suggests XGBoost performs well from one to seven days ahead, but rolling-origin evaluation reverses rankings: XGBoost is not consistently better than persistence at short and intermediate horizons, whereas SARIMA remains positively skilled across the full range. (c) For researchers, static splits can overstate operational usefulness and change rankings. For practitioners, rolling-origin, persistence-referenced skill profiles show which methods stay reliable at each lead time.

Medical Vision-Language Models (VLMs) often hallucinate by generating responses based on language priors rather than visual evidence, posing risks in clinical applications. We propose Visual Grounding Score Guided Decoding (VGS-Decoding), a training-free method to mitigate hallucinations during inference. Our key insight is that hallucinated tokens maintain or increase their probability when visual information is degraded, while visually grounded tokens decrease in probability. We introduce the Visual Grounding Score (VGS), which measures each token's visual dependency by comparing distributions from original and distorted images. During decoding, we reweight probabilities by amplifying visually grounded tokens while suppressing hallucinations. Unlike fixed-weight contrastive methods, VGS-Decoding provides per-token adaptive control. Experiments on MIMIC-Diff-VQA and VQA-RAD across LLaVA-Med, CheXagent, and MedGemma demonstrate consistent improvements, with up to +9.12% overall gain and $+8.98\%$ in open-ended recall, while introducing only $2\times$ inference overhead and no additional training, making it practical for clinical deployment. Upon acceptance, code will be released publicly to facilitate reproducibility.

Large Language Models (LLMs) with tool-calling capabilities have demonstrated remarkable potential in executing complex tasks through external tool integration. The Model Context Protocol (MCP) has emerged as a standardized framework for connecting LLMs to diverse toolsets, with individual MCP servers potentially exposing dozens to hundreds of tools. However, current implementations face a critical scalability challenge: providing all available tools to the LLM context results in substantial token overhead, increased costs, reduced accuracy, and context window constraints. We present a semantic tool discovery architecture that addresses these challenges through vector-based retrieval. Our approach indexes MCP tools using dense embeddings that capture semantic relationships between tool capabilities and user intent, dynamically selecting only the most relevant tools (typically 3-5) rather than exposing the entire tool catalog (50-100+). Experimental results demonstrate a 99.6% reduction in tool-related token consumption with a hit rate of 97.1% at K=3 and an MRR of 0.91 on a benchmark of 140 queries across 121 tools from 5 MCP servers, with sub-100ms retrieval latency. Contributions include: (1) a semantic indexing framework for MCP tools, (2) a dynamic tool selection algorithm based on query-tool similarity, (3) comprehensive evaluation demonstrating significant efficiency and accuracy improvements, and (4) extensibility to multi-agent and cross-organizational tool discovery.

Real-world financial decision-making is a challenging problem that requires reasoning over heterogeneous signals, including company fundamentals derived from regulatory filings and trading signals computed from price dynamics. Recently, with the advancement of Large Language Models (LLMs), financial analysts have begun to use them for financial decision-making tasks. However, existing financial question answering benchmarks for testing these models primarily focus on company balance sheet data and rarely evaluate reasoning over how company stocks trade in the market or their interactions with fundamentals. To take advantage of the strengths of both approaches, we introduce FinTradeBench, a benchmark for evaluating financial reasoning that integrates company fundamentals and trading signals. FinTradeBench contains 1,400 questions grounded in NASDAQ-100 companies over a ten-year historical window. The benchmark is organized into three reasoning categories: fundamentals-focused, trading-signal-focused, and hybrid questions requiring cross-signal reasoning. To ensure reliability at scale, we adopt a calibration-then-scaling framework that combines expert seed questions, multi-model response generation, intra-model self-filtering, numerical auditing, and human-LLM judge alignment. We evaluate 14 LLMs under zero-shot prompting and retrieval-augmented settings and witness a clear performance gap. Retrieval substantially improves reasoning over textual fundamentals, but provides limited benefit for trading-signal reasoning. These findings highlight fundamental challenges in the numerical and time-series reasoning for current LLMs and motivate future research in financial intelligence.

We introduce Nemotron-Cascade 2, an open 30B MoE model with 3B activated parameters that delivers best-in-class reasoning and strong agentic capabilities. Despite its compact size, its mathematical and coding reasoning performance approaches that of frontier open models. It is the second open-weight LLM, after DeepSeekV3.2-Speciale-671B-A37B, to achieve Gold Medal-level performance in the 2025 International Mathematical Olympiad (IMO), the International Olympiad in Informatics (IOI), and the ICPC World Finals, demonstrating remarkably high intelligence density with 20x fewer parameters. In contrast to Nemotron-Cascade 1, the key technical advancements are as follows. After SFT on a meticulously curated dataset, we substantially expand Cascade RL to cover a much broader spectrum of reasoning and agentic domains. Furthermore, we introduce multi-domain on-policy distillation from the strongest intermediate teacher models for each domain throughout the Cascade RL process, allowing us to efficiently recover benchmark regressions and sustain strong performance gains along the way. We release the collection of model checkpoint and training data.

Modern machine learning systems rely on complex data engineering workflows to extract, transform, and load (ELT) data into production pipelines. However, constructing these pipelines remains time-consuming and requires substantial expertise in data infrastructure and orchestration frameworks. Recent advances in large language model (LLM) agents offer a potential path toward automating these workflows, but existing approaches struggle with under-specified user intent, unreliable tool generation, and limited guarantees of executable outputs. We introduce kRAIG, an AI agent that translates natural language specifications into production-ready Kubeflow Pipelines (KFP). To resolve ambiguity in user intent, we propose ReQuesAct (Reason, Question, Act), an interaction framework that explicitly clarifies intent prior to pipeline synthesis. The system orchestrates end-to-end data movement from diverse sources and generates task-specific transformation components through a retrieval-augmented tool synthesis process. To ensure data quality and safety, kRAIG incorporates LLM-based validation stages that verify pipeline integrity prior to execution. Our framework achieves a 3x improvement in extraction and loading success and a 25 percent increase in transformation accuracy compared to state-of-the-art agentic baselines. These improvements demonstrate that structured agent workflows with explicit intent clarification and validation significantly enhance the reliability and executability of automated data engineering pipelines.

Cooperative perception among autonomous agents overcomes the limitations of single-agent sensing, but bandwidth constraints in vehicle-to-everything (V2X) networks require efficient communication policies. Existing approaches rely on reactive mechanisms, such as confidence maps, learned gating, or sparse masks, to decide what to transmit, without reasoning about why a message benefits the receiver. We introduce Reason-to-Transmit (R2T), a framework that equips each agent with a lightweight transformer-based module that reasons over local scene context, estimated neighbor information gaps, and bandwidth budget to make per-region transmission decisions. Trained end-to-end with a bandwidth-aware objective, R2T is evaluated against nine baselines in a multi-agent bird's-eye-view perception environment. Any communication improves performance by about 58% AP over no communication. At low bandwidth, all selective methods perform similarly, but R2T shows clear gains under high occlusion, where information asymmetry is greatest, approaching oracle performance. All methods degrade gracefully under packet drops up to 50%, showing robustness to communication failures. These results indicate that while fusion design dominates performance, deliberative communication provides additional gains in challenging scenarios. R2T introduces a reasoning-based approach to communication, enabling more efficient and context-aware information sharing in cooperative perception.

Audio-driven talking head generation aims to create vivid and realistic videos from a static portrait and speech. Existing AR-based methods rely on intermediate facial representations, which limit their expressiveness and realism. Meanwhile, diffusion-based methods generate clip-by-clip, lacking fine-grained control and causing inherent latency due to overall denoising across the window. To address these limitations, we propose EARTalking, a novel end-to-end, GPT-style autoregressive model for interactive audio-driven talking head generation. Our method introduces a novel frame-by-frame, in-context, audio-driven streaming generation paradigm. For inherently supporting variable-length video generation with identity consistency, we propose the Sink Frame Window Attention (SFA) mechanism. Furthermore, to avoid the complex, separate networks that prior works required for diverse control signals, we propose a streaming Frame Condition In-Context (FCIC) scheme. This scheme efficiently injects diverse control signals in a streaming, in-context manner, enabling interactive control at every frame and at arbitrary moments. Experiments demonstrate that EARTalking outperforms existing autoregressive methods and achieves performance comparable to diffusion-based methods. Our work demonstrates the feasibility of in-context streaming autoregressive control, unlocking a scalable direction for flexible, efficient generation. The code will be released for reproducibility.

Language models increasingly appear to learn similar representations, despite differences in training objectives, architectures, and data modalities. This emerging compatibility between independently trained models introduces new opportunities for cross-model alignment to downstream objectives. Moreover, it unlocks new potential application domains, such as settings where security, privacy, or competitive constraints prohibit direct data or model sharing. In this work, we propose a privacy-preserving framework that exploits representational convergence to enable cross-silo inference between independent language models. The framework learns an affine transformation over a shared public dataset and applies homomorphic encryption to protect client queries during inference. By encrypting only the linear alignment and classification operations, the method achieves sub-second inference latency while maintaining strong security guarantees. We support this framework with an empirical investigation into representational convergence, in which we learn linear transformations between the final hidden states of independent models. We evaluate these cross-model mappings on embedding classification and out-of-distribution detection, observing minimal performance degradation across model pairs. Additionally, we show for the first time that linear alignment sometimes enables text generation across independently trained models.

Popular language learning applications such as Duolingo use large language models (LLMs) to generate lessons for its users. Most lessons focus on general real-world scenarios such as greetings, ordering food, or asking directions, with limited support for profession-specific contexts. This gap can hinder learners from achieving professional-level fluency, which we define as the ability to communicate comfortably various work-related and domain-specific information in the target language. We surveyed five employees from a multinational company in the Philippines on their experiences with Duolingo. Results show that respondents encountered general scenarios more frequently than work-related ones, and that the former are relatable and effective in building foundational grammar, vocabulary, and cultural knowledge. The latter helps bridge the gap toward professional fluency as it contains domain-specific vocabulary. Each participant suggested lesson scenarios that diverge in contexts when analyzed in aggregate. With this understanding, we propose that language learning applications should generate lessons that adapt to an individual's needs through personalized, domain specific lesson scenarios while maintaining foundational support through general, relatable lesson scenarios.

Agent Control Protocol (ACP) is a formal technical specification for governance of autonomous agents in B2B institutional environments. ACP acts as an admission control layer between agent intent and system state mutation: before execution, every agent action must pass a cryptographic admission check that validates identity, capability scope, delegation chain, and policy compliance. ACP defines mechanisms for cryptographic identity, capability-based authorization, deterministic risk evaluation, verifiable chained delegation, transitive revocation, and immutable auditing, enabling autonomous agents to operate under explicit institutional control. ACP operates as an additional layer on top of RBAC and Zero Trust, without replacing them. It addresses a gap these models do not solve: governing what autonomous agents can do, under what conditions, with what limits, and with full traceability for external auditing, including across organizational boundaries. The specification includes a multi-organization interoperability model in which independently governed systems validate cross-organizational execution requests through a shared verification pipeline. Divergence between policy evaluations is detected and reported, but not resolved by the protocol, preserving institutional sovereignty. All cryptographic operations use Ed25519 with JCS canonicalization. The specification is language-agnostic, with a reference implementation in Go.

Currently, the methods called Iterative Ensemble Smoothers, especially the method called Ensemble Smoother with Multiple Data Assimilation (ESMDA) can be considered state-of-the-art for history matching in petroleum reservoir simulation. However, this approach has two important limitations: the use of an ensemble with finite size to represent the distributions and the Gaussian assumption in parameter and data uncertainties. This latter is particularly important because many reservoir properties have non-Gaussian distributions. Parameterization involves mapping non-Gaussian parameters to a Gaussian field before the update and then mapping them back to the original domain to forward the ensemble through the reservoir simulator. A promising approach to perform parameterization is through deep learning models. Recent studies have shown that Generative Adversarial Networks (GAN) performed poorly concerning data assimilation, but generated more geologically plausible realizations of the reservoir, while the Variational Autoencoder (VAE) performed better than the GAN in data assimilation, but generated less geologically realistic models. This work is innovative in combining the strengths of both to implement a deep learning model called Variational Autoencoder Generative Adversarial Network (VAE-GAN) integrated with ESMDA. The methodology was applied in two case studies, one case being categorical and the other with continuous values of permeability. Our findings demonstrate that by applying the VAE-GAN model we can obtain high quality reservoir descriptions (just like GANs) and a good history matching on the production curves (just like VAEs) simultaneously.

Flow Matching (FM) has recently emerged as a leading approach for high-fidelity visual generation, offering a robust continuous-time alternative to ordinary differential equation (ODE) based models. However, despite their success, FM models are highly sensitive to dataset biases, which cause severe semantic degradation when generating out-of-distribution or minority-class samples. In this paper, we provide a rigorous mathematical formalization of the ``Bias Manifold'' within the FM framework. We identify that this performance drop is driven by conditional expectation smoothing, a mechanism that inevitably leads to trajectory lock-in during inference. To resolve this, we introduce InjectFlow, a novel, training-free method by injecting orthogonal semantics during the initial velocity field computation, without requiring any changes to the random seeds. This design effectively prevents the latent drift toward majority modes while maintaining high generative quality. Extensive experiments demonstrate the effectiveness of our approach. Notably, on the GenEval dataset, InjectFlow successfully fixes 75% of the prompts that standard flow matching models fail to generate correctly. Ultimately, our theoretical analysis and algorithm provide a ready-to-use solution for building more fair and robust visual foundation models.

Voice privacy approaches that preserve the anonymity of speakers modify speech in an attempt to break the link with the true identity of the speaker. Current benchmarks measure speaker protection based on signal-to-signal comparisons. In this paper, we introduce an attribute-based perspective, where we measure privacy protection in terms of comparisons between sets of speaker attributes. First, we analyze privacy impact by calculating speaker uniqueness for ground truth attributes, attributes inferred on the original speech, and attributes inferred on speech protected with standard anonymization. Next, we examine a threat scenario involving only a single utterance per speaker and calculate attack error rates. Overall, we observe that inferred attributes still present a risk despite attribute inference errors. Our research points to the importance of considering both attribute-related threats and protection mechanisms in future voice privacy research.

The No-U-Turn Sampler (NUTS) is the computational workhorse of modern Bayesian software libraries, yet its qualitative and quantitative convergence guarantees were established only recently. A significant gap remains in the theoretical comparison of its two main variants: NUTS-mul and NUTS-BPS, which use multinomial sampling and biased progressive sampling, respectively, for index selection. In this paper, we address this gap in three contributions. First, we derive the first necessary conditions for geometric ergodicity for both variants. Second, we establish the first sufficient conditions for geometric ergodicity and ergodicity for NUTS-mul. Third, we obtain the first mixing time result for NUTS-BPS on a standard Gaussian distribution. Our results show that NUTS-mul and NUTS-BPS exhibit nearly identical qualitative behavior, with geometric ergodicity depending on the tail properties of the target distribution. However, they differ quantitatively in their convergence rates. More precisely, when initialized in the typical set of the canonical Gaussian measure, the mixing times of both NUTS-mul and NUTS-BPS scale as $O(d^{1/4})$ up to logarithmic factors, where $d$ denotes the dimension. Nevertheless, the associated constants are strictly smaller for NUTS-BPS.

Software systems have traditionally been designed for human interaction, emphasizing graphical user interfaces, usability, and cognitive alignment with end users. However, recent advances in large language model (LLM)-based agents are changing the primary consumers of software systems. Increasingly, software is no longer only used by humans, but also invoked autonomously by AI agents through structured interfaces. In this paper, we argue that software engineering is undergoing a paradigm shift from human-oriented interfaces to agent-oriented invocation systems. We formalize the notion of agent interfaces, introduce invocable capabilities as the fundamental building blocks of AI-oriented software, and outline design principles for such systems, including machine interpretability, composability, and invocation reliability. We then discuss architectural and organizational implications of this shift, highlighting a transition from monolithic applications to capability-based systems that can be dynamically composed by AI agents. The paper aims to provide a conceptual foundation for the emerging paradigm of AI-native software design.

Existing Retrieval-Augmented Generation (RAG) methods for code struggle to capture the high-level architectural patterns and cross-file dependencies inherent in complex, theory-driven codebases, such as those in algorithmic game theory (AGT), leading to a persistent semantic and structural gap between abstract concepts and executable implementations. To address this challenge, we propose Hierarchical Code/Architecture-guided Agent Generation (HCAG), a framework that reformulates repository-level code generation as a structured, planning-oriented process over hierarchical knowledge. HCAG adopts a two-phase design: an offline hierarchical abstraction phase that recursively parses code repositories and aligned theoretical texts to construct a multi-resolution semantic knowledge base explicitly linking theory, architecture, and implementation; and an online hierarchical retrieval and scaffolded generation phase that performs top-down, level-wise retrieval to guide LLMs in an architecture-then-module generation paradigm. To further improve robustness and consistency, HCAG integrates a multi-agent discussion inspired by cooperative game. We provide a theoretical analysis showing that hierarchical abstraction with adaptive node compression achieves cost-optimality compared to flat and iterative RAG baselines. Extensive experiments on diverse game-theoretic system generation tasks demonstrate that HCAG substantially outperforms representative repository-level methods in code quality, architectural coherence, and requirement pass rate. In addition, HCAG produces a large-scale, aligned theory-implementation dataset that effectively enhances domain-specific LLMs through post-training. Although demonstrated in AGT, HCAG paradigm also offers a general blueprint for mining, reusing, and generating complex systems from structured codebases in other domains.

Accurate calibration is essential for instruments whose measurements must remain traceable, reliable, and compliant over long operating periods. Fixed-interval programs are easy to administer, but they ignore that instruments drift at different rates under different conditions. This paper studies calibration scheduling as a predictive maintenance problem: given recent sensor histories, estimate time-to-drift (TTD) and intervene before a violation occurs. We adapt the NASA C-MAPSS benchmark into a calibration setting by selecting drift-sensitive sensors, defining virtual calibration thresholds, and inserting synthetic reset events that emulate repeated recalibration. We then compare classical regressors, recurrent and convolutional sequence models, and a compact Transformer for TTD prediction. The Transformer provides the strongest point forecasts on the primary FD001 split and remains competitive on the harder FD002--FD004 splits, while a quantile-based uncertainty model supports conservative scheduling when drift behavior is noisier. Under a violation-aware cost model, predictive scheduling lowers cost relative to reactive and fixed policies, and uncertainty-aware triggers sharply reduce violations when point forecasts are less reliable. The results show that condition-based calibration can be framed as a joint forecasting and decision problem, and that combining sequence models with risk-aware policies is a practical route toward smarter calibration planning.

Recent advances in collaborative knowledge distillation have demonstrated cutting-edge performance for resource-constrained distributed multimedia learning scenarios. However, achieving such competitiveness requires addressing a fundamental mismatch: high-dimensional teacher knowledge complexity versus heterogeneous client learning capacities, which currently prohibits deployment in edge-based visual analytics systems. Drawing inspiration from curriculum learning principles, we introduce Federated Adaptive Progressive Distillation (FAPD), a consensus-driven framework that orchestrates adaptive knowledge transfer. FAPD hierarchically decomposes teacher features via PCA-based structuring, extracting principal components ordered by variance contribution to establish a natural visual knowledge hierarchy. Clients progressively receive knowledge of increasing complexity through dimension-adaptive projection matrices. Meanwhile, the server monitors network-wide learning stability by tracking global accuracy fluctuations across a temporal consensus window, advancing curriculum dimensionality only when collective consensus emerges. Consequently, FAPD provably adapts knowledge transfer pace while achieving superior convergence over fixed-complexity approaches. Extensive experiments on three datasets validate FAPD's effectiveness: it attains 3.64% accuracy improvement over FedAvg on CIFAR-10, demonstrates 2x faster convergence, and maintains robust performance under extreme data heterogeneity (α=0.1), outperforming baselines by over 4.5%.

Uncovering causal structures from observational data is crucial for understanding complex systems and making informed decisions. While reinforcement learning (RL) has shown promise in identifying these structures in the form of a directed acyclic graph (DAG), existing methods often lack efficiency, making them unsuitable for online applications. In this paper, we propose MARLIN, an efficient multi agent RL based approach for incremental DAG learning. MARLIN uses a DAG generation policy that maps a continuous real valued space to the DAG space as an intra batch strategy, then incorporates two RL agents state specific and state invariant to uncover causal relationships and integrates these agents into an incremental learning framework. Furthermore, the framework leverages a factored action space to enhance parallelization efficiency. Extensive experiments on synthetic and real datasets demonstrate that MARLIN outperforms state of the art methods in terms of both efficiency and effectiveness.

Text-attributed graphs, where nodes are enriched with textual attributes, have become a powerful tool for modeling real-world networks such as citation, social, and transaction networks. However, existing methods for learning from these graphs often assume that the distributions of training and testing data are consistent. This assumption leads to significant performance degradation when faced with out-of-distribution (OOD) data. In this paper, we address the challenge of node-level OOD detection in text-attributed graphs, with the goal of maintaining accurate node classification while simultaneously identifying OOD nodes. We propose a novel approach, LLM-Enhanced Energy Contrastive Learning for Out-of-Distribution Detection in Text-Attributed Graphs (LECT), which integrates large language models (LLMs) and energy-based contrastive learning. The proposed method involves generating high-quality OOD samples by leveraging the semantic understanding and contextual knowledge of LLMs to create dependency-aware pseudo-OOD nodes, and applying contrastive learning based on energy functions to distinguish between in-distribution (IND) and OOD nodes. The effectiveness of our method is demonstrated through extensive experiments on six benchmark datasets, where our method consistently outperforms state-of-the-art baselines, achieving both high classification accuracy and robust OOD detection capabilities.

In the context of the rapid development of digital finance, some financial technology companies exhibit the phenomenon of "AI washing," where they overstate their AI capabilities while underinvesting in actual AI resources. This paper constructs a corporate-level AI washing index based on CHFS2019 data and AI investment data from 15-20 financial technology companies, analyzing and testing its impact on farmers' digital financial behavior response. The study finds that AI washing significantly suppresses farmers' digital financial behavior; the higher the degree of AI washing, the lower the response level of farmers' digital financial behavior. Moreover, AI washing indirectly inhibits farmers' behavioral responses by exacerbating knowledge exclusion and risk exclusion. Social capital can positively moderate the negative impact of AI washing; among farmer groups with high social capital, the suppressive effect of AI washing on digital financial behavior is significantly weaker than that among groups with low social capital. In response, this paper suggests that regulatory authorities establish a strict information disclosure system for AI technology, conduct differentiated digital financial education to enhance the identification capabilities of vulnerable groups, promote digital financial mutual aid groups to leverage the protective effects of social capital, improve the consumer protection mechanism for farmers in digital finance, and set up pilot "Digital Inclusive Finance Demonstration Counties," etc.

At a time when the phenomenon of 'AI washing' is quietly spreading, an increasing number of enterprises are using the label of artificial intelligence merely as a cosmetic embellishment in their annual reports, rather than as a genuine engine driving transformation. A test regarding the essence of innovation and the authenticity of information disclosure has arrived. This paper employs large language models to conduct semantic analysis on the text of annual reports from Chinese A-share listed companies from 2006 to 2024, systematically examining the impact of corporate AI washing behaviour on their green innovation. The research reveals that corporate AI washing exerts a significant crowding-out effect on green innovation, with this negative relationship transmitted through dual channels in both product and capital markets. Furthermore, this crowding-out effect exhibits heterogeneity across firms and industries, with private enterprises, small and medium-sized enterprises (SMEs), and firms in highly competitive sectors suffering more severe negative impacts from AI washing. Simulation results indicate that a combination of policy tools can effectively improve market equilibrium. Based on this, this paper proposes that the government should design targeted support tools to 'enhance market returns and alleviate financing constraints', adopt a differentiated regulatory strategy, and establish a disclosure mechanism combining 'professional identification and reputational sanctions' to curb such peer AI washing behaviour.

In incremental classification tasks for hyperspectral images, catastrophic forgetting is an unavoidable challenge. While memory recall methods can mitigate this issue, they heavily rely on samples from old categories. This paper proposes a teacher-based knowledge retention method for incremental image classification. It alleviates model forgetting of old category samples by utilizing incremental category samples, without depending on old category samples. Additionally, this paper introduces a mask-based partial category knowledge distillation algorithm. By decoupling knowledge distillation, this approach filters out potentially misleading information that could misguide the student model, thereby enhancing overall accuracy. Comparative and ablation experiments demonstrate the proposed method's robust performance.

Causal representation learning (CRL) aims to learn low-dimensional causal latent variables from high-dimensional observations. While identifiability has been extensively studied for CRL, estimation has been less explored. In this paper, we explore the use of empirical Bayes (EB) to estimate causal representations. In particular, we consider the problem of learning from data from multiple domains, where differences between domains are modeled by interventions in a shared underlying causal model. Multi-domain CRL naturally poses a simultaneous inference problem that EB is designed to tackle. Here, we propose an EB $f$-modeling algorithm that improves the quality of learned causal variables by exploiting invariant structure within and across domains. Specifically, we consider a linear measurement model and interventional priors arising from a shared acyclic SCM. When the graph and intervention targets are known, we develop an EM-style algorithm based on causally structured score matching. We further discuss EB $g$-modeling in the context of existing CRL approaches. In experiments on synthetic data, our proposed method achieves more accurate estimation than other methods for CRL.

Bus holding control is challenging due to stochastic traffic and passenger demand. While deep reinforcement learning (DRL) shows promise, standard actor-critic algorithms suffer from Q-value instability in volatile environments. A key source of this instability is the conflation of two distinct uncertainties: aleatoric uncertainty (irreducible noise) and epistemic uncertainty (data insufficiency). Treating these as a single risk leads to value underestimation in noisy states, causing catastrophic policy collapse. We propose a robust ensemble soft actor-critic (RE-SAC) framework to explicitly disentangle these uncertainties. RE-SAC applies Integral Probability Metric (IPM)-based weight regularization to the critic network to hedge against aleatoric risk, providing a smooth analytical lower bound for the robust Bellman operator without expensive inner-loop perturbations. To address epistemic risk, a diversified Q-ensemble penalizes overconfident value estimates in sparsely covered regions. This dual mechanism prevents the ensemble variance from misidentifying noise as a data gap, a failure mode identified in our ablation study. Experiments in a realistic bidirectional bus corridor simulation demonstrate that RE-SAC achieves the highest cumulative reward (approx. -0.4e6) compared to vanilla SAC (-0.55e6). Mahalanobis rareness analysis confirms that RE-SAC reduces Oracle Q-value estimation error by up to 62% in rare out-of-distribution states (MAE of 1647 vs. 4343), demonstrating superior robustness under high traffic variability.

Out-of-plane and in-plane electric polarization, which rarely coexist in a two-dimensional (2D) ferroelectric material, offer different advantages in ferroelectricity-based devices. Here, we report the coexistence of in-plane and out-of-plane electric dipoles, along with various phase transitions, in 2D van der Waals layered TlGaS2 single crystal. Quantum paraelectricity was observed along both in-plane and out-of-plane directions of the TlGaS2 crystal. Detailed investigation of the quantum paraelectric soft-mode behavior reveals a close correlation between the electric dipoles and the off-center displacement of Tl1+ ions with 6s2 lone pairs in TlGaS2. Anomalies near temperatures of about 120 K and 60-75 K in dielectric and/or infrared spectra indicate the existence of local or weak long-range structural transitions in TlGaS2. Our results provide important experimental evidence for elucidating the phase transitions and coexistence of in-plane and out-of-plane electric dipoles in 2D layered TlGaS2.

Elasticity has long been regarded as a property exclusive to material media. Here we uncover its hidden existence in the spin degree of freedom. We introduce spin elasticity-an intrinsic mechanism that governs recoverable deformation of spin morphology. This discovery reveals a previously unrecognized universality: elasticity operates in both matter and spin spaces, underpinning structural integrity across physical realms. By establishing the missing spin counterpart, this work completes the elastic picture and points toward a broader paradigm where elasticity transcends its conventional boundaries.

In moiré graphene systems, electronic interactions lift spin and valley degeneracies, leading to symmetry-broken ground states. In helical trilayer graphene (HTG), we uncover a distinct interaction-driven mechanism in which the roles of sublattice-polarized valence and conduction bands are cyclically reversed. Using scanning nano-SQUID magnetometry, we detect a series of sharp magnetic signatures consistent with seesaw-like transitions, where occupied and unoccupied valence and conduction bands interchange repeatedly with doping, accompanied by a novel form of magnetic hysteresis. These transitions occur entirely within metallic regimes and leave only weak fingerprints in transport measurements. Self-consistent Hartree-Fock calculations reveal that interactions reorganize all eight low-energy flat bands, driving abrupt changes in orbital magnetization. Our results establish HTG as the first system where electronic interactions provide doping-controlled access to all three internal degrees of freedom - spin, valley, and sublattice - introducing a new class of correlated phase transitions.

Chiral-induced spin selectivity (CISS) is a striking phenomenon in which spin-unpolarized electrons become spin-polarized after traversing a chiral medium. Theoretical studies have shown that spin-orbit coupling, geometric chirality, and dephasing act cooperatively for this effect to emerge. Inspired by this, we demonstrate a solid-state realization of CISS in an engineered InAs/GaSb quantum well where geometric chirality and dephasing can be introduced controllably. Introducing a chiral structure produces a clear spin polarization whose sign reverses when the chirality is flipped, and whose magnitude grows systematically with the number of dephasing electrodes, while achiral configurations exhibit no spin selectivity. The polarization remains robust even under strong Anderson disorder, showing that the engineered chiral structures provides an intrinsically stable route to spin-selective transport. These results establish a solid-state platform in the topological quantum well system for controllably generating the CISS effect.

Superconductivity in the iron-chalcogenide series FeSe-Fe(Te, Se)-FeTe has been restricted to the near neighbor of iron selenide (FeSe), with a general consensus that iron telluride (FeTe) is not superconducting. In this study, we report the method to grow FeTe islands with atomically flat surface and hexagonal lattice on SrTiO3 (001) substrates, in which a gap structure with a gap-filling temperature close to 40 K is detected by scanning tunneling spectroscopy. Such signature is examined under various conditions and reminiscent of a superconducting gap structure. This work might offer a potential platform to explore new superconductors at ambient pressure.

Higher order topology, in the form of the emergence of corner modes, is observed in two dimensions when crystalline symmetries are superposed on the Altland-Zirnbauer classification of topological insulators. It occurs in Benalcazar-Bernevig-Hughes (BBH) model on a 2D square lattice, which owing to an embedded $\mathbb{Z}_2$ gauge field, features a bulk quadrupole moment with localized zero-energy corner states. Further, as a dividend, the BBH model transmutes the general notion of the space-time inversion ($\mathcal{PT}$) symmetry and behaves as a spinful system, without having to invoke `real' spin degrees of freedom. A two-fold engineering of the model, namely a periodic drive, followed by a non-reciprocal hopping render intriguing consequences. As a first, the drive activates first-order topology, and the resulting Floquet phase hosts a coexistence of first-order conducting edges at both zero and $π$ quasienergies along with higher order corner states, which qualifies the coexisting state to be denoted as a hybrid-order topological phase. Further, inclusion of non-reciprocal couplings features a $\mathbb{Z}_2$-like skin effect which demonstrates a drive-induced transition from a unipolar to a bipolar localized phase, and is evidenced via the generalized Brillouin zone (GBZ) theory. While depiction of a GBZ in 2D is challenging, a corresponding 1D map is still possible and can be implemented by exploiting the mirror symmetry. We further uncover conditions under which the skin effect is completely suppressed in our system. Putting together, our results manifest an efficient technique to dynamically engender and control Hermitian and non-Hermitian topological features, that remain otherwise masked in a static scenario.

We investigate the nonequilibrium dynamics of an open photon Bose-Einstein condensate in a dye-filled microcavity using a Lindblad master-equation approach, treating the condensate and the noncondensed fluctuations on the same footing. The driven-dissipative condensate exhibits a long-lived, metastable plateau stabilized by a ghost attractor, a fixed point that lies outside the physical domain in configuration space, yet stalls the condensate dynamics for exceedingly long times before it dephases to zero [Phys. Rev. Lett. 135, 053402 (2025)]. Despite the nonequilibrium origin of this dynamical stabilization, the condensate exhibits quasithermal fluctuations in the plateau in that the relative order-parameter fluctuations scale as the inverse square root of the system size. A linear stability analysis further reveals the presence of exceptional points, resulting in multiple non-Hermitian phase transitions associated with the relaxation dynamics into and out of the metastable condensate.

State-of-the-art metallic terahertz (THz) emitters rely predominantly on spintronic heterostructures, where heavy metals serve as passive spin-to-charge converters. Here, we demonstrate efficient THz radiation from standalone Pt nanofilms at cryogenic temperatures and under external magnetic fields. The governing mechanism is identified as the ultrafast photo-Nernst effect, wherein a transient thermal gradient drives a transverse charge current. The THz emission polarity is directly dictated by the sign of the Nernst coefficient, as verified by the phase reversal observed between Pt and W or Ta. Remarkably, both thickness scaling and alloying-induced suppression of thermal conductivity independently amplify the single-layer emission to levels comparable with benchmark spintronic bilayers. These findings redefine the established role of heavy metals from passive spin-sinks to active THz emitters, uncovering a universal emission paradigm applicable across diverse spintronic and quantum materials.

Actin flow in the cortical cytoskeleton underneath the cell membrane generates mechanical stresses that shape the cell surface. We study this mechanism using an hydrodynamic model of a compressible active gel polymerising at the membrane and undergoing turnover. We determine how actin flow, density relaxation and friction of actin with the membrane generate stress on a corrugated membrane at the linear order in deformation. Analytical solutions in limiting regimes, combined with finite element methods in the general case, provide a map of normal and tangential stresses as functions of compressibility, interfacial friction and actin turnover, and determine the conditions under which actin polymerisation can render the membrane linearly unstable. The non-linear regime is also briefly discussed.

The appearance of emergent symmetries in complex systems with components that can form composite units provides us with opportunities for design and control of exotic phase behaviour, for example by exploiting the dynamical symmetry breaking associated with them. We present a novel mechanism for the emergence of non-reciprocal interactions in a single-species suspension of chemically active colloids made out of semi-permeable vesicles, which encapsulate enzymes that catalyze a non-linear chemical reaction. Bistable chemical dynamics enables the colloidal reaction chamber to act as a net producer or consumer of a chemical, depending on the selected values of the chemical concentrations inside and around it. Since the internal chemical state of the colloid depends on the dynamic chemical concentrations rather than the material parameters, two identically produced colloids can present different effective chemical interactions within the same system upon responding to the corresponding gradients via diffusiophoresis. Furthermore, the colloids can spontaneously and reversibly switch between being effective consumers or producers. As a consequence, the colloids can dynamically switch between ignoring, attracting, repelling, and chasing each other, in a non-reciprocal manner. This flexibility can be exploited by manipulation of tuning parameters to induce bifurcations in the chemical dynamics, resulting in a robust control over the interaction motifs, and rich emergent dynamics such as spontaneous many-body polar swarming.

Ion channels regulate many essential properties of biological cells, especially the membrane potential. Despite decades of efforts on artificial channels, it remains a great challenge to mimic the dipole potential-an indispensable constituent of the membrane potential, due to its angstrom-scale characteristic length. Here, we explore nanopores in monolayer molybdenum sulfide selenide (MoSSe) considering its intrinsic dipole and atomic thickness. Remarkably, an invariant ionic conductance was observed over salt concentrations spanning six orders of magnitude, distinct from all known conductance-concentration scaling laws and reminiscent of the current saturation in cell membranes at high concentrations. Molecular dynamics simulations revealed the fundamental role of the dipole-modulated dielectric properties of nanoconfined water. Our findings highlight an exotic conductance scaling law and open up a novel avenue for controlling ion transport in unprecedented ways.

The ultrawide-bandgap semiconductor $β$-Ga2O3 holds exceptional promise for next-generation power electronics and deep-ultraviolet optoelectronics, yet its widespread application is hindered by the lack of cost-effective, high-quality heteroepitaxial thin films. Here, we demonstrate an interpretable machine learning framework that efficiently navigates the complex, multiparameter process space of pulsed laser deposition (PLD) to achieve high-crystallinity $β$-Ga2O3 epitaxy on c-plane sapphire. By systematically benchmarking nine regression algorithms under limited experimental data conditions, we identify quadratic polynomial ridge regression as the optimal surrogate model, which combines predictive accuracy (R$^2$ $\approx$ 0.86) with full physical transparency through explicit analytical coefficients. Coupling this model with SHAP (SHapley Additive exPlanations) analysis and iterative experimental design, we construct a closed-loop optimization workflow that progressively refines the process-performance landscape over only three experimental rounds. This data-efficient strategy reduces the X-ray rocking curve (RC) full-width at half-maximum (FWHM) by 70$\%$ from > 3$^{\circ}$ to 0.92$^{\circ}$, which is the best reported value for PLD-grown $β$-Ga2O3 on sapphire. Intriguingly, concurrent modeling of surface roughness reveals that crystalline quality and surface morphology are governed by distinct dominant factors: temperature primarily controls bulk crystallinity, whereas oxygen pressure dictates surface kinetics. This decoupled mechanism, quantitatively captured for the first time via feature importance analysis, provides actionable physical insight for independent optimization of structural and morphological properties. Our work establishes a generalizable, resource-efficient paradigm for intelligent process development in oxide epitaxy and beyond.

We study the many-body localization (MBL) transition in interacting fermionic systems on disordered one-dimensional lattices using a physics-informed machine-learning framework. Instead of feeding full many-body wave functions into the model, we construct a compact feature representation based on four physically motivated observables: the inverse participation ratio, the Shannon entropy, the many-body hybridization parameter, and the mean level-spacing ratio. These quantities capture complementary aspects of localization, entanglement, and spectral correlations, and are used to train a Kolmogorov--Arnold Network (KAN) classifier on eigenstates deep in the weak and strong disorder regimes. The resulting KAN achieves a validation accuracy exceeding $99.9\%$, comparable to that of convolutional neural networks trained directly on high-dimensional wave-function data, while requiring substantially reduced input dimensionality and significantly shorter training time. Applying the trained classifier across the full energy spectrum yields energy-resolved phase diagrams that reveal a clear many-body mobility edge and provide a consistent estimate of the critical disorder strength. The approach is inherently extensible: additional physically relevant observables can be incorporated into the feature space in a systematic manner without altering the overall architecture. Our results demonstrate that feature-based learning with KAN provides an efficient, scalable, and interpretable methodology for identifying many-body localization transitions, offering a practical alternative to raw-data-based neural network approaches.

We analyze states of stationary activity in randomly coupled quadratic integrate-and-fire neurons using stochastic mean-field theory. Specifically, we consider the two cases of Gaussian random coupling and Cauchy random coupling, which are representative of systems with light- or with heavy-tailed synaptic strength distributions. For both, Gaussian and Cauchy coupling, bistability between a low activity and a high activity state of self-sustained firing is possible in excitable neurons. In the system with Cauchy coupling we find analytically a directed percolation threshold, i.e., above a critical value of the synaptic strength, activity percolates through the whole network starting from a few spiking units only. The existence of the directed percolation threshold is in agreement with previous numerical results in the literature for integrate-and-fire neurons with heavy-tailed synaptic strength distribution. However, we have found that the transition can be continuous or discontinuous, depending on the excitatory-inhibitory imbalance in the network. Networks with Gaussian coupling and networks with Cauchy coupling and additional additive noise lack the percolation transition in the thermodynamic limit.

Recent studies have identified materials and devices whose behavior lies beyond the scope of conventional electronic-structure theory. Such theories are formulated entirely in terms of Hamiltonian evolution and therefore describe only unitary dynamics and thus only a restricted class of quantum systems. In contrast, electron systems that incorporate quantum measurement as an intrinsic dynamical element undergo Hamiltonian evolution interleaved with projection-induced state updates. This unitary-projective dynamics breaks constraints imposed by purely unitary evolution and permits stochastic population transfer between symmetry-related transport channels, thereby enabling fundamentally new material functionalities. This insight motivates the deliberate design of materials and devices that harness unitary-projective dynamics. This article explores the foundations of unitary-projective electron dynamics and charts the resulting landscape of quantum materials and their functionalities. Model calculations demonstrate passive mesoscopic structures with intrinsic nonreciprocal single-electron transmission, materials exhibiting a novel category of magnetism, and possible platforms for energy harvesting and conversion with efficiencies that exceed the standard Carnot limit.

This paper is accountable only to explicitly stated physical assumptions and strict logical inference. Its goal is to run a rigorous stress test of second-law claims within the Clausius framework. We work directly with \textbf{Clausius's entropy definition} for an isolated composite with energy-form conversion. Heat is withdrawn from a cold releasing subsystem with relatively small heat capacity, converted to electrical energy, and then delivered as heat to a hotter subsystem. In the ideal limit, the electrical leg contributes negligibly to Clausius entropy accounting, so the modeled reservoir Clausius sum is \[ ΔS_{\mathrm{Cl}} = Q\!\left(\frac{1}{T_B}-\frac{1}{T_A}\right) < 0. \] The paper provides a derivation, numerical illustrations, and a scope analysis; any claimed contradiction should be interpreted as a compatibility issue between different axiom sets, not as an algebraic error in the Clausius bookkeeping above.

Spin currents can be generated through various mechanisms, including the piezospintronic effect, which arises when strain or lattice distortions induce a change in the dipolar spin moment, causing a pure spin current without necessarily being accompanied by net charge transport. This opens new possibilities for low-power information processing and novel device architectures. In this work, we propose a novel effect, the spintronic-magneto-impedictive effect, as the theoretical basis for a pure spin-current memory-like device based on antiferromagnetic components. We focus on materials that can be modeled by the so-called spin-Rice-Mele Hamiltonian, incorporating a magnetic field gradient that explicitly breaks inversion symmetry. Our results shed light on how spin currents are generated and controlled, providing new insights into the potential of these materials for next-generation spintronic technologies.

Band structure engineering in surface acoustic wave (SAW) metamaterials could advance both classical telecommunications and quantum information processing. However, no imaging technique has demonstrated the necessary capability to resolve sub-$μ$m traveling SAWs across wide GHz bandwidths. Existing methods capture only fragments of the dispersion at discrete frequencies, preventing systematic characterization and control of SAW-based metamaterials. Here, we develop electrostatic force microscopy (EFM) to enable real-space imaging of traveling SAWs in honeycomb metamaterials on LiNbO$_3$. Our application leverages sub-200 nm spatial resolution, broad GHz bandwidth, and non-contact imaging to map complex band structures with continuous frequency resolution and expanded frequency range, while preserving sub-lattice detail. Using EFM, we map the full relevant frequency range around the Dirac point of a SAW graphene analog, including the acoustic Dirac cones, and the transition from ballistic to diffusive SAW transport regime. Furthermore, by breaking sublattice symmetry, we tune the opening of a band gap at the Dirac point, and image frequency-dependent wave localization on sublattice sites. Our EFM technique closes the loop between design and real-space validation, streamlining the engineering of arbitrary SAW landscapes for next-generation applications spanning telecommunications, microfluidics, and quantum acoustics.

Accurate descriptions of reference systems are a central task in liquid-state theories for the study of more complex systems. Using scaled particle theory (SPT), we derive a fully analytical description of the thermodynamic properties of a hard-sphere (HS) fluid confined in size-polydisperse HS random porous media, extending the existing approaches to higher matrix packing fractions. We calculate chemical potentials for a wide range of porous-matrix parameters, including the matrix packing fraction, degree of polydispersity, and particle-size distributions. Within the proposed framework, our results show excellent agreement with available Monte Carlo simulations and previous integral-equation theories over a broad range of matrix packing fractions, $0.1 \leqslant η_0 \leqslant 0.3$, and degrees of polydispersity.

Non-equilibrium dynamics of strongly and rapidly driven quantum many-body systems is poorly understood beyond periodic driving, where heating is exponentially slow in the drive frequency (Floquet Prethermalization). In contrast, non-periodic drives were found to exhibit widely different heating scalings with no unifying principle. This work identifies a resonance-suppression principle governing slow heating up to a prethermal lifetime $τ_*$: When the drive's spectral arithmetic structure restricts multiphoton resonances, $τ_*$ is controlled by low-frequency spectral suppression. The principle distinguishes (i) Single-photon suppression, quantified by a low-frequency suppression law $f(Ω)$ for the drive's Fourier Transform weight near $Ω=0$, from (ii) Multi-photon suppression, where nested commutators remain controlled if exceptional arithmetic structure satisfies a subadditive property. Remarkably, if multi-photon suppression holds, $τ_*$ scaling with drive speed $λ$ is governed by $f(Ω)$. This law of $τ_*$ is found through a small-divisor mechanism in this work's iterative rotating frame scheme. Multi-photon suppression breakdown separates $λ$-scaling of $τ_*$ in linear response and non-perturbative theory, shown by a case study of Quasi-Floquet driving. The principle is applied to (i) Resolve inconsistencies in literature on non-periodic driving, and (ii) Provide design principles for engineering prethermal phases of matter in programmable quantum simulators, exemplified by new non-periodic `Factorial' drives with tunable $τ_*$.

Perovskite oxides display correlated electrical, magnetic, and thermal properties that can be further tuned in the thin-film limit, making them contenders for next-generation electronics. Measuring thermal transport in thin films is challenging, because traditional techniques are dominated by the substrate. Here, frequency-domain thermoreflectance (FDTR) of an epitaxial NdNiO$_3$ thin film reveals a sharp change in out-of-plane thermal conductivity across the metal-insulator transition. Complementary frequency-domain photoreflectance (FDPR) reveals a large change in ambipolar diffusivity of photoexcited carriers. While the in-plane electrical resistance shows large hysteresis, out-of-plane thermal and charge transport shows negligible hysteresis. We attribute this discrepancy to anisotropy in the percolation of nanoscale domains across the transition as the film thickness approaches the domain length scale. We establish FDTR and FDPR as sensitive probes of quantum material phase transitions and highlight NdNiO$_3$ for thermal control and memory applications.

The Large Deviation Principle (LDP) and the Central Limit Theorem (CLT) are concepts of information theory and probability. While their formulations are established under the i.i.d. assumption, the probabilistic foundation for power-law distributions has primarily evolved through descriptive models or variational principles, rather than a constructive derivation comparable to the classical binomial process. This paper establishes a constructive probabilistic framework for power-law distributions, proceeding from the nonlinear differential equation $dy/dx = y^q$ without assuming a specific distribution a priori. We build the algebraic and combinatorial foundations, which lead to a generalized binomial distribution based on finite counting. We prove the LDP for this generalized binomial distribution in the regime $0 < q < 1$, demonstrating that the $α$-divergence is identified as the rate function, and clarify the breakdown of this macroscopic scaling for heavier tails ($q > 1$). This result connects our constructive framework to the structures of information geometry. Furthermore, we prove a generalized de Moivre-Laplace theorem, showing that the generalized binomial distribution converges to a heavy-tailed limit distribution (the $q$-Gaussian distribution). We derive that the scaling law follows the order of $n^{q/2}$ as a consequence of the underlying nonlinearity. These analytical results are numerically verified for distinct values of $q \in (0, 2)$. This framework provides a constructive basis that unifies the shift-invariant exponential family and the rescaling-invariant power-law family.

The prevalence of sparsity in interacting many-body systems motivates an investigation into the spectral statistics of sparse random matrices with on-site disorder. We numerically demonstrate that the Anderson transition can be identified through the statistical properties of the ground state. By analytically deriving the energy moments and calculating the shifted kurtosis, we estimate the critical sparsity threshold for this localization-delocalization transition. The short-range energy correlation in the bulk indicates that the Anderson transition at infinite temperature coincides with the quantum phase transition. Furthermore, long-range energy correlations in the bulk spectrum reveal a Thouless energy scale, suggesting a broad nonergodic regime within the delocalized phase.

We perform a differential resistance study in the hydrodynamic regime of electron liquid in GaAs/AlGaAs quantum wells. At zero magnetic field ($B$) a Lorentzian profile occurs in the nonlinear transport driven by a U-turn (ac) current loop, in (ac + dc) measurements a minimum deepens with the external dc current bias ($j_{dc}$). Our analysis shows that the observed electronic transport valley induced by $j_{dc}$ is attributed to Joule heating effect on the electron temperature ($T_{e}$) of electron liquid. Quantitatively, we demonstrate that the viscosity resistivity ($Δρ$) is proportional to $T^{-2}$ and is consistent with the dc-current induced electronic Gurzhi effect in various configurations of measurement.

Ground states of materials with orientational order ranging from solid ferromagnets and ferroelectrics to liquid crystals often contain spatially varying vector-like order parameter caused by inner factors such as the shape of building units or by the geometry of confinement. This review presents examples of how the shapes, chirality, and polarity of molecules and spatial confinement induce deformed equilibrium and polydomain states with parity breaking, splay, bend, and twist-bend deformations of the order parameter in paraelectric and ferroelectric nematic liquid crystals. Parity breaking results either from chirality of the constituent molecules, as a replacement of energetically costly splay and bend in paraelectric nematics, or in response to depolarization field in the ferroelectric nematic. Both paraelectric and ferroelectric nematics exhibit a splay cancellation effect, in which the elastic and electrostatic energies of splay along one direction are reduced by an additional splay along orthogonal directions.

We turn the long time puzzle of the free volume, known for its highly irregular form, into exact analytical formulae and develop statistical mechanics of the hard disk model. The free volume is exactly expressed in terms of the intersection areas of up to five exclusion circles, which can be computed analytically as functions of disk coordinates. In turn, the free volume determines the partition function and entropy. The partition function is shown to factorize into a product of free volumes and admits two exact limiting forms corresponding to gaslike and liquidlike regimes. From this construction, using Monte Carlo-generated disk coordinates, the entropy and pressure are obtained analytically and recover the known equation of state of hard disks in almost entire density range up to the close packing. At intermediate densities, the theory reveals a mixed liquid regime associated with defect formation preceding the hexagonal ordering. The intersection area of five disks emerges as a scalar measure of the local hexagonal order. The theory can be directly adopted for the hard sphere model.

Chiral surface waves are surface-localized modes that propagate unidirectionally along a boundary, enabling directed transport and minimal back-scattering. While first identified in quantum systems, they were recently shown to emerge in classical metamaterials in the presence of `odd elasticity'. Owing to the non-reciprocality of odd elasticity, these waves exhibit growing amplitudes during propagation, reminiscent of the non-Hermitian skin effect. To date, studies of odd elastic systems have mainly focused on ordered structures. Whether structurally-disordered materials can host non-Hermitian chiral surface waves (NHCSW) remains unexplored. We address this question using a minimal model of torque-driven disordered odd solids. Such solids are abundant, from biological gels such as the cytoskeleton driven by motor-proteins to synthesized systems such as magnetic colloidal gels. We find that torque-driven disordered odd solids have unique NHCSW with stronger surface localization and stable boundary velocity, in contrast to previous lattice models of odd solids. These distinct features stem from an intrinsic interplay between boundary torques and odd elasticity in torque-driven odd solids. Our results offer a new strategy to control NHCSW using active torques.

We introduce a new approach to disordered two-dimensional Ising models based on the extension of the combinatorial solution to randomized supercells. Applying it to the site-diluted Ising model on the square lattice, we resolve the full phase boundary $T_c(p)$ from the pure-Ising point to the percolation limit $T_c(p_c)=0$ with, in principle, arbitrary precision. The critical eigenvalue governing the transition is found to follow a remarkably accurate linear interpolation between the Ising and percolation endpoints, whose small but systematic deviations reveal the nontrivial fine structure of the phase boundary. Near the percolation threshold, we confirm the crossover exponent $φ_{\rm RSIM}=1$ and extract the nonuniversal amplitude ${α_{\rm RSIM}\simeq 1.616}$.

Microorganisms ofter move in confined, disordered environments, where hydrodynamic couplings can modify their transport behavior. Using extensive finite-element simulations, we investigate the dynamics of microswimmers -- modeled as squirmers -- in two-dimensional disordered porous media by resolving the full hydrodynamic interactions. We reveal that the deterministic coupling between activity, hydrodynamics, and disorder is sufficient to generate effective diffusive transport. Strong pushers and pullers become localised in the porous medium either by trapping at corners or dynamic trapping, depending on swimmer type and obstacle packing fraction. Squirmers can escape from dynamic traps, leading to a prominent ``hopping-and--trapping'' dynamics. Strikingly, we find a pusher-puller asymmetry in the trapping probability that can be reversed by short-range swimmer-obstacle interactions, highlighting the sensitivity of transport to near-field effects.

Quantum geometry governs a wide range of transport and optical phenomena in quantum materials. Recent works have explored analogue electromagnetism and gravity in terms of the quantum geometric tensor, whose real and imaginary parts correspond to the quantum metric and the Berry curvature. By treating real- and momentum-space geometries on an equal footing, we develop a comprehensive and general formalism based on an expansion in $\hbar$, equivalent to an expansion in spatial derivatives. We derive the quantum-metric corrections to the wave-packet energy, the Berry connection, and the phase-space density of states, similar to the field-induced corrections in nonlinear response. A kinetic equation that captures quantum-metric effects across the full phase space then follows naturally. We further identify a polarization induced by gradients of the metric and a linear Hall response originating from its mixed components. Our framework provides a foundation for investigating thermodynamic and transport properties in systems where real- and momentum-space quantum geometries coexist.

Green parafermions, originally introduced by Green and extended by Greenberg and Messiah through trilinear and relative trilinear commutation relations beyond Bose-Fermi statistics, are generally regarded as mathematical curiosities without physical realization. We show that these paraparticles can in fact emerge as composite excitations in a broad class of condensed-matter systems undergoing spontaneous symmetry breaking with type-B Goldstone modes. The key ingredient is the introduction of auxiliary Majorana fermions defined on emergent unit cells produced by partial translational-symmetry breaking. When the auxiliary Majoranas are treated as physical degrees of freedom, the resulting Green parafermion states (up to a projection operator) correspond to flat-band excitations, whose creation and annihilation operators satisfy the trilinear algebra. When they are regarded as fictitious, the same construction explains the appearance of exponentially many degenerate ground states and reveals a surprising correspondence between Green parafermions and self-similar geometric objects, such as the golden spiral. Explicit realizations are demonstrated for the ferromagnetic spin-1 biquadratic model and the ferromagnetic $\rm {SU}(2)$ flat-band Tasaki model, showing that condensed-matter systems with type-B Goldstone modes provide a natural setting for Green parafermions as emergent, possibly observable quasiparticles.

The non-uniform (or inhomogeneous) electron gas has received much attention in many-body quantum mechanics and quantum chemistry in the early days of density functional theory, mainly as a theoretical device to construct gradient approximations via linear response theory. In this article, motivated by the recent works of Lewin, Lieb and Seiringer, we propose a definition of the quantum (resp. classical) non-uniform electron gas through the use of the grand-canonical Levy-Lieb functional (resp. the grand-canonical strictly correlated electrons functional), establish these systems as rigorous thermodynamic limits and analyze their basic properties. The non-uniformity of the gas comes from an arbitrary lattice-periodic background density.

Geometric incompatibility, the inability of a material's rest state to be realized in Euclidean space, underlies shape formation in natural and synthetic thin sheets. Classical Gauss and Mainardi-Codazzi-Peterson (MCP) incompatibilities explain many patterns in nature, but they do not exhaust the mechanisms that frustrate thin elastic sheets. We identify a new incompatibility that forbids any stretching-free configuration, even when the rest state of the elastic sheet locally satisfies the Gauss and MCP compatibility conditions. We demonstrate this principle in a model of surface growth with positive Gaussian curvature, where a geometric horizon forms, leading to the onset of frustration. Experiments, simulations, and theory show that the sheet responds by nucleating periodic d-cone-like dimples. We show that this obstruction to stretching-free configurations is topological, and we point to open questions concerning the origin of frustration.

Practical performance of quantum sensors is often curtailed by uncontrolled environmental drift (bias-field instability, temperature fluctuations, mechanical vibration), background fields, and imperfect control pulses. This motivates developing physical mechanisms that intrinsically compensate for such perturbations while retaining high sensitivity to target fields. We introduce an interaction-protected magnetometry scheme where periodic driving steers the collective magnetization onto two long-lived, prethermal Floquet "orbit" axes well-separated on the Bloch sphere. Rapid toggling between these axes encodes target fields as a differential signal, whereas background fields appear as common-mode motion that is strongly rejected, achieving >1000-fold suppression while canceling prethermal transients. This enables accurate reconstruction of rapidly varying audio-band magnetic signals without predictive filtering or spectral tuning. We provide an experimental proof-of-principle using a dense ensemble of coupled nuclear spins, operated here as a broadband (0-1 kHz) magnetometer. The protocol is remarkably tolerant to imperfections, operating robustly across millions of pulses under pulse-angle (~10°) and pulse frequency (>1 kHz) errors, large bias-field drifts (>50 $\mathrmμ$T), temperature variations over 150 K, and harsh mechanical vibrations. These results establish Floquet prethermalization as a resource for robust quantum sensors that combines broadband magnetic-field sensitivity with intrinsic immunity to diverse environmental and control perturbations, opening a path toward stable quantum metrology beyond controlled laboratory conditions.

Cooperating first then mimicking the partner's act has been proven to be effective in utilizing reciprocity in social dilemmas. However, the extent to which this, called Tit-for-Tat strategy, should be regarded as equivalent to unconditional cooperators remains controversial. Here, we introduce a biased Tit-for-Tat (T) strategy that cooperates differently toward unconditional cooperators (C) and fellow T players through independent bias parameters. The results show that, even under strong dilemmas in the donation game framework, this three-strategy system can exhibit diverse phase diagrams on the parameter plane. In particular, when T-bias is small and C-bias is large, a ``hidden T phase'' emerges, in which the weakest T strategy dominates. The dominance of the weakened T strategy originates from a counterintuitive mechanism characterizing non-transitive ecological systems: T suppresses its relative fitness to C, rapidly eliminates the cyclic dominance clusters, and subsequently expands slowly to take over the entire population. Analysis in well-mixed populations confirms that this phenomenon arises from structured populations. Our study thus reveals the subtle role of bias regulation in cooperative modes by emphasizing the ``survival of the weakest'' effect in a broader context.

We investigate the equilibrium magnetization in ferromagnetic nanoscrews (NSw) using micromagnetic simulations. These systems consist of elongated three-dimensional magnetic membranes with helicoidal geometry, combining curvature, torsion ($\mathrm{w}$), and eccentricity ($ε$) along their length. We focus on the influence of these geometric parameters, together with membrane thickness and inner diameter, on remanent states and coercive fields. Our results, obtained over a broad range of eccentricities and torsions, reveal bistable magnetic behavior, with vortex-domain-wall propagation during magnetization reversal. We identify four degenerate configurations of a remarkably stable mixed remanent state. The coercive field is found to increase with eccentricity for structures with a major axis (larger inner diameter) approximately 30\% larger than the minor axis (smaller inner diameter), while remaining largely insensitive to variations in torsion. These findings are interpreted in terms of geometry-induced modifications of surface magnetostatic charges on the membrane mantle. Overall, our results demonstrate that nanoscrews exhibit robust bistability under systematic geometric deformation, together with enhanced coercivity, highlighting their potential for applications in three-dimensional nanomagnetism.

We introduce a two-dimensional model of spin-1/2 Dirac fermions in graphene subjected to a highly tunable electric field, which exhibits super-Klein tunneling. The electric field can be continuously interpolated between two limiting configurations: a uniform electrostatic Lorentzian barrier with translational invariance and a chain of well-separated electrostatic scatterers. We demonstrate that super-Klein tunneling arises naturally as a direct consequence of the intrinsic connection of the model to free-particle dynamics, a relation that is established through methods of supersymmetric quantum mechanics, which provide an elegant and analytically tractable framework. Besides the mentioned super-Klein tunneling, scale invariance of the model and invisibility of the potential for particles of specific energy are revealed, and possible routes toward experimental realization are discussed.

Glassiness occurs when disorder and frustration cause local degrees of freedom to freeze despite the lack of long-range order. In systems of interacting bosons, such glassiness may involve a purely quantum degree of freedom$\unicode{x2014}$local phases of particle wave functions$\unicode{x2014}$partly analogous to spins in spin glasses. However, experimental identification of such phases is difficult because it requires prohibitively long measurement times or recourse to the elusive Edwards-Anderson order parameter. Moreover, the off-diagonal character of the phase makes it seemingly even harder to capture via typical observables. To address this issue, we study a system of strongly interacting bosons with random hoppings that features off-diagonal glassiness exhibiting replica symmetry breaking (RSB). We find that the glass phase is compressible, which distinguishes it from the Mott insulator. Thus, we establish a direct correspondence between phase-based glassy order and a measurable density-based thermodynamic observable. We use a framework adopted from spin glasses, including the replica trick within the one-step RSB scheme, to obtain meaningful results in the glass phase and to characterize the order parameters, RSB structure, slow relaxation, and compressibility. Glassiness in particle systems could thus be experimentally identified via measurements of compressibility, such as probing density fluctuations or the particle-number response to a trapping potential.

Under coarse observation, detectability of unresolved slow forcing can be projection-controlled: only the component of the hidden-induced deformation normal to a reduced null manifold remains locally visible. We establish this exactly in a solvable driven AR$(1)$-by-AR$(1)$ benchmark. The local Whittle/Kullback--Leibler distance from the true spectrum to the best nearby one-pole surrogate obeys $\Dloc(λ)=Cλ^4+O(λ^6)$, even though the observed spectrum itself is perturbed at $O(λ^2)$; detectability is therefore quartic, not quadratic, in coupling. The coefficient $C$ is obtained in closed form and vanishes as $(a-b)^2$ when the hidden and intrinsic timescales coalesce, identifying a spectrally \emph{dark} regime in which the leading perturbation is tangent to the reduced manifold. This yields a population boundary $\lcpop(N)\propto(\log N/N)^{1/4}$, with Whittle-BIC crossover near that scale. The benchmark exposes a broader geometric principle in reduced inference: tangent hidden effects are absorbed by reparametrization, whereas only surviving normal components control local distinguishability.

We propose an approach to induce nontrivial bands with non-zero Chern numbers by utilizing strong spin-orbit coupling in transition metal dichalcogenides with dopants. We demonstrate that a doped state near the valence-band edge induces band inversion with the hybridized host band, leading to topologically non-trivial properties. Calculations for V-doped WSe2 and WS2 confirm this mechanism. The coexistence of magnetic order and nontrivial topology in these systems offers a promising platform for exploring the quantum anomalous Hall effect.

Flat bands in moiré superlattices provide a fertile ground for correlated and topological phases, governed by their quantum geometric properties. While the valley-based paradigm captures key features in select materials, it breaks down in a growing class of systems lacking valley structure, where exotic phenomena such as twist-angle-tunable numbers of flat bands emerge. In this work, we develop and analyze tight-binding models for twisted heterobilayers of bipartite lattices, with a focus on the role of interlayer hybridization in generating flat-band quantum geometry. We demonstrate that sublattice-selective interlayer tunnelings in twisted dice lattice and graphene heterobilayers induce isolated flat bands at zero energy, whose number is tunable by the twist angle. Most importantly, these flat bands exhibit finite Berry curvature and a quantum metric of the Chern-insulator scale, generated through interlayer hybridization. This establishes a mechanism to induce quantum geometry in moiré flat bands beyond the valley paradigm. Our results chart a route to flat-band quantum geometry engineering in twisted bilayer bipartite lattices, with potential material realizations in oxide heterostructures, molecular lattices, and synthetic quantum matter.

The interplay between band topology, Berry curvature, and moiré flat bands lies at the heart of recent advances in quantum materials. In well-studied moiré systems such as twisted bilayer graphene and transition metal dichalcogenides, the quantum geometry of moiré flat bands typically reflects that of the monolayer, with Berry curvature originating from the band edge at the same valley. Whether this correspondence persists in systems with complex monolayer band structures and broken symmetries remains unclear. Here, we study twisted bilayers of loop-current-ordered kagome lattices (tb-LCK), which have been proposed in the context of vanadium-based kagome materials, using tight-binding models, and uncover a twist-induced reconfiguration of quantum geometry. By tuning the phase of the loop-current order, we identify the suppression of monolayer Berry curvature through twist-driven band reconstruction. We attribute these effects to strong interlayer hybridizations, enabled by the unusually large interlayer tunneling inherent to vanadium-based kagome materials, which mix energetically distant states and reshape quantum geometry. These results reveal that twist in tb-LCK suppresses quantum geometric inheritance from the monolayer, and establish loop-current-ordered moiré systems as promising platforms for exploring unconventional quantum geometry in moiré flat bands. We further comment on the experimental feasibility of the proposed system via vanadium-based kagome materials.

Frontal polymerization (FP) enables rapid curing of thermosets via a self-sustaining thermal wave, but its propagation mechanism can shift dramatically depending on processing conditions. In this study, we investigate the effect of trigger direction and monomer viscosity - controlled via hold time - on the front velocity in frontal ring-opening metathesis polymerization (FROMP) of dicyclopentadiene (DCPD). Our experiments reveal that at low viscosities, bottom-triggered FP fronts propagate significantly faster, ~50% faster front speed compared to top-triggered ones, driven by buoyancy-enhanced convection that preheats the unreacted monomer ahead of the front, that can have important implications for manufacturing applications. However, with increasing hold time, the monomer viscosity rises steeply, suppressing convection and causing the front velocity for top and bottom triggering to converge. This behavior reflects a convection-to-conduction (thermal-diffusion) transition in heat transport during FP. Complementary simulations incorporating buoyancy-driven advection reproduce the observed trends and highlight the importance of fluid flow in front dynamics. These results provide new insight into the coupled thermo-fluid-chemical mechanisms in FP offer strategies to tailor front behavior through viscosity and initiation geometry.

Onsager reciprocity $L_{ij}=L_{ji}$ is a cornerstone of near-equilibrium thermodynamics derived from microscopic time-reversal symmetry. We develop a geometric framework in which entropy-weighted reparameterization of thermodynamic response functions leads to an effective asymmetry in cross-couplings without violating the microscopic Onsager theorem. Motivated by the parallel structure of heat capacities $C_p$ and $C_v$, we introduce entropy-weighted response variables $λ_p$, $λ_v$, $λ_s$, and $λ_t$. Their ratios $Γ_c=λ_p/λ_v=C_v/C_p$ and $Γ_m=λ_s/λ_t=κ_T/κ_S$ form thermodynamic invariants whose product equals unity in equilibrium. Within a differential-form representation of thermodynamic state space, equilibrium corresponds to exactness of the accessibility form $ω=λ_p\,dp+λ_v\,dv$ with $dω=0$, while non-equilibrium processes generate curvature $Ω=dω$, producing an effective asymmetry in the transformed coupling matrix. A microscopic theorem shows that entropy-weighted statistical ensembles with time-reversal asymmetry $χ(Γ)=W(ΘΓ)/W(Γ)\neq1$ generate an antisymmetric contribution to the Green--Kubo transport matrix. Atomic-scale analysis using the Transforma model reveals cross-derivative asymmetries across the $3d$ transition series, peaking at configuration anomalies in Cr and Cu. Temperature-dependent Raman spectroscopy of monolayer graphene exhibits statistically significant hysteresis loops (up to $30σ$), providing experimental evidence for thermodynamic curvature. These results unify microscopic irreversibility, atomic structure anomalies, and macroscopic hysteresis within a geometric interpretation of entropy-weighted thermodynamic coupling.

Active particles locally transduce energy into motion, leading to unusual and emergent behaviors. However, current synthetic particles lack sensing and adaptation mechanisms. Here, we demonstrate a novel regulation pathway, through the combined use of thermophoretic propulsion and nanometric building blocks. We build an active fluid composed of artificial nanomotors and study its three-dimensional (3D) dynamics. We use laser-induced photo-thermal effect to actuate nanoparticles, and probe their self-propulsion within assemblies. Despite significant thermal fluctuations at the nanoscale, our results reveal a strong dependence of the thermophoretic propulsion on the concentration of nanomotors, leading to ultrafast velocities of up to ~ 800 um/s. This unique behavior originates from a strong coupling of the local concentration of nanomotors and the temperature field, which feeds back on the thermophoretic mobility of the nanoparticles. We rationalize our results from independent modeling of all thermal effects, accounting for nonlinearities of thermophoretic self-propulsion. Our results open novel routes for the design and self-regulation of 3D active fluids by thermal processes.

We numerically study confined channel foam flow around an obstacle using a two-dimensional bubble model, inspired by experiments performed in the same geometry. We systematically vary the polydispersity, the external driving force, and the packing fraction of the system. Our simulations capture a broad range of plastic flow phenomenologies, from highly directional, sliding-like motion characteristic of crystalline materials to more isotropic and localized rearrangements typical of amorphous systems. We identify a threshold value of polydispersity that marks the crossover between crystalline-like and amorphous-like plasticity. In addition, we observe the existence of a critical external force, associated with the phenomenon of yield drag, above which the system reaches steady flow and below which it remains arrested. We determine a critical packing fraction above which such yield-drag behavior emerges. Our results provide a comprehensive framework for understanding the interplay between disorder, driving, and the presence of an obstacle in foam flows.

We use in situ transmission electron microscopy with automated tracking to study supported gold nanoparticles (NPs) during high-temperature vacuum annealing. \rev{The average mass loss per NP is governed by a flat, nearly size-independent substrate-mediated evaporation profile.} On top of \rev{this mean shrinkage}, individual NPs show significant fluctuations in apparent growth or shrinkage, and NP volume follows a \rev{random-walk-like trajectory. To rationalize both the ensemble-mean behavior and the particle-resolved variability, we develop a self-consistent theory that couples substrate-mediated evaporation to collective 2D Ostwald-type mass exchange through a shared adatom field, described in terms of a renormalized screening length and background concentration. In the experimentally relevant regime, the theory predicts an approximately size-independent mean shrinkage rate and clarifies how net mass loss suppresses classical coarsening.} \rev{Superimposed on this deterministic drift, we quantify stochastic volume trajectories and capture their fluctuation spectrum with a minimal Langevin description consistent with intermittent adatom attachment and detachment events.} In addition, we characterize the lateral diffusive motion of NPs, which is responsible for their coalescence. Altogether, our results highlight that stochasticity is intrinsic at the nanoscale \rev{and that predicting the evolution of supported NPs at early and intermediate times requires a unified framework combining substrate-mediated evaporation, collective mass exchange, and stochastic fluctuations.

Pedestrian congestion at corridor intersections often originates from localized fluctuations in motion rather than from a macroscopic collapse of flow. Understanding pedestrian instability at corridor intersections remains challenging because existing studies mainly rely on density, average speed, or flow-based measures and limited datasets, making it difficult to separate geometric turning effects from interaction induced fluctuations in merging flows. In particular, the mechanism underlying the turning angle dependence in T junctions has not been resolved. Here, we analyze more than 300 controlled experiments conducted in L corridors with turning only and T corridors with turning and merging. Using Voronoi-based speed variance $V_s$ and velocity variance $V_v$, we systematically compare geometric and interaction effects. $V_s$ effectively captures interaction driven instability, while $V_v$ reflects directional adjustments due to geometry. The comparison reveals distinct fluctuation mechanisms and identifies a critical transition near $90°$, demonstrating the advantage of variance-based indicators for diagnosing pedestrian dynamics.

In classical systems, chaos is clearly defined via the behavior of trajectories. In quantum systems with a classical analogue one finds that the transition from regular to chaotic dynamics is signified by a change in the spectral statistics. This has been found to remain true for quantum systems with no classical analogue, including many-body systems. Furthermore, quantum chaotic systems explore all the allowed configurations in the Hilbert space, i.e. they are ergodic, while integrable systems, and systems in the many-body localized phase, are restricted to a certain subspace of the available phase space, and hence strongly break ergodicity. In this dissertation, we study the intermediate behavior between ergodicity and localization, i.e. the weak breaking of ergodicity. The model examined is the PXP spin chain model, where spins are allowed to flip only under certain kinetic constraints. We start by reproducing some already established results. First, we explore the eigenstate thermalization hypothesis (ETH) for this model and demonstrate the existence of a small number of states, throughout the PXP spectrum, that violate the ETH. Then we study the level-spacing statistics of the model, a well-known quantum chaos diagnostic, which turns out to be close to semi-Poisson and approach Wigner--Dyson statistics for large system sizes. Moreover, we examine various aspects of the model that have not been studied before. For example, the eigenvector component statistics, another quantum chaos diagnostic, for the PXP model turn out to be non-Gaussian. Finally, we perform a quench, in order to study how the energy spreads throughout the system, and observe ballistic fronts.

In this work, we formulate a theoretical model based on a cellular automaton (CA) to study thermal transport in low-dimensional nanostructures across ballistic, diffusive, and transition regimes. Unlike computationally intensive methods such as the Boltzmann Transport Equation (BTE), our model stands out for its geometrical robustness, allowing the seamless integration of substitutional impurities, vacancies, and irregular edges. We validated the model using graphene nanoribbons (AGNRs), successfully replicating the dependence of thermal conductivity on ribbon width and temperature. Results demonstrate that the model captures critical scattering and confinement effects with a linear scalability O(N). Given the increasing pressure to optimize computational resources and reduce the carbon footprint associated with AI infrastructure, this CA model emerges as a highly efficient tool for the parametric exploration and design of next-generation thermal devices.

Understanding and predicting the behavior of liquid matter across length scales, using only the microscopic interactions encoded in the Schrödinger equation, remains a central challenge in the physical sciences. Achieving this goal requires not only an accurate and efficient description of intermolecular forces but also a consistent framework that bridges the micro-, meso-, and macroscales. Here, by combining machine-learned interatomic potentials (MLIPs) with neural classical density functional theory (neural cDFT), we present such a framework. The underlying idea is simple: MLIPs trained on quantum-mechanical energies and forces are used to generate inhomogeneous microscopic density profiles, which in turn serve as the training data for neural cDFT. The resulting ab initio neural cDFT is not only significantly more computationally efficient than molecular simulations, but also provides a conceptually transparent route to the thermodynamics of both homogeneous and inhomogeneous systems. We demonstrate the approach for both water and carbon dioxide using several exchange-correlation functionals. Beyond accurately reproducing bulk equations of state and liquid-vapor phase diagrams, ab initio neural cDFT predicts, from first principles, how confinement modifies liquid-vapor coexistence in water. It also captures complex behavior in supercritical carbon dioxide such as the Fisher-Widom and Widom lines. Ab initio neural cDFT establishes a general first-principles route to multiscale modeling of fluids within a single unified conceptual framework.

We propose encapsulating type-A antiferromagnetic semiconductors between graphene layers to realize a gate-tunable synthetic antiferromagnet with nonrelativistic spin splitting, enabling efficient spintronic transport via graphene. Ab initio calculations and tight-binding models of graphene/MnS/graphene heterostructure reveal that gate-tuning of the heterostructure breaks top/bottom graphene equivalence, inducing opposite ferromagnetic proximity exchange that lifts spin degeneracy to yield nonrelativistic spin splitting at the Fermi level, dominating over relativistic effects. The induced effects manifest as conductance dips in spin-resolved transport through proximitized graphene nanoribbons, observable as giant magnetoresistance within a narrow energy window around the Fermi level. Our graphene/type-A antiferromagnetic heterostructure, a readily synthesizable platform incorporating antiferromagnets with nonrelativistic spin splitting, pave the way for gate-manipulated, low-dimensional antiferromagnetic devices.

A unified semiclassical framework is presented to describe the evaporative cooling of trapped atomic gases, accounting for both classical and quantum statistics. By combining global thermodynamics with phase-space distributions, general analytic expressions for the particle number and internal energy are derived for a broad family of confining potentials. Building on these results, a recursive evaporation protocol is formulated based on truncated energy distributions, enabling stepwise mapping between successive thermodynamic states and revealing the system's degree of freedom governance over cooling efficiency. Numerical simulations of the systems highlight the contrasting behavior of classical and quantum systems as they approach degeneracy, with particularly distinctive signatures in quadrupole traps, due to their nonstandard phase-space scaling. The results provide a versatile theoretical tool for modeling evaporative cooling across experimentally relevant geometries and offer quantitative guidance for optimizing cooling trajectories in ultracold atomic systems.

The forward derivation of stochastic differential equations in individual physical domains has proceeded independently for over a century without generalising across disciplines. A generalisation of Takens' embedding theorem to stochastic systems, the Stochastic Embedding Sufficiency Theorem, closes this gap as an inverse methodology enabling non-parametric recovery of drift and diffusion fields from scalar time series without prior assumptions about the governing physics. A blind recovery protocol, receiving only raw time series and sampling interval, is applied to nine domains: classical mechanics, statistical mechanics, nuclear physics, quantum mechanics, chemical kinetics, electromagnetism, relativistic quantum mechanics, quantum harmonic oscillator dynamics, and quantum electrodynamics. The pipeline recovers the governing equations of each domain with errors from 0.026% to ~1%, with no null hypothesis rejected at the 5% level. Physical constants emerge in both channels without prior specification. The recovered diffusion coefficients constitute an empirical pattern, the σ-continuum, in which the Planck constant, Boltzmann constant, and speed of light play structurally distinct roles. Three independent uniqueness arguments determine the gravitational diffusion coefficient as one Planck length per square root of Planck time, non-parametrically derived from first principles. Four canonical axioms formalise the framework. Physical time emerges as a monotone functional of the stochastic evolution. Within these axioms and the short-memory limit, the drift, covariance operator, and fluctuation amplitude are all fixed. The resulting superspace diffusion hypothesis generates non-parametric, first-principles, falsifiable predictions against galactic kinematic data as developed in a companion paper (Part II).

We study how matrix-product-operator (MPO) truncation errors evolve when simulating two setups: (1) 1D Haar-random circuits under either depolarizing noise or amplitude-damping noise, and (2) 1D Lindbladian dynamics of a non-integrable quantum Ising model under either depolarizing or amplitude-damping noise. We first show that the average purity of the system density matrix relaxes to a steady value on a timescale that scales inversely with the noise rate. We then show that truncation errors contract exponentially in both system size $N$ and the evolution time $t$, as the noisy dynamics maps different density matrices toward the same steady state. This yields an empirical bound on the $L_1$ truncation error that is exponentially tighter in $N$ than the existing bound. Together, these results provide empirical evidence that MPO simulation algorithms may efficiently sample from the output of 1D noisy random circuits [setup (1)] at arbitrary circuit depth, and from the steady state of 1D Lindbladian dynamics [setup (2)].

Non-Hermitian disordered systems have emerged as a central arena in modern physics, with ramifications spanning condensed matter, quantum, statistical, and high energy contexts. The same principles also underlie phenomena beyond physics, such as network science, complex systems, and biophysics, where dissipation, nonreciprocity, and stochasticity are ubiquitous. Here, we review the physics and mathematics of non-Hermitian disordered systems, with particular emphasis on non-Hermitian random matrix theory. We begin by presenting the 38-fold symmetry classification of non-Hermitian systems, contrasting it with the 10-fold way for Hermitian systems. After introducing the classic Ginibre ensembles of non-Hermitian random matrices, we survey various diagnostics for complex-spectral statistics and distinct universality classes realized by symmetry. As a key application to physics, we discuss how non-Hermitian random matrix theory characterizes chaos and integrability in open quantum systems. We then turn to the criticality due to the interplay of disorder and non-Hermiticity, including Anderson transitions in the Hatano-Nelson model and its higher-dimensional extensions. We also discuss the effective field theory description of non-Hermitian disordered systems in terms of nonlinear sigma models.

We investigate the interplay between the non-Hermitian skin effect and Aubry-André-Harper (AAH) quasiperiodic disorder in a one-dimensional Su-Schrieffer-Heeger (SSH) chain with nonreciprocal hopping. By exact diagonalization, transfer-matrix analysis, and an analytical similarity-transformation argument, we map the full ( , $δ$) phase diagram, where A is the AAH modulation strength and the nonreciprocity parameter. We identify five distinct regimes: ( ) topological with extended bulk, (II) AAH-localized, (III) skin-localized, (IV) fully localized, and a previously unreported (V) competition regime exhibiting reentrant partial delocalization, in which intermediate quasiperiodic disorder disrupts the directional skin accumulation before ultimately Anderson-localizing all states. Using phase-averaged diagnostics and finite-size scaling, we confirm that the reentrant regime is robust, characterized by a non-monotonic inverse participation ratio that sharpens with increasing system size. We derive an analytical expression for the modified localization boundary $λ_{c}(δ)=2\sqrt{v_{eff}w}$ with $v_{vff}=\sqrt{v^{2}-δ^{2}}$, which agrees with numerical Lyapunov exponent calculations. We further show that quasiperiodic disorder progressively unwinds the complex spectral loops, destroying the point-gap topology at a critical strength distinct from the band-topological transition ; that the skin effect suppresses entanglement entropy to near-zero values while sufficiently strong AAH disorder partially restores it ; and that the SSH sublattice structure absent in the widely studied non-Hermitian AAH chain is essential for producing the five-phase landscape, as demonstrated by direct comparison with the non-dimerized limit.

Linear Response theory aims to predict how added forcing alters the statistical properties of an unforced system. These kinds of questions have been studied predominantly for autonomous dynamical systems, yet many systems in the physical, natural, and social sciences are inherently nonautonomous, evolving in time under external forcings of various kinds (a canonical example being the climate system). In such settings, one would like to understand how the system's time dependent statistical properties change when additional infinitesimal forcings are applied. This question is of clear practical relevance, but from a rigorous mathematical viewpoint it has been addressed only for a few specific classes of systems/perturbations. Here we provide a rigorous linear response theory for a rather general class of deterministic and random nonautonomous systems satisfying a specific set of assumptions that in some sense extend the standard assumptions used in the autonomous setting. A central ingredient is rapid loss of memory, i.e. sufficiently fast forgetting of initial conditions along the nonautonomous evolution. Our main strategy is to reformulate the sequential dynamics as a fixed-point problem for a global transfer operator acting on an extended sequence space of measures. This yields explicit and readily implementable response formulas for predicting the effect of small perturbations on time-dependent statistical states. We illustrate the theory on two representative classes: sequential compositions of C3 expanding maps and sequential compositions of noisy random maps, where uniform positivity of the noise induces exponential loss of memory.

Scientists have long been fascinated by magnetoreception, the innate capacity of many animals to sense and use the Earth's magnetic field for navigation. In eusocial insects like honey bees, magnetoreception has been linked to communication and foraging. However, little is known about magnetoreception's phylogenetic patterns and relationship to species traits and natural history. Here, we demonstrate that putative magnetoreception based on ferromagnetic particles is widespread across a diversity of bee species (72 out of 96 species tested), with no phylogenetic signal. We also detected such putative magnetoreception in non-bee outgroups, suggesting this magnetic capacity predates the evolution of the Anthophila. While magnetic signals were found across a diversity of life history traits, the strength of the magnetic signal varied within and between species, and increased with body size and social behavior.

Statistical mechanics reveals that the properties of a macroscopic physical system emerge as an average over an ensemble of statistically independent microscopic subsystems, each occupying a specific microstate. In the study of quantum systems, these microstates can be chosen to correspond to the pure state wavefunctions of individual quantum systems. However, the physical principles that govern the distribution of a pure state wavefunction ensemble, even under conditions of thermal equilibrium, are not well established. For instance, the canonical Boltzmann distribution cannot be applied to wavefunctions because they lack a definite energy. In this manuscript, we present a maximum entropy principle for the quantum wavefunction ensemble at thermal equilibrium, the so-called Scrooge ensemble. We highlight that a constraint on the energy expectation value, or even the shape of the associated eigenstate distribution, fails to yield a valid equilibrium state. We find that in addition to these constraints, one must also constrain the measurement entropy to be equal to the Rényi divergence of the ensemble with respect to the Gibbs state, indicating that the Rényi divergence may have uninvestigated physical importance to thermal equilibrium in quantum systems.

Solitons are localized nonlinear wave packets that propagate without spreading because nonlinearity balances dispersion. Their robustness is well understood in effectively one-dimensional systems, but introducing additional spatial dimensions is generally expected to destabilize them or destroy their coherent character. Here we experimentally investigate how deep-water gravity-wave solitons behave when a controlled transverse degree of freedom is introduced through diffraction. Using a large-scale water-wave facility, we generate solitonic wave packets whose transverse structure is imposed across a segmented wavemaker through either a sharp slit or a smooth Gaussian apodization. The resulting two-dimensional wave fields are measured with high spatial resolution. Diffraction reshapes the transverse profile of the wave packet while its longitudinal dynamics retain the characteristic features of a soliton. Nonlinear spectral analysis confirms that the solitonic content is preserved along the direction of propagation, whereas the transverse evolution follows the linear Fresnel laws of diffraction. These observations reveal an unexpected coexistence of nonlinear soliton dynamics and classical wave diffraction.

We construct the dynamical symmetry of the quantum Calogero model with particle exchange in a confining Coulomb field. This symmetry is governed by the algebra $so(N+1,2)$, deformed by exchange (Dunkl) operators, with its invariant sector generated by the Dunkl angular momentum tensor and the modified Laplace-Runge-Lenz vector. The equidistant analogue of the Hamiltonian, with a linear spectrum, is expressed in terms of the conformal subalgebra $so(1,2)$. In addition, the wave functions of the Calogero-Coulomb Hamiltonian are classified into infinite-dimensional lowest-weight $so(1,2)$ multiplets.

We introduce a framework for the data-driven discovery of stochastic differential equations (SDEs) that unifies, for the first time, the weak-form integration-by-parts approach of Weak SINDy with the stochastic system identification goal of stochastic SINDy. The central novelty is the adoption of spatial Gaussian test functions $K_j(x)=\exp(-|x-x_j|^2/2h^2)$ in place of temporal test functions. Because the kernel weight $K_j(X_{t_n})$ is $\mathcal{F}_{t_n}$-measurable and the Brownian innovation $ξ_n$ is independent of $\mathcal{F}_{t_n}$, every noise term in the projected response has zero conditional mean given the current state -- a property that guarantees unbiasedness in expectation and prevents the structural regression bias that afflicts temporal test functions in the stochastic setting. This design choice converts the SDE identification problem into two sparse linear systems -- one for the drift $b(x)$ and one for the diffusion tensor $a(x)$ -- that share a single design matrix and are solved jointly via $\ell_1$-regularised regression with grouped cross-validation. A two-step bias-correction procedure handles state-dependent diffusion. Validated on the Ornstein--Uhlenbeck process, the double-well Langevin system, and a multiplicative diffusion process, the method recovers all active polynomial generators with coefficient errors below 4\%, stationary-density total-variation distances below 0.01, and autocorrelation functions that faithfully reproduce true relaxation timescales across all three benchmarks.

We study transport and escape in the Stochastic Web Map (SWM), an area-preserving system with phase-space structure controlled by a symmetry parameter $q$ and nonlinearity $K$. By analyzing the survival probability $P_{\text{S}}(n)$ and escape frequency $P_{\text{E}}(\ln n)$, we show that in the chaotic regime escape dynamics is governed by a single time scale $n_{\text{typ}}\propto K^{-2}h^{2}$; here $h$ is the size of the escape horizon. Deviations at large $K$ and small $h$ indicate a breakdown of the quasilinear approximation. Then, upon rescaling the time by $n_{\text{typ}}$, escape statistics becomes universal, independent of $q$. These results demonstrate that escape is controlled by global transport rather than symmetry.

The Out-of-Time-Order Correlator (OTOC) is a standard algebraic diagnostic of quantum information scrambling, but it offers limited direct geometric intuition. In this note, we propose a Bohmian, trajectory-based framework for constructing a geometric diagnostic of scrambling-related sensitivity using Lagrangian Descriptors (LDs). To avoid the uncertainty-principle obstruction to assigning independent initial position and momentum within a single wave function, we evaluate Bohmian dynamics over a two-dimensional preparation space of localized Gaussian wavepackets labeled by their initial center and momentum kick. For the inverted harmonic oscillator, this construction is analytically tractable: the wavepacket-center dynamics and their dependence on preparation parameters can be written explicitly. In particular, away from the equilibrium origin, the exponential growth of the associated preparation-space stability matrix yields an $\mathcal{O}(e^{ωT})$ bound on the sensitivity of the wavepacket-center LDs, motivating a semiclassical comparison with sensitivity structures associated with OTOC growth. In this sense, the LD provides a geometric indicator of scrambling-related sensitivity. We conclude by discussing how this preparation-space picture suggests a program for future work regarding the distinct microcanonical regimes previously reported for the inverted harmonic oscillator.

We study the integrability of a two-dimensional Hamiltonian system with a gyroscopic term and a non-homogeneous potential composed of two homogeneous components of different degrees. The model describes the motion of a particle in a plane under the combined influence of a central (Kepler-type) potential, a uniform magnetic field, and a superposition of homogeneous forces. By combining the Levi--Civita regularization with the so-called coupling constant metamorphosis transformation, and employing differential Galois theory, we derive analytical necessary conditions for integrability in the Liouville sense. They put restrictions on the degrees of homogeneity of the potential terms and their values in particular points. The obtained results encompass and generalize several classical galactic and astrophysical models, including the generalized Hill model, the Hénon--Heiles and Armbruster--Guckenheimer--Kim systems, providing a unified framework for studying non-homogeneous Hamiltonians. We demonstrate the effectiveness of the derived integrability obstructions by proving the non-integrability of these models in the presence of a uniform rotational field. The numerical analysis via the Poincaré cross-sections further confirms the analytical results, illustrating the transition from regular to chaotic dynamics as the rotational and non-homogeneous terms are introduced. Moreover, we show that, without the Kepler-type term, a generalized non-homogeneous extension of the exceptional potential remains integrable. The explicit forms of the first integrals are given.

Addressing both natural and societal challenges requires collective cooperation. Studies on collective-risk social dilemmas have shown that individual decisions are influenced by the perceived risk of collective failure. However, existing feedback evolving game models often focus on a single feedback mechanism, such as the coupling between cooperation and risk or between cooperation and cost. In many real-world scenarios, however, the level of cooperation, the cost of cooperating, and the collective risk are dynamically interlinked. Here, we present an evolutionary game model that considers the interplay of these three variables. Our analysis shows that the worst-case scenario, characterized by full defection, maximum risk, and the highest cost of cooperation, remains a stable evolutionary attractor. Nevertheless, cooperation can emerge and persist because the system also supports stable equilibria with non-zero cooperation. The system exhibits multistability, meaning that different initial conditions lead to either sustained cooperation or a tragedy of the commons. These findings highlight that initial levels of cooperation, cost, and risk collectively determine whether a population can avert a tragic outcome.

Turbulence remains one of the central open problems in classical physics, largely due to the absence of a closed dynamical description of the Reynolds stress. Existing approaches typically rely either on local constitutive assumptions or on high-dimensional statistical representations, without identifying a minimal set of dynamical variables governing the cascade response. Here we show that the non-local stress response implied by the Navier-Stokes equations admits a systematic reduction onto a low-dimensional anisotropic sector of the turbulent cascade. This reduction leads to a minimal dynamical system with the structure of a damped oscillator, arising from the coupling between the leading angular mode and its nonlinear transfer to higher-order sectors. Within this framework, classical turbulent behaviors --including inertial-range scaling, shear-driven transport, and wall-bounded logarithmic profiles-- emerge as different realizations of the same underlying dynamical structure. Universal quantities such as the Kolmogorov constant and the von Kármán constant appear as leading-order consequences of internal consistency conditions applied across homogeneous and shear-driven regimes. These results suggest that turbulence admits a minimal dynamical backbone governed by non-local cascade response, providing a unified perspective that connects spectral transfer, anisotropy, and mean-flow interaction within a single reduced framework.

The control and steering of light at nanometre length scales is crucial for the development of both fundamental science and nanophotonic technologies. Recent advancements have been achieved by exploiting various crystalline anisotropies, allowing for subdiffractional and diffraction-less canalisation of energy. These studies in particular benefit from stacking and twisting of 2D materials, whereas corresponding capabilities of anisotropic bulk crystals are rather unexplored. In this work, we show that ferroelastic twin walls - crystallographically perfect 2D-sheets that separate regions of differently oriented domains - in the distorted perovskite LaAlO3 provide a natural platform for broadband lateral confinement and superb canalisation of light at the nanoscale. Without fabrication processes, the electromagnetic fields localised at such walls exhibit lateral optical sizes up to 260 times smaller than the free-space wavelength. Depending on the adjacent domain orientation and frequency, the twin wall pattern preferentially concentrates or repels the electromagnetic energy, constituting a natural building block towards broadband MIR and THz nanophotonics for polaritonic circuitry.

Femtosecond laser micromachining (FLM) fabricated waveguides inherently form elliptical cores due to differences in focal spot size and the Rayleigh range of the microscope objective. Consequently, it is essential to study their propagation characteristics, which differ from those of conventional circular-core waveguides. In this work, we present the results of a parametric optimization of these waveguides to identify fabrication parameters that lead to minimal loss. A propagation loss characterization study revealed that, for a laser wavelength of 1030 nm, a pulse width of $\sim$300 fs, a pulse energy of 600 nJ, a scan speed of 2 mm/s, and a repetition rate of 100 kHz, a transparent and micro-bubble-free waveguide with a propagation loss of $\sim$0.4 dB/cm was formed. The modal analysis further demonstrated that the V-number depends on the core aspect ratio. The waveguide modes were compared with computationally generated modes, revealing a correlation that aligns well with existing literature.

Quantum-correlated networks distribute quantum resources such as squeezed and entangled states. These states are central to modern quantum technology, including photonic quantum computing, quantum communications, non-destructive biological sensing and gravitational-wave detection. Even for squeezed states of light - the most robust quantum-correlated resource - loss-induced decoherence remains the dominant obstacle to strong quantum advantage in in large-scale interferometric and networked quantum systems. Common design assumption in these applications is treating mismatches between spatial modes as a small, incoherent loss. Here we show that this picture can fail: coherent spatial-mode mixing with higher-order spatial modes can produce an apparent loss exceeding 100% relative to the initial squeezing, a regime we term hyperloss. We experimentally demonstrate hyperloss in a minimal two-node quantum network: with only 8% mode mismatch, a 5.8dB squeezed state is converted into an effectively thermal state with no quadrature squeezing, eliminating the quantum advantage. Because the effect is coherent, it is controllable: lost correlations can be recovered by tuning differential spatial-mode phases (e.g., Gouy-/propagation-phase). We demonstrate this recovery experimentally, not only eliminating the hyperloss, but even significantly suppressing the mode mismatch loss, with 15% geometric mismatch acting like only ~2.8% effective loss. Hyperloss is a design-limiting mechanism for all quantum networks with squeezed light, from from photonic quantum processors to large-scale interferometers and distributed quantum-sensing networks. Our results provide a practical route to avoid hyperloss and turn mode mismatch into an explicit, phase-aware design parameter for future quantum technologies.

High power end pumped solid state lasers often operate in regimes where pump induced heating creates a strong refractive index gradient (thermal lensing) that governs resonator stability and mode quality. When the pump is absorbed according to the Beer Lambert law, the thermal load, and hence the GRIN strength, vary along the crystal length, so the standard ABCD matrix of a constant-gradient GRIN element is no longer directly applicable. Here, we derive a closed-form ABCD transmission matrix for a thermally loaded laser crystal pumped by atop-hat beam while explicitly accounting for axial absorption. Starting from the steady-state heat equation, we obtain the temperature field and the associated thermo-optic index profile. We then solve the paraxial eikonal ray equation analytically and express the transfer-matrix elements in terms of Bessel and Neumann functions. The resulting matrix is validated against the conventional slab product method and shown to recover the uniform-medium and constant gradient GRIN limits. Finally, we illustrate its utility by model ing Bessel Gaussian beam propagation through the axially varying thermally induced GRIN medium.

We investigate finite-width optical-beam propagation in a medium with uniform torsion described by the geometric theory of a continuous distribution of screw dislocations. Starting from the Riemann--Cartan framework that yields torsion-induced circular birefringence for local plane waves, we construct a minimal paraxial beam model in which the same contortion-driven helicity splitting remains explicit. We show that uniform torsion breaks the degeneracy between the two circular-polarization sectors and induces a geometric rotation of the polarization that scales with both the propagation distance and the radial position in the beam. As a consequence, a finite-width beam develops spatially varying polarization textures across its transverse profile, naturally described by the Stokes parameters. We introduce beam-level observables based on the integrated Stokes vector, the transverse inhomogeneity of the polarization texture, and the number of resolved radial polarization domains, thereby connecting the torsion parameter to experimentally accessible beam diagnostics. The paper combines two complementary levels of description: an analytic short-distance regime, used to isolate the geometric mechanism, and full paraxial propagation including diffraction, used to test the robustness of the predicted textures. Within the cylindrically symmetric minimal model, the most robust structured-light signature of uniform torsion is beam-resolved polarization structuring, whereas strong orbital-angular-momentum conversion is not expected without additional azimuthal structure. We also identify the geometric ingredient required for genuine torsion-assisted spin--orbit conversion beyond the minimal radial model: an effective azimuthal geometric connection.

We investigate the mechanisms of formation of stable (2+1)-dimensional optical soliton molecules (SMs) and breather molecules (BMs) in a Rydberg atomic gas, highlighting the distinct roles of nonlocality. The underlying giant, nonlocal nonlinearity induced via Rydberg electromagnetically induced transparency (EIT), supports diverse, large-size lattice SMs (rhombic, square, checkerboard, hexagonal lattice SMs). Crucially, we identify two distinct formation regimes: In the nonlocal regime, long-range interactions alone stabilize the SMs without requiring initial motion. In contrast, within the strongly nonlocal regime, an initial velocity is essential to generate a centrifugal force that counteracts the strong attraction, resulting in rotating SMs. Furthermore, specific initial velocities can induce a periodic breathing instability, leading to the formation of BMs. Our study offers a new scheme for engineering SMs with diverse configurations and opens new avenues for data processing and transmission in optical systems.

Matrix elements of spherical tensor operators are fundamental to the analysis of lanthanide spectra in both amorphous and crystalline host materials. In the intermediate coupling scheme, the eigenvectors of the Hamiltonian define the electronic structure, while the eigenvalues determine the energy levels of the $f^N$ configuration. By utilizing these eigenvectors to evaluate electric and magnetic dipole operators, one can identify the radiative line strengths for all transitions in both absorption and emission. This work presents a comprehensive framework for the direct calculation of angular matrix elements using a Slater determinant basis and their subsequent transformation to the traditional $LS$-coupling scheme. Unlike conventional indirect methods, this approach is more universally applicable, though it is computationally more intensive. A concise set of general rules is prepared to enable the calculation of angular matrix elements for virtually any spherical tensor operator within an $f^N$ configuration. The computational overhead of this direct approach is well within the capabilities of modern desktop computing. Furthermore, since these configuration-specific angular matrices are mathematical constants independent of the host environment, they need only be calculated once. The Python package AMELI is introduced, which employs exact arithmetic to generate the matrix elements with absolute mathematical precision. Both the underlying algorithms and the calculated matrices for all lanthanide ions are provided in open-access repositories. This removes a significant barrier for experimentalists, providing the necessary operator matrices without requiring them to navigate the intricate theory and algorithmic implementation.

This paper examines how the Joint Institute for Nuclear Research (JINR), an international organization formally committed to peaceful science, is deeply embedded in an ecosystem of military-industrial enterprises in the city of Dubna in Russia, contributing to training specialists and developing technologies used in Russia's military operations, including attacks on civilian facilities in Ukraine. It also shows how JINR collaborates with scientific institutions on the Ukrainian territories occupied by Russia, legitimizing the occupation and exposing international partners to legal and ethical risks. Despite these ties, JINR maintains broad international collaborations, allowing its scientists and engineers to access advanced technologies and indirectly support Russia's military capabilities, highlighting the need for greater awareness in the global scientific community and coordinated sanctions enforcement.

A number of recent contributions have put forward the topological structure of production networks as a key determinant of macro-economic dynamics. However, firm-to-firm production networks data is generally not available. Against this background, reconstruction method based on firms' size have been developed. This paper enriches this set of reconstruction methods by integrating input-output sectoral flows in the reconstruction process. We derive analytical expressions for the maximum entropy solutions to the firm network reconstruction problem with sectoral input-output constraints, first for binary networks and then for weight reconstruction. We perform a numerical analysis comparing standard and input-output based reconstruction methods using Hungarian production network data. Our results show that adding input-output constraints substantially reduces deviations from the input-output structure compared with standard methods. Our augmented method provides an almost perfect fit to input-output data, though all methods have difficulties reproducing other structural characteristics.

We introduce $π$-Girsanov, a new method for constructing Markov state models from biased enhanced-sampling molecular dynamics simulations based on Girsanov reweighting. The key idea behind this new method is to separate the reweighting stationary density from the reweighting of the correlation function. We evaluate the effectiveness of this approach on several analytical potentials and on a model biomolecular system, comparing its performance with the original method. Our results show that $π$-Girsanov not only improves the estimation in a single-ensemble setting, but also resolves key challenges in estimating transition matrices from multiensemble and non-equilibrium biased trajectories. Overall, $π$-Girsanov represents a substantial advance in kinetic reweighting, strengthening the connection between enhanced sampling techniques and Markov state modeling.

Real-space singularities underpin diverse wave phenomena yet remain largely unexplored in elastic wave systems. We report the observation of real-space topological singularities in structured flexural waves on finite-sized solids. These singularities are robust against perturbations and annihilate only through topological phase transitions. Moreover, they imprint dislocation lines on the radiated sound field, generating acoustic vortices in free space from an achiral source and structure. Our findings bridge continuum mechanics and topological physics, establishing elastic waves as a platform for exploring complex topological textures in real space and paving the way towards singular phononics.

We present a unified heterogeneous computing framework for real-time time-dependent density functional theory (RT-TDDFT) based on numerical atomic orbitals (NAOs), implemented in the ABACUS package. We introduce three co-designed abstraction layers, including unified data containers, unified linear algebra operators, and unified grid integration interfaces. These layers collectively accelerate the two most demanding parts of NAO-based RT-TDDFT: explicit real-time wavefunction propagation and real-space grid operations such as Hamiltonian construction and force evaluation under external fields. We validate the method by computing optical properties for systems ranging from finite molecules to periodic solids, showing excellent agreement with standard benchmarks. Performance evaluations on bulk silicon demonstrate that a single GPU can achieve substantial wall-clock speedup over a fully utilized dual-socket CPU node. Furthermore, distributed multi-GPU strong-scaling tests confirm high parallel efficiency over tens of GPUs. This work establishes a high-performance, portable platform for large-scale first-principles simulations of ultrafast electron dynamics.

In the present study, wakefield amplification via coherent resonant excitation using two co propagating laser pulses in a homogeneous plasma is investigated. The proposed scheme is based on linearly polarized leading seed pulse followed by a trailing pulse with identical or controlled parameters, enabling phase synchronized energy transfer to the plasma wave. By systematically varying the temporal pulse widths and inter pulse separation, conditions for resonant enhancement of the wakefield are established. Analytical modelling, supported by particle in cell simulations, reveals that maximum amplification occurs when the pulse separation approaches a quarter of the plasma wavelength, ensuring constructive interference of the plasma oscillations driven by successive pulses. Under optimal conditions, the coherent resonant excitation leads to a significant enhancement of the wakefield amplitude, reaching up to three times of that produced by a single laser pulse. The results demonstrate that precise control of pulse spacing and duration enables efficient energy coupling into plasma waves, providing a robust pathway for enhanced wakefield generation in laser plasma interaction regimes.

Integrated visible photonic engines for solid-state quantum defects provide a foundation for scalable quantum networks. While miniaturization is advancing, active manipulation remains limited by the difficulty of achieving simultaneous milliwatt-scale visible light generation and high-contrast modulation. Despite extensive efforts, the concurrent chip-scale realization of nonlinear frequency conversion and fast temporal gating for high-fidelity quantum control has remained elusive. Here, we demonstrate a monolithic thin-film lithium niobate (TFLN) platform integrating periodically poled frequency conversion with GHz-bandwidth electro-optic (EO) switching. The device delivers off-chip green-light power exceeding 1 mW with an extinction ratio (ER) of 42.2 dB, enabling coherent spin control and time-resolved lifetime measurements of individual nitrogen-vacancy (NV) centers in diamond through nanosecond gating. System performance is validated through pulsed optically detected magnetic resonance (ODMR), Rabi oscillations, and Ramsey interference, supported by time-tagged photon counting with nanosecond resolution. By unifying sufficient nonlinear light generation with high-speed active manipulation, this platform establishes a scalable framework for the realization of high-rate quantum communication nodes.

Raman scattering underlies a broad range of spectroscopic and light-generation techniques, yet its conventional description, based on the Raman gain spectrum, accurately describes only long-pulse, steady-state dynamics. We present and develop a time-domain theoretical approach that provides a unified and physically-transparent description of Raman interactions across all temporal regimes. It enables direct visualization of Raman temporal dynamics and accounts for spectrotemporal aspects of Raman phenomena. We apply this theory specifically to Raman-shifting with ultrashort light pulses in gases, where the excitation is in the impulsive Raman regime and dephasing of Raman transitions is weak. The analysis, for the first time, exposes temporal and spectral distortions that arise from Raman scattering and which impact frequency-shifting performance detrimentally. Crucially, it also identifies how these distortions can be suppressed through temporal control of the nonlinear response. Numerical simulations of the soliton self-frequency shift (SSFS) in gas-filled hollow-core fibers show that molecules with strong Raman responses do not yield efficient frequency conversion, and predict that reducing the relative Raman contribution (compared to the electronic response) enhances the process. Experiments using gas mixtures with tunable Raman fraction of the nonlinear response confirm these predictions. An analytic expression for impulsive SSFS in gases, which departs significantly from the well-known formula for glasses, predicts the observed behavior when Raman-induced temporal distortion is suppressed. The new time-domain framework uncovers phenomena and provides physical insight that are inaccessible through the decades-old frequency-domain treatment of Raman scattering.

Since the publication of our review article Hyperspectral imaging solutions for brain tissue metabolic and hemodynamic monitoring: past, current and future developments in 2018, the technological and applicational landscape of the use of hyperspectral imaging (HSI) in brain sciences has evolved and transformed significantly. The number of studies and works where HSI has been deployed in its many forms to map and monitor the haemodynamic and metabolic states of cerebral tissues have grown exponentially, to such a point where an update on the cur-rent state of the art is timely, and we believe would be desirable for both long-term experts in the field, as well as for any new researcher approaching it for the first time. In this commentary, we provide a renewed perspective on the newest and latest developments in brain haemodynamic and metabolic monitoring with HSI over the past eight years. Our hope is that even greater breakthroughs and broader, more numerous novel applications will come forward in the future for the technology, that may benefit from this new overview, as they did from the original one.

Stimulated Brillouin microlasers offer chip-scale light sources with high spectral purity and low phase noise--key attributes for applications spanning precision metrology, quantum technologies, and coherent information processing. However, simultaneously bringing both pump and scattered waves into resonance often compromises photon confinement or modal volume, resulting in limited conversion efficiency and elevated thresholds. In this work, a novel approach is proposed to generate Brillouin microlasers with high efficiency, low threshold, and narrow linewidth, by combining a cross-polarized stimulated Brillouin scattering scheme with intentional Stokes mode splitting to compensate for mode detuning. Triple-resonance and phase-matching conditions are simultaneously achieved in a 114-um-diameter thin-film lithium niobate (TFLN) microresonator, enabling precise alignment with both the ~10-GHz Brillouin shift and the ~100-MHz narrow gain bandwidth. The resulting Brillouin microlaser achieves a narrow intrinsic linewidth of 2.88 Hz, a short-term integral linewidth of 185 Hz, an on-chip conversion efficiency of 57.92%, and a pump threshold as low as 1.03 mW. Both the conversion efficiency and the lasing threshold represent record-high performance for the TFLN platform to date.

We study the use of a learned forward operator in the inverse problem of photoacoustic tomography. The Fourier neural operator to approximate the photoacoustic wave propagation is used. Further, the inverse problem is solved using a gradient-based approach with automatic differentiation. The methodology is evaluated using numerical simulations, and the results are compared to a conventional approach, where the forward operator is approximated using the pseudospectral $k$-space method. The results show that the learned forward operator can be used to approximate the photoacoustic wave propagation with good accuracy, and that it can be utilised as a computationally efficient forward operator in solving the inverse problem of photoacoustic tomography.

We demonstrate a single-arm optical platform for phase-retrieval-free, quantitative dynamic phase mapping of continuous transparent media via field-correlation ghost imaging. By modeling the medium as a dynamic pure-phase object, we spatially encode and compress its two-dimensional (2D) complex transmittance into a single bucket detector. Balanced heterodyne detection downconverts the optical frequencies for direct digitization. Crucially, by mapping spatial information into the temporal domain, this single-pixel architecture exploits high-speed digitization to continuously resolve 2D phase dynamics, effectively bypassing the frame-rate bottlenecks of traditional array sensors. Coupled with intermediate-frequency spectral analysis, this establishes a direct linear mapping from the recorded signal to the physical phase. The complex amplitude is thus deterministically extracted via field-correlation, enabling the spatial reconstruction of 2D acoustic pressure distributions using a pseudo-inverse algorithm. Experimental validations in an acoustic levitator confirm that the optically extracted acoustic wavelengths strictly match theoretical dispersion models, exhibiting a robust linear correlation between the retrieved phase shift and local sound pressure levels. This deterministic methodology provides a real-time-capable metrological tool for characterizing rapidly evolving phenomena, including transient aeroacoustic flows, shockwaves, and microfluidic biological dynamics.

The conventional description of time-varying media assumes that electromagnetic fields evolve according to fixed continuity conditions during parameter jumps. Here we reveal that these conditions are not physical constraints but tunable design degrees of freedom. By developing a unified framework that treats continuity rules as engineerable parameters, we expand the scope of time-varying metamaterials and enable wave phenomena previously considered impossible. For instance, non-resonant, reflectionless wave amplification without momentum bandgaps, and reversible conversion between propagating waves and static fields for optical memory, etc. This work opens a new dimension for controlling light-matter interactions.

The constancy of the speed of light (the maximum velocity of interaction) is the second postulate of Albert Einstein's special theory of relativity. Currently, there is no correct theoretical proof of this constancy in all inertial frames of reference. This paper presents such a proof, demonstrating that quantum mechanics (quantum field theory) can only be formulated under the condition of the constancy of the speed of light in a vacuum. It has been established that this constancy is determined by the minimum energy of the particles. When this minimum is reached, two identical solutions emerge -- one with positive and one with negative energies. Thus, within the framework of classical physics, the existence of particles and antiparticles is demonstrated. It is shown that matter dominates over antimatter.

Inspired by biology's use of vesicles for targeted transport, many studies have propelled liposomes with active matter, creating synthetic systems that can be viewed as microscale biohybrid robots. Nevertheless, the underlying motility mechanisms from a hydrodynamic perspective are often unresolved, and reliable velocity control remains challenging. Here we present a chlamylipo formed by encapsulating the motile alga Chlamydomonas reinhardtii within a giant liposome. We quantify how the characters of swimming change under controlled perturbations and, from a fluid-mechanical perspective, derive a deformation-velocity expression that incorporates liposome radius, beating frequency, and membrane protrusion. We further show that motility can be reversibly switched by incorporating light-responsive lipids, with the liposome acting as a "clutch" that modulates membrane-coupled propulsion. Thus, liposome encapsulation can function not only as a cargo compartment but also as a tunable motility regulator, enabling speed adjustment and reversible transitions between motile and non-motile states.

A fundamental question in nonequilibrium statistical physics is whether effective equilibrium behavior can emerge at coarse-grained scales in strongly driven systems. Here, we investigate this question in the context of human mobility by analyzing five years of intercity flow data covering millions of travelers. While short-term flows are highly asymmetric, temporal coarse-graining reveals that over half of all city pairs converge toward effective flow balance, with normalized directional imbalance decaying as a power law. The remaining pairs either exhibit persistent drift-dominated currents or a crossover between these two extremes. A stochastic model decomposing mobility into directional drift and correlated fluctuations quantitatively captures the coexistence of all three regimes. Directly measured variance scaling of the fluctuation process confirms near-diffusive behavior with regime-dependent deviations. These results demonstrate that large-scale mobility networks exhibit a scale-dependent transition from broken to restored flow symmetry, with direct implications for modeling transport and spreading dynamics.

The evolution toward next-generation intelligent sensing requires microwave systems to move beyond static detection and achieve high-speed and adaptive perception of dynamic scenes. However, the existing microwave sensing systems have bottlenecks owing to their sequential digital processing chain, limiting the refresh rates to hundreds of hertz, while the existing integrated microwave processors are lack of programmable and scalable capabilities for robust and open-world deployment. To break the bottlenecks, here we report a programmable surface plasmonic neural network (P-SPNN) that enables real-time microwave sensing and automatic recognition of dynamic objects in open-world environment. With a perception latency of 25 ns and a refresh rate exceeding 10 kHz, the P-SPNN system operates more than two orders of magnitude faster than the conventional millimeter-wave sensors, while achieving an energy efficiency of 17 TOPS per W. With 288 programmable phase-modulated neurons, we demonstrate real time and robust classification of persons and cars with 91-97% accuracy in the open road scenarios. By further integrating beam-scanning function, P-SPNN enables multi-dimensional spatial temporal frequency sensing without the digital preprocessing. These results establish P-SPNN as a programmable, scalable, and low-power platform for high-speed perception tasks in realistic world, with broad implications for autonomous driving, intelligent sensing, and next-generation artificial intelligence hardware.

Current studies on activity space are limited by the conceptualization of absolute physical space that fails to consider the heterogeneity of relational spaces reconstructed from spatial interactions of human movements between locations and falls short in incorporating the inherent hierarchical property of human mobility. Consequently, these approaches cannot faithfully reflect how people interact with urban spaces through travels. From the lens of relational space, this study proposes the new Hierarchical Activity Region Model (HARM) to derive the space and hierarchical properties of activity spaces perceived by various urban groups. We demonstrate the enhanced validity of our model on travel behavior in Manhattan, New York City, before, during, and after Hurricane Sandy on the basis of taxi data. Empirical results show that intra-urban travel retains clear hierarchical organization, even under disruption of a major weather event. Yet, travel undergoes a compression effect in travel hierarchies, characterized by fewer hierarchical levels and enlarged characteristic scales, followed by a rebound. Clustering the derived hierarchies reveals pronounced heterogeneity that stems from differences in population profiles; some groups sustain deeper structures or recover quickly, while others experience a persistent loss of levels. This study provides valuable insights into the functional hierarchies of urban mobility, which could inform more sustainable, resilient and equitable urban planning. The proposed methodological framework is generic for studying human mobility in broader contexts.

The rapid evolution of next-generation communications and the Internet of Things (IoT) has catalyzed an urgent demand for governing expansive spatial environments as functional electromagnetic (EM) entities. However, deterministically programming such open EM spaces remains a formidable challenge, as current methodologies are largely confined to localized interfaces that lack the collective coordination required to orchestrate unbounded environments. Here, we introduce a general framework for the deterministic programming of EM space via cooperative metasurface clusters, achieved by mapping volumetric field interference landscapes onto a virtual nodal network. By representing excitations and meta-atoms as fully interconnected nodes, we transform intricate non-local interactions into tractable nodal states, enabling the precise quantitative synthesis of spatial scattering. This framework bridges local meta-atoms with global EM environment to program space as a functional entity, as demonstrated by a deeply coupled meta-emitter for programmable collective radiation and metasurface clusters that sculpt angle-resolved illusion spaces. By transitioning from individual components to cooperative multi-body assemblies, our work provides a scalable foundation for next-generation wireless networks, wave-based analog computing, and ambient intelligence, where space itself becomes a coherent functional and reconfigurable entity capable of holistic information management.

Topological concepts have been at the forefront of materials research in recent years, driving a revolution in our understanding of the response of quantum materials and enabling new ways to manipulate light and sound in topological metamaterials. Topological defects and topological boundaries of different dimensions have driven a paradigm shift in photonics, where topological photonic crystals and metamaterials can be engineered to create one-way flow of energy robust to defects or to control such flows with synthetic degrees of freedom along topological domain walls. More recently, topological point singularities encoded into photonic structures have been shown to enable confinement of optical modes with the topologically nontrivial nature of the cavity imprinted into the vorticity of optical far fields. Here we demonstrate that the two latter concepts - domain wall and point singularities - can be unified into an even more powerful tool to enable arbitrarily shaped resonant cavities of any dimension supporting spectrally stable zero-energy modes. We experimentally confirm that such modes, whose existence is guaranteed by topological principles, allow an unprecedented degree of control over the optical field, which appears to have no phase modulation across space, can have any desirable radiation pattern, and enables spectral stability regardless of shape or length.

Negative and complex numbers are so familiar in modern mathematics, physics, and engineering that it is easy to forget how uncertain their status once was. They did not become established through a single route. This article follows four linked processes in their stabilization: operational use, formal legitimation, pedagogical normalization, and physical naturalization. Negative quantities appear early in Chinese rod arithmetic and Indian debt--fortune rules, were reshaped in medieval Islamic algebra, and remained conceptually unstable in early modern Europe even when they worked in practice. Complex quantities followed a different path: they first appeared as troubling by-products of algebraic formulas, then gained stability through Bombelli's rules, geometric representation, nineteenth-century analysis, and later applications in circuits, wave theory, optics, and quantum mechanics. Franklin's electrical plus and minus helped make sign physically intelligible, while electrical engineering turned impedance and complex amplitudes into routine tools. The broader lesson is that these quantities became natural through repeated interaction among calculation, representation, teaching, and experiment.

This paper presents a physics-based channel modeling and optimization framework for reconfigurable intelligent surface (RIS)-assisted downlink multi-user multiple-input single-output (MU-MISO) communication systems in site-specific environments. A hybrid ray-tracing (RT) and full-wave electromagnetic analysis approach is developed to construct a deterministic channel model that explicitly captures multipath propagation, RIS scattering behavior, and mutual coupling effects through a non-diagonal load impedance representation. Based on this model, an alternating optimization scheme jointly updates the base-station (BS) beamformer and RIS load impedances to maximize the minimum achievable rate under a total transmit power constraint and practical capacitance limits. The objective of the proposed framework is to provide a reliable initial assessment of the system-level impact of RIS deployment in realistic propagation scenarios. To evaluate this capability, the RIS is operated in a column-paired 1-bit control mode that enables exhaustive evaluation of all realizable configurations in both simulation and measurement. Performance is compared at the distribution level through achievable-rate histograms across all configurations and further examined under small user-location variations. The observed agreement between simulation and measurement demonstrates that the proposed framework reliably captures practical performance trends and provides useful guidance for the design and deployment of RIS-assisted MU-MISO systems in site-specific environments.

Thin films are preferred for high photocurrent conversion efficiency, but strong photon absorption at photon energies below the bandgap (near and shortwave infrared) typically requires thicker semiconductor layers. To address this tradeoff, various optical approaches have been proposed, including light scattering within the active layer, reducing surface reflection, and using resonant structures to improve light confinement, trapping, and coupling. However, resonant structures often operate over a narrow spectral range, limiting their use of the full solar spectrum, and can involve complex fabrication and careful structural design. In this work, I propose a new method to enhance absorption in semiconductor solar cells across wide angular and spectral ranges for both polarization states (transverse electric (TE) and transverse magnetic (TM)). The method is based on a coupled plasmonic waveguide resonance (CPWR) configuration excited in a planar layered structure that can be fabricated using simple deposition techniques. Using the proposed approach, as an example, the thickness of the required Silicon (Si) layer can be reduced from approximately 130 to 180 μm (the typical Si thickness in commercial solar cells) to only a few microns. The method enables efficient harvesting of the infrared portion of the solar spectrum. By exciting CPWRs, the method overcomes the sharp drop in the absorption spectrum of conventional Si solar cells at wavelengths longer than 1100 nm. Field calculations demonstrate that light is efficiently absorbed in the Si layer at the resonant wavelengths. The proposed approach is general and can be applied to different types of semiconducting and prism materials. To maintain high absorption at wavelengths below 1100nm, an additional semiconductor metal semiconductor configuration is proposed, in which a thinner Si layer is added beneath the metal layer.

All-optical logic gates have significantly advanced over a diverse range of photonic systems, boosted by intricate nonlinearities that facilitate the engineering of complex logic operations. Here, we demonstrate that in semiconductor microcavities, polariton condensates trapped in Ouroboros-shaped rings form specifically charged vortices, determined by the strength of nonlinearity and the excitation method. Quantized vortex phases encode binary digits that can be nonresonantly controlled by optical pulses incident directly upon the ring, enabling logic operations. By interconnecting three polariton Ouroboros rings, we realize a universal set of logic gates (AND, OR, NIMPLY) fundamental to functional polaritonic devices. The Ouroboros structures are highly customizable, providing a robust and promising platform for exploring more complex logic operations.

The rapid growth of AI and accelerator-driven workloads is forcing a fundamental rethinking of optical interconnect architectures in datacenters. Co-packaged optics and three-dimensional photonic integration have emerged as promising solutions to overcome the energy and bandwidth limitations of electrical I/O. Yet, as optics move closer to compute, packaging, thermal management, and system-level robustness increasingly dominate performance and scalability. Here, we argue that co-packaged optics should not be viewed as a component-level optimization, but as an architectural commitment that reshapes the boundaries between photonics, electronics, and system design. We examine how heterogeneous integration strategies, chiplet-based optics, and emerging packaging platforms redefine scaling laws for AI systems, often introducing trade-offs that are underappreciated in device-centric analyses. Looking forward, we discuss why standardization, serviceability, and thermal-aware co-design will be decisive in determining whether co-packaged optics can transition from early deployment to widespread adoption in AI-scale datacenters.

Graphene supports surface plasmon polaritons (SPPs) with extreme field confinement and electrical tunability, but these waves are typically short-lived due to ohmic loss in the sheet. We show that embedding graphene in an active dielectric can counteract this loss and we derive closed-form design rules to do so, based on gain-assisted plasmonics and plasmonic amplification concepts. Specifically, from the full Maxwell model of a conductive sheet we obtain (i) the exact gain required for lossless plasmon propagation, and (ii) a second critical gain that marks the $\mathcal{PT}$-symmetric threshold, the exceptional point separating propagating and forbidden SPP regimes. The formulas are expressed directly in terms of the complex conductivity of graphene and the surrounding media, making them easy to evaluate and implement. We verify the theory with full-wave eigenmode calculations (COMSOL), showing dispersion and attenuation/amplification trends with and without gain for our plasmonic structures, finding a practical route to engineer long-range, tunable, lossless graphene plasmonics and to map/target non-Hermitian operating phases for device design in single- and double- layer graphene surfaces.

Coupled partial differential equations underpin a wide range of multiphysics systems, yet existing neural PDE solvers still struggle to resolve localized high-risk regions and often fail to preserve structural admissibility across coupled fields. To address these limitations, we propose the Partition-of-unity Information-geometric Certified Solver (PICS), a closed-loop framework that strictly enforces structural admissibility at the level of representation rather than relying on an additional soft penalty. By constructing a gate-structured admissible manifold coupled with a restricted jet prolongation, PICS ensures that geometry-sensitive approximations and closure-essential differential coordinates enter the solver as a strongly enforced, structure-preserving ansatz. Furthermore, the framework integrates entropic tail-risk control and \textit{a posteriori} certificate-driven empirical measure transport, dynamically reallocating training efforts toward uncertified, error-prone transition zones. Evaluated against standard baseline methods across three two-dimensional coupled benchmarks, PICS achieves more consistently accurate and balanced cross-field recovery while retaining practical computational efficiency, thereby providing a rigorous route toward highly reliable multiphysics simulation.

We investigate the evolution of Anomalous Hall Conductivity (AHC) in a coplanar and collinear antiferromagnetic system with varying spin canting angles. A tight-binding model based on three t2g-orbitals in a body-centered tetragonal lattice is constructed, where the inclusion of third-nearest neighbor hopping is demonstrated to be essential for capturing the characteristic energy band splitting of altermagnetic materials. By employing a symmetry analysis based on spin space groups and treating spin-orbit coupling (SOC) as a perturbation, we theoretically distinguish and numerically verify two origins of the transverse transport: the conventional anomalous Hall effect (AHE) induced by net magnetization and the Crystal Hall Effect (CHE) arising from specific crystal symmetries. Our results show that the conductivity components driven by these two mechanisms follow distinct trigonometric dependencies on the canting angle. Crucially, we identify a hidden C110 rotational symmetry that has been previously overlooked in static magnetic group analyses. By expanding the AHC in terms of spin orientation vectors, we demonstrate that this symmetry acts as a bridge connecting distinct magnetic configurations with different canting angles, thereby strictly protecting the equivalence of orthogonal conductivity components in the collinear system.

Accurately resolving steady electrohydrodynamic (EHD) flows presents a formidable computational challenge due to the strong nonlinear coupling between charged-particle density, velocity fields, and electric potential. These interactions frequently induce sharp transition layers, crossing fronts, and multiscale spatial structures, which notoriously degrade the predictive accuracy of standard mesh-free solvers like Physics-Informed Neural Networks (PINNs). To systematically address this bottleneck, we formulate a unified four-variable operator framework and develop a comprehensive benchmark suite for two-dimensional steady EHD shock-like problems. The benchmark comprises eight rigorously designed cases featuring diverse front geometries, such as oblique, curved, and intersecting layers, alongside complex multiscale patterns. Under strictly identical configurations, including governing equations, source terms, sampling strategies, and loss formulations, we evaluate a Standard MLP-based PINN, a Residual Attention PINN (ResAtt-PINN), and an LSTM-PINN that leverages pseudo-sequential spatial encoding. Extensive numerical experiments demonstrate that the LSTM-PINN consistently achieves the highest predictive accuracy across all eight cases. It successfully reconstructs sharp gradients and intricate multiscale structures where other architectures fail or over-smooth. Furthermore, the LSTM backbone efficiently captures long-range spatial correlations while maintaining an exceptionally low computational overhead and GPU memory footprint. These findings not only establish the LSTM-PINN as a robust and efficient solver for strongly coupled PDEs with shock-like features, but also provide the computational physics community with a standardized, reproducible benchmark for future algorithmic evaluations.

Fundamental questions related to the roles of fuel type, combustion parameters, and turbulence transport interactions in the inception and growth of contrails have remained intractable in remote sensing and in-flight measurements. Consequently, we developed a novel laboratory-scale facility for studying the inception, growth and persistence of contrails for aircraft-relevant conditions. The exhaust gas generated using an inverted co-flow soot generator at a set of global equivalence ratios for two fuels - ethylene and propane is supplied to the contrail tunnel, which then mixes with an ambient flow emulating long-haul aircraft cruise conditions (20.8 kPa and 190 K). Detailed soot characterization using a scanning mobility particle sizer and transmission electron microscopy is coupled with measurements of instantaneous and averaged scattering intensities from the generated contrails. The experimental results are complemented by numerical simulations of the contrail tunnel using solutions of the Favre-averaged Navier-Stokes (FANS) equation and a two-equation model for handling particulate matter, including soot and ice. Results show, for the first time, the cross-section of a contrail, and the interaction of turbulent mixing and microphysical growth scales involved in ice nucleation across the shear layers. The average scattering cross sections of contrails increase with equivalence ratio, due to higher soot number concentrations and water vapor content. Comparisons between ethylene and propane exhausts indicate that the scattering propensity of contrails is more sensitive to exhaust water vapor content than to soot concentrations. Finally, depolarization measurements are used to show asphericity in ice crystal habits. Thus, our study present a unique window into contrail formation, theoretical modeling and simulation.

We report the first plasma atomic layer etching (ALE) process for diamond using a cyclic plasma sequence composed of two separated steps: oxygen surface modification and krypton ion removal. The process is implemented in an inductively coupled plasma reactor using alternating O$_2$ plasma exposure and low-energy Kr ion bombardment. This cyclic process exhibits the characteristic self-limiting behavior of ALE and enables controlled material removal with atomic-scale precision. An etch depth per cycle of \SI{6.85}{\angstrom} was achieved. Surface analysis reveals that the etched diamond surfaces exhibit lower roughness than the pristine material, while XPS confirms the preservation of the diamond bonding structure and indicates essentially damage-free etching. These results demonstrate that plasma ALE based on O$_2$/Kr chemistry provides a viable route toward damage-controlled nanoscale processing of diamond, opening new opportunities for advanced device fabrication in power electronics, photonics, quantum sensing and quantum computing technologies.

The introduction of network science approaches into public transport research has seen great advances in the past 15 years. However, it has become apparent that monolayer networks are often not sufficient to model and analyse real-world systems in sufficient detail. In the last decade, the theory of multilayer networks has proven to be an invaluable tool in various disciplines, including transport. Multilayer networks consist of layers of networks that are coupled among themselves. This enables modelling of complex systems with heterogeneous elements and relations between them. Although there is a body of work in public transport research that uses multilayer networks, the related literature is scattered, lacking unified terminology and agreed-upon approaches. We posit that there is vast uncovered potential in using multilayer network approaches to public transport modelling, planning, and operations. We first present the basic formalisms of multilayer networks with a focus on how they (may) relate to public transport networks. We then provide a systematic review of the literature on multilayer networks in public transport research. We identify and taxonomise ways in which public transport systems are modelled as multilayer networks. Based on the survey and drawing from the state and history of network science in public transport research as well as multilayer approaches across other application domains, we propose a research agenda for multilayer public transport networks for the upcoming decade(s).

In this paper, the global $χ^2$ function for the simultaneous fitting of correlated energy-dependent cross sections is constructed, where the correlations between the measured cross sections of different processes and/or at different center-of-mass energy points, as well as the contributions from the integrated luminosity measurement and the center-of-mass energy measurement, are taken into account.

The exotic physics emerging at singularities has long attracted intense theoretical and experimental attention. In non-Hermitian systems, exceptional points (EPs), unique spectral singularities, have given rise to a host of intriguing wave phenomena and enabled a broad range of promising applications across diverse physical platforms. Recently, considerable effort has been devoted to combining nonlinearity with exceptional points (EPs) to enable flexible control, overcome the limitations of linear EPs, discover previously unexplored singularities, and reveal novel physical phenomena and application potentials. In this review, we provide a detailed overview of the interplay between nonlinearity and EPs, highlighting key developments such as noise suppression for enhanced sensing, emerging mechanisms for chiral-like state transfer, the realization of optical isolators in nonlinear EP systems, applications including wireless energy transfer and frequency comb generation, among others. We also offer a perspective on future research directions and opportunities in this rapidly evolving field.

Neuromorphic electronics and spiking neural networks (SNNs) offer energy-efficient data processing, essential for real-time and edge-computing applications. In particular, interfacing and processing biological signals require devices that combine electronic performance with ionic sensitivity, which are capabilities uniquely provided by organic electrochemical transistors (OECTs). However, realizing a simple, fully integrated OECT-based neuron with rich dynamics and adaptability remains challenging. Most reported implementations rely on current-driven operation, which complicates large-scale integration and neuron-neuron coupling due to the need for precise matching of operating currents and bias voltages. Here we present a voltage-driven neuron circuit based on a multivibrator oscillator architecture, entirely fabricated from poly(3,4-ethylenedioxythiophene):polystyrene sulfonate (PEDOT:PSS). The neuron exhibits tunable adaptability through an additional control input, enabling switching between burst latency and length encoding modes. We further demonstrate a hardware-implemented two-neuron unit consisting of an inhibitory and a readout neuron, where readout activity is suppressed depending on the relative timing of the inhibitory input. Finally, we demonstrate that the fabrication process is compatible with polymer dendrite growth, enabling on-chip integration of synaptic elements on the same substrate. Owing to its structural simplicity and compatibility with a single, available material, this approach offers a scalable and accessible route toward integrated OECT-based SNNs.

We report on a novel type of dual-wavelength fiber laser with a structured-core design inside silica glass, forming a spatial separation of the several core areas doped with ytterbium and erbium ions. We have optimised the key parameters of the fiber core, such as the concentration of rare earth elements, and the optimal length of active fiber to operate simultaneously at two different wavelengths. Using the Modified Chemical Vapor Deposition method to obtain initial optical fiber preforms, and using the stack and draw technique, we have fabricated two types of active fibers, one with 7 and one with 19 rare-earth-doped rods (elements) forming the fiber core. We characterized the drawn fibers by investigating their structure by scanning electron microscopy, confirming the spatial separation of the elements within the core. Measuring absorption shows that concentration ratios Nt Yb: Nt Er were approximately 52: 48 for 7 core fibers and 56:44 for 19 core fibers. Lifetimes for both active fibers were 0.84 ms for Yb3+ ion and 10.30 ms for Er3+ ion. The performance of fiber lasers was determined, proving that fibers are capable of laser emission simultaneously at 1042 nm and 1550 nm. We have shown experimentally that the output power ratio between both lasing wavelengths can be controlled by the length of the fiber.

Plasma-based mid-infrared (MIR) laser generation has garnered significant interest owing to its advantage of high output power, continuous wavelength tunability, and ultrashort pulse durations. However, existing methodologies predominantly depend on high-intensity inputs at the hertz frequency level, with spectral energy concentrated near the central frequency, rendering them unsuitable for spectroscopic applications. This paper proposes and demonstrates a cascaded deceleration scheme that enables the generation of broadband MIR lasers with low energy inputs compatible with high-repetition-rate laser systems. By confining the input laser within a plasma channel, this approach preserves the laser intensity, which not only sustains the decelerating field strength but also enables the cumulative effect of deceleration across multiple distinct bubbles. Numerical simulations demonstrate that more than 30% of the 23 mJ input energy is converted into a broadband MIR output spanning wavelength from 0.58 to 6.86 μm, achieving peak powers on the order of gigawatts. The output exhibits unique time-frequency characteristics, defined by spectral sub-bands organized in a temporal sequence, wherein each sub-band comprises few-cycle pulses. Parametric analyses reveal that the spectral bandwidth broadens with increasing laser intensity, provided that the plasma density being adequate to ensure a sufficiently short deceleration length. This approach provides a practical, efficient route to broadband ultra-intense mid-infrared sources, promising for applications in Fourier transform spectroscopy and laser-induced electron diffraction.

Optical metasurfaces composed of metallic or dielectric scatterers (meta-atoms) promise a powerful way of tailoring light-matter interactions. Phase-change materials (PCMs) are prime candidates for non-volatile resonance tuning of metasurfaces based on a refractive index change. Precise resonance control can be achieved by locally applying laser pulses to crystallize a PCM, modifying the dielectric surrounding of meta-atoms. However, the complex crystallization kinetics of PCMs in the vicinity of metallic meta-atoms have not been studied yet. Here, we experimentally investigate metallic dimer antennas on top of the PCM Ge3Sb2Te6 and address these nanoantennas with laser pulses. Our study reveals inhomogeneous crystallization caused by the absorption and heat conduction of the metallic nanoantennas. A self-consistent multiphysics model, including electromagnetic, thermal, and phase-transition processes, is employed to simulate the crystallization and understand the resulting resonance shift of the antennas. This model enables the optimization of the laser parameters and the geometry of the meta-atoms to achieve an optimal crystallization pattern and resonance shift. Our work paves the way towards complex antenna geometries optimized for local addressing of PCMs to achieve sophisticated crystallization patterns, enabling on-demand programming of individual nanoantennas within metasurfaces.

Laser-driven ion acceleration from nanostructured targets offers a promising route to compact, high-energy ion sources. In this work, we demonstrate through particle-in-cell simulations that rectangular nanoring targets significantly enhance energy absorption and increase the cutoff energy of laser-accelerated ions. The nanoring geometry enables strong field confinement within its hollow core when optimally oriented relative to the laser polarization, leading to hotter electron populations and more robust sheath acceleration. These results demonstrate that rectangular nanorings offer a versatile platform for controlling laser-plasma interactions at solid densities and advancing compact, high-repetition-rate particle sources.

How microorganisms respond to and interact with their environment can vary significantly from individual to individual, which can have important microbiological and ecological implications. However, most microscopy techniques can only observe motile microorganisms for short times because of their limited fields of view. Using Lagrangian tracking, a single microorganism can be followed in 3D, potentially indefinitely, allowing to decipher individual phenotypical traits. Current Lagrangian tracking methods use the fluorescence signal emitted by the microorganism as feedback to keep it in focus. However, over long times, epifluorescent imaging can induce photobleaching and photodamage, and importantly, not all microorganisms can easily be made fluorescent. Additionally, traditional algorithms used in feedback loops to determine microorganism position are prone to errors, especially in optically complex media. Here, we present a faster, more reliable, and versatile Lagrangian tracking method that uses deep learning to determine the 3D position of the microorganism. This new method demonstrates enhanced accuracy and speed in tracking fluorescent bacteria with fluorescence microscopy also in optically complex media. Furthermore, we track bacteria with other microscopy modalities, such as brightfield microscopy -- for example, this enables us to track magnetotactic bacteria, which cannot be made fluorescent without degrading their magnetotactic properties. These novel capabilities allow to extract previously inaccessible quantitative information, significantly advancing the study of microorganism behavior -- and thus opening new avenues for research in complex biological and ecological systems.

Indirect control of ultraviolet (UV) pulse phase through nonlinear frequency conversion is attractive when direct UV pulse shaping is limited by material loss, dispersion, and damage threshold. Here we cast dispersive four-wave mixing (DFWM) as a pump-conditioned spectral kernel and show that, in a locally one-to-one mapping regime, the signal-to-idler conversion admits a practical transfer function description. Starting from the exact frequency domain expression, we rewrite the idler field as a linear operator acting on the conjugated signal spectrum, with a two-frequency kernel set by the pump self-convolution and phase matching. Linearization around a reference operating point then yields a spectral phase-response kernel for small input perturbations. By probing this response with sinusoidal spectral-phase modulation of different spatial frequencies, we define a spectral phase-modulation transfer function (SPMTF), an MTF-like measure of the phase-transfer bandwidth of the nonlinear interaction. Simulations with different pump group-delay dispersion (GDD) values produce distinct SPMTF curves, showing that pump chirp directly controls how much fine spectral-phase structure survives the conversion. This framework provides a simple way to compare operating conditions and identify regimes favorable for programmable NIR-to-UV phase transfer.

Terahertz bands enable ultra-broadband wireless communications but require compact, low-cost, and efficient transceiver modules. Conventional implementations based on metallic waveguides or silicon lenses suffer from high loss, bulkiness, and fabrication complexity. Here, we present a compact terahertz transceiver module enabled by a resonant tunneling diode (RTD) integrated with a photonic-electronic antenna chain. The RTD on InP is coupled to a modified Vivaldi antenna and an all-silicon effective-medium-clad waveguide, terminating in a rod antenna interfaced with a 3D-printed cyclic olefin copolymer lens. This architecture enables broadband directive radiation without matching networks or anti-reflection coatings. Packaged in a low-cost 3D-printed PLA enclosure, the module achieves realized gains of 28-33 dBi (E11x) and 30-33 dBi (E11y) across 220-330 GHz. As a receiver, it exhibits a noise voltage density of 5.6 x 10^-9 V/sqrt(Hz), a minimum noise equivalent power of 1.8 pW/sqrt(Hz), and an average responsivity of 6.8 kV/W. It supports error-free transmission up to 30 Gbit/s (OOK) and 80 Gbit/s (16-QAM) over 10 cm, and enables real-time uncompressed high-definition video streaming over 1 m. As a transmitter, it achieves error-free OOK transmission up to 12 Gbit/s at 332 GHz. These results demonstrate a promising terahertz transceiver architecture for 6G systems.

In this paper, we develop a Localized Orthogonal Decomposition (LOD) method for the two-dimensional time-dependent nonlinear Schrödinger equation with a wave operator. We prove that our method preserves conservation laws and admits a unique numerical solution; furthermore, we obtain unconditional (i.e., time-step restriction-free) optimal-order superconvergent \(L^p\) error estimates. To complement the theoretical analysis, we present a series of numerical simulations that verify the analytical results and further illustrate structural aspects of the problem.

In various engineering fields including mechanical, aerospace, and civil engineering, the identification of modal parameters, including natural frequencies, damping ratios, and mode shapes, is crucial for determining the vibration characteristics of engineered structures. A common method for identifying the modal parameters of structures involves experimental modal analysis using frequency response functions (FRFs) obtained from forced vibration tests. The least squares complex frequency (LSCF) domain method is a widely-used frequency-domain curve-fitting method for the FRFs using the polynomials of high order, which can extract modal parameters with high accuracy. However, increasing the polynomial order tends to result in the generation of non-physical spurious poles that need to be eliminated from the stability diagrams. To overcome this issue, we propose a method that strategically destabilize the stable yet spurious poles of the characteristic polynomials by making their coefficients as sparse as possible, via orthogonal matching pursuit (OMP). This results in sparse stability diagrams because unstable poles can be eliminated from the diagrams. In this paper, the proposed method is first applied to a numerically-obtained FRFs of a rectangular plate using finite element model, and its validity is discussed. Then, the method is applied to experimentally-obtained FRFs of rectangular plates with low-damping and with high-damping. Furthermore, to confirm its applicability to industrial applications with realistic complexity, it has also been applied to the FRFs of the electric machine's stator core used for electric vehicles. Based on the results, we have confirmed that the spurious roots can be eliminated from the stability diagrams without compromising accuracy for the cases considered.

Active wavefront control in high-power laser illumination systems is important for technologies such as additive manufacturing, free-space laser communication, and power transmission. Conventional spatial light modulators (SLMs) and mechanical beam-steering devices are unsuitable for such applications as they rely on metal mirrors and electrical contacts which are damaged under high laser irradiances. Here, we report on the design and realization of an optically addressable metasurface liquid crystal (LC)-based SLM for the modulation of high-power transmitted light. Our device uses a photoactive top contact which is optically addressed with a patterned 435 nm laser, creating a transient electrical contact that selectively switches the underlying LC medium. A TiO$_2$ metasurface, resonant in the 915-985 nm wavelength range, is embedded within a thin (~2 $μ$m) LC layer and enables large optical tunability. We demonstrate 90$^\circ$ linear polarization rotation in reconfigurable patterns across a 5x5 mm$^2$ active area with an overall transmittance of >60%. Additionally, we develop a multiphysics approach to simulate transmittance modulation in our device by modeling the LC interactions with TiO$_2$ nanopillars under an applied electrostatic field. This model exhibits good agreement with measurements and provides improved understanding of how LCs interact with both transmitted light and nanoscale metastructures in active devices. We show that our design and fabrication approach can yield high-efficiency transmissive metasurface SLM devices and lay the groundwork for the design of future LC-based active nanophotonics.

The development of many scalable quantum technologies requires single-photon nonlinearity, such as single-photon blockade, in solid-state systems. Recently, it has been shown that single-photon Fock states can, in principle, be unconditionally generated using arbitrarily small intrinsic optical nonlinearities in photonic cavities. We investigate the feasibility of such a scheme in achieving photon blockade in an on-chip silicon photonics platform. We show that a triply resonant nanobeam cavity pumped with three monochromatic lasers could achieve such functionalities with quality factors $\sim 10^7$ and effective mode volumes $\sim 10^{-2} μm^3$, for experimentally feasible incident powers. Using quantum optical simulations, we propose an experimental protocol to generate single photons under this scheme. The constraints on the cavity design and experimental conditions are thoroughly explored to determine feasible regimes of operation.

Previous studies of reverse-bias stability in perovskite solar cells have focused primarily on voltage controlled reverse-bias tests. Here we instead present an investigation of perovskite solar cell degradation under well-defined, constant reverse-current stress. We show that the choice of hole-transport layer dictates the dominant degradation pathway: cells using thick poly(triphenylamine) (PTAA) layers with better indium-doped tin oxide (ITO) coverage can tolerate high reverse bias but quickly undergo catastrophic breakdown under fixed reverse current near their one-sun maximum power-point. In contrast, cells modified with the phosphonic-acid interface layer MeO-2PACz, with poorer ITO coverage compared to PTAA, exhibit soft, gradual, and largely recoverable degradation, regardless of the shading conditions. For MeO-2PACz devices, degradation increases with both current magnitude and duration. Importantly, when normalized by injected charge (current times duration), lower currents applied over longer times cause more severe degradation than higher currents over shorter periods. Combining electrical measurements with spatially resolved photoluminescence imaging, we argue against shunt formation and instead support an ion- and charge-mediated interfacial electrochemical degradation mode.

Copper fire-through metallization is a cost-effective alternative to Ag counterpart for industrial high efficiency solar cells. The fire through dielectric metallization relies on belt speed, which dictates the ramp up and ramp down rates for effective contact formation. In this paper three belt speeds (325oC, 360oC, 390oC) at constant peak firing temperature, were used to process PERC (homogeneous emitter) cells. After the contact firing the electrical parameters were dependent on belt speed, but after LECO treatment, they were identical. The SEM/EDS cross sectional analyses showed increased elemental Cu with belt speed, and the series resistance was lowest for the middle belt speed before LECO. However, after the LECO treatment, the series resistance dropped, respectively, to 0.503 ohm-cm-2, 0.428 ohm-cm-2 and 0.500 ohm-cm-2 leading to efficiency of 20.8% on homogeneous PERC emitter.

The formation of stable copper-silicide (Cu3Si) interfaces is crucial for cost-effective, high-efficiency solar cells. However, copper's diffusivity and electromigration issues pose challenges for contact stability. This study employs Laser-Enhanced Contact Optimization (LECO) to induce localized nano-scale Joule heating at the Cu-Si interface in phosphorus-doped p-PERC solar cells. High-resolution STEM and bright field analyses confirm stable Cu3Si formation in LECO-treated samples, with significantly reduced material segregation compared to nonLECO samples. SEM and post-etch EDS mapping demonstrate improved chemical resistance and interface cleanliness. Electrically, LECO treatmenet reduces series resistance by a factor 3, enhancing fill factor and efficiency while preserving diode quality. These results highlight LECO as a scalable method for reliable, silver-free solar cell metallization.

As many research papers in astronomy have been written since the beginning of the 21st century as had been written previously. This exponential growth has been accompanied by substantial changes in the structure of astrophysics research, which organizations perform it and where they are located. Using data from the Smithsonian/NASA Astrophysics Data System/Science Explorer (ADS/SciX) we have obtained an article number and citation based set of metrics as a function of the institutional affiliation of the first author; nearly every organization which has produced recent astronomy research is included. We use these data to examine changes in where astronomy research is being done. We demonstrate how to create custom rankings for the organizations. We develop a dashboard of key performance indicators (KPI) to examine the relative and absolute changes in the research performance for each of the 1949 organizations which have produced at least one first authored, refereed astronomy journal article since 1997. We also present KPI dashboards for 65 countries and three regions.

For a long periodic chain of Bose condensates prepared in the free space, the subsequent evolution and interference dramatically depend on the difference between the phases of the adjacent and more distant condensates. If the phases are equal, the initial periodic density distribution reappears at later times, which is known as the Talbot effect. For randomly-related phases, we have found that a spatial order also appears in the interference, while the evolution of the fringes differs with the Talbot effect qualitatively. Even a small phase disorder is sufficient for qualitatively altering the interference, though maybe at long evolution times. This effect may be used for measuring the amount of coherence between adjacent condensates and the correlation length along the chain.

Converting mid-infrared (MIR) radiation to visible or near-infrared wavelengths is essential for imaging and sensing, yet achieving sensitive, low-power, and scalable detection remains challenging. Lanthanide nanocrystals provide an alternative through ratiometric luminescence but are typically constrained by Boltzmann statistics, which tie population distributions to lattice temperature and limit signal contrast. Here we show that MIR irradiation rebalances dissipative relaxation pathways, driving lanthanide emitters into a non-Boltzmann steady state that enables non-thermal control of population distributions. This allows emission behaviors inaccessible under thermal equilibrium. We exploit this regime to achieve linear MIR detection with respect to MIR power across 6.8 to 8.6 micrometers. The ratiometric response is intrinsically independent of the pump power, enabling operation at an ultralow excitation power of 10 uW, several orders of magnitude lower than conventional approaches. Using standard silicon photodetectors, we then demonstrate room-temperature MIR imaging with detection limits approaching 4 nW um-2. Our results establish lanthanide nanoparticles as an efficient platform for MIR conversion and sensing in nanophotonic systems.

We study population dynamics through a general growth/degrowth-fragmentation process, with resource consumption and unbounded growth/degrowth, birth and death rates. Our model is structured in a positive trait called energy (which is a proxy for any biological parameter such as size, age, mass, protein quantity...), and the jump rates of the process can be arbitrarily high depending on individual energies, which has not been considered yet in the literature. After a preliminary study to construct well-defined objects (which is necessary contrary to similar works, because of the explosion of individual rates), we consider a classical sequence of renormalizations of the underlying process and obtain a tightness result for the associated laws in large-population asymptotics. We characterize the accumulation points of this sequence as solutions of an integro-differential system of equations, which proves the existence of measure solutions to this system. Furthermore, if such a measure solution is unique, then our tightness result becomes a convergence result towards this unique process. We illustrate our work with the case of allometric rates (i.e. they are assumed to be power functions) and eventually present numerical simulations in this allometric setting.

The human brain achieves its remarkable computational prowess not despite its inherent non-ideal factors noise, heterogeneity, structural irregularities, decentralized plasticity, systematic errors, and chaotic dynamics but precisely because of them. This paper systematically demonstrates that these traits, long dismissed as imperfections in classical neuroscience and eliminated in digital engineering, are evolutionary design principles that endow the brain with robustness, adaptability, and creativity.

The costly provision of public goods serves as a model problem for the evolution of cooperative behavior, presenting a social dilemma between the collective benefits of shared resources and the individual incentive to free-ride in resource production. The spatial structure of populations can also impact cooperation over public goods, as diffusion of public goods and intentional motion of individuals towards regions with greater resources can interact with population and public goods dynamics to produce heterogeneous patterns in the spatial distribution of strategies and resources. In this paper, we build off a model introduced by Young and Belmonte for the reaction dynamics of interacting individuals and explicit public good, deriving a system of PDEs that describes the spatial profiles of strategies and the public good in the presence of both diffusive motion of individuals and resources and chemotaxis-like directed motion of individuals in response to gradients in the concentration of public goods. Through linear stability analysis, we show that spatial patterns in strategic and public goods profiles can emerge due to either Turing instability with high defector diffusivity or a directed-motion instability through strong sensitivity of cooperators towards increasing resource concentration. We further explore the emergent spatial patterns with a mix of weakly nonlinear stability analysis and numerical simulation, showing that diffusion-driven instability appears to increase cooperation and public goods across the spatial domain, while directed motion of cooperators towards regions with great public goods provision tends to decrease cooperation and environmental quality across the environment.

Finite element analysis of knee joint contact mechanics is computationally expensive, which has motivated the development of graph neural network surrogate models. However, effectively representing long-range dependencies in joint mechanical responses remains challenging. This study systematically compared topology diffusion, global routing, and their hybridization for surrogate modeling of knee joint contact mechanics. Using kinematic and force data from nine soccer players performing change-of-direction maneuvers, finite element simulations were used to generate graph-structured samples for training and evaluation under a grouped three-fold cross-subject evaluation framework. Five architectures were compared: standard MeshGraphNet, hierarchical MeshGraphNet, a routing-only transformer, a topology-biased routing transformer, and a hybrid model. The hybrid model achieved the best overall performance, yielding the lowest full-field error and peak stress error, together with the highest spatial agreement for high-risk regions. Among the non-hybrid models, the standard topology-diffusion model performed best overall, whereas routing-only strategies were less effective. These findings indicate that topology diffusion provides a robust basis for surrogate modeling of knee joint contact mechanics within the present benchmark, while the addition of global routing can further improve reconstruction of clinically relevant high-stress patterns.

We develop a unified spectral framework for finite ultrametric phylogenetic trees, grounding the analysis of phylogenetic structure in operator theory and stochastic dynamics in the finite setting. For a given finite ultrametric measure space $(X,d,m)$, we introduce the ultrametric Laplacian $L_X$ as the generator of a continuous time Markov chain with transition rate $q(x,y)=k(d(x,y))m(y)$. We establish its complete spectral theory, obtaining explicit closed-form eigenvalues and an eigenbasis supported on the clades of the tree. For phylogenetic applications, we associate to any ultrametric phylogenetic tree $\mathcal{T}$ a canonical operator $L_{\mathcal{T}}$, the ultrametric phylogenetic Laplacian, whose jump rates encode the temporal structure of evolutionary divergence. We show that the geometry and topology of the tree are explicitly encoded in the spectrum and eigenvectors of $L_{\mathcal{T}}$: eigenvalues aggregate branch lengths weighted by clade mass along ancestral paths, while eigenvectors are supported on the clades, with one eigenspace attached to each internal node. From this we derive three main contributions: a spectral reconstruction theorem with linear complexity $O(|X|)$; a rigorous geometric interpretation of the spectral gaps of $L_{\mathcal{T}}$ as detectors of distinct evolutionary modes, validated on an empirical primate phylogeny; an eigenmode decomposition of biological traits that resolves trait variance into contributions from individual splits of the phylogeny; and a closed-form centrality index for continuous-time Markov chains on ultrametric spaces, which we propose as a mathematically grounded measure of evolutionary distinctiveness. All results are exact and biologically interpretable, and are supported by numerical experiments on empirical primate data.

This study aims to better understand the functional geometry of the motor cortex, starting from different sources of experimental evidence. Recent studies have proved that cells of the primary motor cortex (M1) are sensitive to short hand trajectories called fragments. Here, we propose a sub-Riemannian higher-dimensional geometry accounting for geometric and kinematic properties. Due to the constraints of the geometry, horizontal curves naturally satisfy a relation between geometric and kinematic properties experimentally observed. In the space of trajectories, we also apply a clustering algorithm based on the Wasserstein distance: we obtain a grouping which nicely fits the observed experimental data much more efficiently than the Sobolev distance.

Many natural ecosystems harbor large numbers of coexisting species competing for far fewer distinct resources, in apparent defiance of the competitive exclusion principle. Various mechanisms have been proposed to explain this apparent paradox, among the most prominent being competition--colonization trade-offs, environmental heterogeneity, and ecological neutrality. We develop a unified stochastic model class that combines all three coexistence narratives in the context of propagule disperser communities and show that this setting encompasses several important classical models. We then prove a general theorem on coexistence at macroscopic equilibria and provide an algorithm that determines equilibrium coalitions solely from readily available matrix spectra, thereby bypassing the costly computation of exact equilibrium states. Using illustrative examples, we demonstrate the potential of this approach for quantifying the relative merits of different coexistence narratives and for studying their synergistic effects.

The present study evaluated the effectiveness of transcranial alternating current stimulation (tACS) treating patients with post-stroke anomia using a picture-naming task and a Single-Case Experimental Design (SCED). A right-handed 38-year-old woman with a left-hemisphere stroke and a left-handed 54-year-old man with a right-hemisphere stroke underwent an eight-week treatment program. Specifically, they participated in a picture-naming task three times a week, alternating between sessions with and without tACS stimulation every two weeks. Electroencephalography (EEG) measurements were taken at the end of each two-week period, and behavioral data were collected before, during and after the treatment. EEG and behavioral assessments were also conducted at one- and three-month follow-ups. Picture-naming performance was significantly faster during tACS sessions compared to sessions without tACS. By the end of the intervention, both participants demonstrated improved accuracy and speed, with positive effects also observed in behavioral measures. EEG analysis showed that post-treatment brain activity resembled that of healthy individuals performing similar tasks. Patients' improvements in picture-naming and behavioral tests showed that the positive effects remained stable even after three months. Thus, preliminary data suggest that tACS might be a promising intervention for anomia, with lasting effects. Large-scale studies are needed to confirm these findings.

Hypoglycemia is a severe condition of decreased blood glucose, specifically below 70 mg/dL (3.9 mmol/L). This condition can often be asymptomatic and challenging to predict in individuals with type 1 diabetes (T1D). Research on hypoglycemic prediction typically uses a combination of blood glucose readings and heart rate data to predict hypoglycemic events. Given that these features are collected through wearable sensors, they can sometimes have missing values, necessitating efficient imputation methods. This work makes significant contributions to the current state of the art by introducing two novel imputation techniques for imputing heart rate values over short-term horizons: Controlled Weighted Rational Bézier Curves (CRBC) and Controlled Piecewise Cubic Hermite Interpolating Polynomial with mapped peaks and valleys of Control Points (CMPV). In addition to these imputation methods, we employ two metrics to capture data patterns, alongside a combined metric that integrates the strengths of both individual metrics with RMSE scores for a comprehensive evaluation of the imputation techniques. According to our combined metric assessment, CMPV outperforms the alternatives with an average score of 0.33 across all time gaps, while CRBC follows with a score of 0.48. These findings clearly demonstrate the effectiveness of the proposed imputation methods in accurately filling in missing heart rate values. Moreover, this study facilitates the detection of abnormal physiological signals, enabling the implementation of early preventive measures for more accurate diagnosis.

With the rapid advancement of 6G, identity authentication has become increasingly critical for ensuring wireless security. The lightweight and keyless Physical Layer Authentication (PLA) is regarded as an instrumental security measure in addition to traditional cryptography-based authentication methods. However, existing PLA schemes often struggle to adapt to dynamic radio environments. To overcome this limitation, we propose the Adaptive PLA with Channel Extrapolation and Generative AI (APEG), designed to enhance authentication robustness in dynamic scenarios. Leveraging Generative AI (GAI), the framework adaptively generates Channel State Information (CSI) fingerprints, thereby improving the precision of identity verification. To refine CSI fingerprint generation, we propose the Collaborator-Cleaned Masked Denoising Diffusion Probabilistic Model (CCMDM), which incorporates collaborator-provided fingerprints as conditional inputs for channel extrapolation. Additionally, we develop the Cross-Attention Denoising Diffusion Probabilistic Model (CADM), employing a cross-attention mechanism to align multi-scale channel fingerprint features, further enhancing generation accuracy. Simulation results demonstrate the superiority of the APEG framework over existing time-sequence-based PLA schemes in authentication performance. Notably, CCMDM exhibits a significant advantage in convergence speed, while CADM, compared with model-free, time-series, and VAE-based methods, achieves superior accuracy in CSI fingerprint generation. The code is available at https://github.com/xiqicheng192-del/APEG

We consider a reconfigurable intelligent surface (RIS)-assisted heterogeneous network comprising legitimate information-harvesting receivers (IHRs) and untrusted energy-harvesting receivers (UEHRs). A multi-antenna base station (BS) transmits confidential information to IHRs while ensuring sufficient energy transfer to UEHRs that may attempt eavesdropping. To enhance physical-layer security, we propose a secure rate-splitting multiple access (RSMA) scheme aided by a UAV-mounted RIS. The objective is to maximize fairness-based secrecy energy efficiency (SEE). Owing to the non-convexity of the formulated problem, we develop an alternating optimization framework that jointly designs the common message allocation, active precoders, and RIS phase shifts under transmit power and energy harvesting constraints, leveraging sequential convex approximation (SCA). Simulation results demonstrate the scalability of the proposed algorithm and its superior SEE performance compared to space-division multiple access (SDMA) and non-orthogonal multiple access (NOMA) benchmarks.

Time series analysis plays a foundational role in a wide range of real-world applications, yet accurately modeling complex non-stationary signals remains a shared challenge across downstream tasks. Existing methods attempt to extract features directly from one-dimensional sequences, making it difficult to handle the widely observed dynamic phase drift and discrete quantization error. To address this issue, we decouple temporal evolution into macroscopic physical periods and microscopic phase perturbations, and inject frequency-domain priors derived from the Real Fast Fourier Transform (RFFT) into the underlying spatial sampling process. Based on this idea, we propose a Frequency-Guided Deformable Module (FGDM) to adaptively compensate for microscopic phase deviations. Built upon FGDM, we further develop an Adaptive Network based on Cascaded Harmonic Offset Routing (ANCHOR) as a general-purpose backbone for time-series modeling. Through orthogonal channel partitioning and a progressive residual architecture, ANCHOR efficiently decouples multi-scale harmonic features while substantially suppressing the computational redundancy of multi-branch networks. Extensive experiments demonstrate that ANCHOR achieves the best performance in most short-term forecasting sub-tasks and exhibits strong competitiveness on several specific sub-tasks in anomaly detection and time-series classification, validating its effectiveness as a universal time-series foundation backbone.

This paper studies wideband channel estimation for OFDM systems assisted by extremely large RIS (XL-RIS). Due to the large aperture of XL-RISs, the user equipment may operate in the near-field region, while the base station-XL-RIS link remains in the far field, leading to a cascaded channel with hybrid near-field and far-field characteristics. Moreover, wideband effects further complicate channel estimation in mmWave/THz systems. To address these challenges, we propose a frequency-independent orthogonal dictionary by augmenting the discrete Fourier transform (DFT) matrix with additional parameters, which enables an efficient representation of the wideband cascaded channel using a two-dimensional block-sparse structure. Based on this property, the considered channel estimation problem is effectively solved within a tailored compressed sensing framework. Simulation results demonstrate that the proposed method significantly outperforms conventional polar-domain channel estimation approaches in terms of estimation accuracy.

The frequency range around 7 GHz has emerged as a promising upper mid-band spectrum for 6th generation (6G), offering a practical balance between coverage and capacity. To fully exploit this band, however, future systems require substantially stronger beamforming and spatial multiplexing capability than today's 5G 64-port commercial deployments. This article investigates extreme multiple-input multiple-output (X-MIMO) with 256 digital ports as a practical 6G architecture for 7 GHz operation. First, through system-level simulations, we examine the throughput benefits and design trade-offs of increasing the number of base station (BS) and user equipment (UE) digital antenna ports, including comparisons between 128-port and 256-port configurations. We then present a 256-port 7 GHz BS and UE prototype and report field-trial results obtained in urban outdoor environments. The measurements demonstrate the feasibility of 8-layer downlink single-user MIMO over a 100 MHz bandwidth, achieving more than 3 Gbps for a single user under urban outdoor propagation conditions. Channel analysis based on measured data further suggests how the large digital aperture of X-MIMO supports high-order spatial multiplexing even with limited dominant angular clusters. Finally, we identify key challenges and outline research directions toward practical deployment of 7 GHz X-MIMO systems for 6G.

Battery state-of-health (SOH) reported by on-board battery management systems (BMS) is the primary metric available to electric vehicle (EV) owners and regulators, yet no study has validated its reliability across manufacturers against independent measurements. Here we show, through an epidemiological study of 1,114 EVs spanning five manufacturers and 375 days, that battery health reporting is fundamentally unreliable: real capacity differences of 12-25% exist within every model, but BMS SOH fails to track them, with correlations ranging from \r{ho} = 0.10 (non-significant) to \r{ho} = 0.62 only under restrictive filtering, while 384 vehicles do not expose SOH at all. A manufacturer-independent electrochemical marker achieves 74-89% degradation classification accuracy across all platforms without requiring BMS data, and a controlled laboratory validation on cells identical to those in the fleet confirms that partial-voltage-window charge measurements track reference capacity with \r{ho} > 0.80 across all 60 voltage windows (p < 0.001). These findings reveal a structural information asymmetry with direct implications for the EU Battery Regulation's 2027 SOH transparency mandate, California's Advanced Clean Cars (ACC) II durability requirements, warranty enforcement, used-vehicle valuation, right-to-repair legislation, and second-life battery markets.

Digital self-interference cancellation (D-SIC) plays a crucial role in in-band full-duplex radios. Unfortunately, its fundamental limit remains unclear. In this paper, we aim to address this problem by exploring the performance limit of the parallel Hammerstein (PH) canceller for D-SIC, which is most commonly used in practice. First, a comprehensive analysis of the power of the residual self-interference (RSI) after the PH canceller with the least squares (LS) estimator is provided, which takes into account the truncation error, reconstruction error and transmitter noise. Specifically, the analysis is greatly simplified by equivalently expanding the PH canceller via generalized Laguerre polynomials (GLP), which enjoys the desirable property of mutual orthogonality among the basis functions. As a by-product of this orthogonal expansion, we establish that the LS estimator for the weights of the GLP canceller is asymptotically \textit{unbiased}, if the pilot sequence is Gaussian distributed. Second, in order to minimize the reconstruction error of the PH canceller, we propose a succinct criterion for optimizing the pilot sequence, which essentially seeks for small eigenvalue spread and large minimum eigenvalue of the Gram matrix corresponding to the pilot sequence. Specifically, the criterion is to minimize the product of the Shannon rank, an effective rank of a positive semidefinite matrix and the minimum eigenvalue of the Gram matrix. Simulation results demonstrate that with the optimized pilot sequence of a single OFDM symbol, over 10 dB gain can be achieved compared to the conventional pilot sequence (HE-LTF) for the PH canceller, and the corresponding RSI can be as low as -87.6 dBm.

Against the backdrop of the global drive to advance the green transformation of the information and communications technology (ICT) industry and leverage technological innovation to facilitate the achievement of Net-Zero carbon goals, research into Rydberg atomic receivers (RAREs) is gaining significant interest. RAREs leverage the electron transition phenomenon for signal reception, offering significant advantages over conventional radio frequency receivers in terms of miniaturized antenna design, high sensitivity, robust interference resistance, and compact form factors, which positions them as a competitive alternative for meeting zero-carbon communication demands. This article systematically elaborates on the basic principle, state-of-the-art progress, and novel experiments of RAREs in quantum wireless communication and sensing. In this first-of-its-kind work, we experimentally verify the RARE-based orthogonal frequency division multiplexing transmission and reveal the potential of deep learning design in optimizing quantum wireless systems. Finally, we delve into the prospect of integrating RARE with existing cutting-edge application scenarios, while mapping out critical pathways for developing Rydberg-based wireless systems.

Radio map estimation (RME), which predicts wireless signal metrics at unmeasured locations from sparse measurements, has attracted growing attention as a key enabler of intelligent wireless networks. The majority of existing RME techniques employ grid-based strategies to process sparse measurements, where the pursuit of accuracy results in significant computational inefficiency and inflexibility for off-grid prediction. In contrast, grid-free approaches directly exploit coordinate features to capture location-specific spatial dependencies, enabling signal prediction at arbitrary locations without relying on predefined grids. However, current grid-free approaches demand substantial preprocessing overhead for constructing the spatial representation, leading to high complexity and constrained adaptability. To address these limitations, this paper proposes a novel cross-attention grid-free based transformer model for RME. We introduce a lightweight spatial embedding module that incorporates environmental knowledge into high-dimensional feature construction. A cross-attention transformer then models the spatial correlation between target and measurement points. The simulation results demonstrate that our proposed method reduces RMSE by up to 6%, outperforming grid-based and gridfree baselines.

This letter examines the impact of oscillator phase noise on sub-terahertz OFDM transceiver architectures, with a focus on the comparison between homodyne and heterodyne designs. Using a Hexa-X compliant phase noise model, we analytically show that heterodyne architectures reduce the total accumulated phase noise variance by distributing frequency translation across lower-frequency oscillators under realistic phase-noise scaling laws, thereby shifting the dominant impairment from inter-carrier interference to common phase error. OFDM simulations at 70 GHz and 140 GHz demonstrate that while homodyne architectures remain competitive at mmWave frequencies, heterodyne designs provide improved robustness to phase noise at higher sub-THz carriers. These results highlight transceiver architecture as a key design lever for relaxing oscillator and phase-locked loop constraints in future sub-THz wireless systems.

Channel capacity lies at the core of wireless communication, yet determining it typically requires detailed channel information between the transmitter and receiver. For near field MIMO systems, obtaining the detailed native channel is often difficult or expensive. This paper develops a scheme to approximate the near field channel in a Gaussian beam domain. Hermite Gaussian (HG) modes are used to approximate the channel between a pair of square antenna arrays in a free space line of sight (LOS) environment. We show that HG modes efficiently represent the dominant singular modes of the native channel, enabling accurate channel estimation and capacity computation in the HG beam space. An iterative algorithm is proposed to approach the maximal channel capacity by gradually expanding the beam space dimension. Simulation results demonstrate that the method converges rapidly and significantly reduces channel estimation overhead.

LLM agents are emerging as a key enabler for autonomous wireless network management. Reliably deploying them, however, demands benchmarks that reflect real engineering risk. Existing wireless benchmarks evaluate single isolated capabilities and treat all errors uniformly, missing both cascaded-chain failures and catastrophic unit confusions (\textit{e.g.}, dB vs.\ dBm). We present \wb{}, the first tolerance-aware, tool-integrated benchmark for LLM-based wireless agents. \wb{} is organized as a three-tier cognitive hierarchy: domain knowledge reasoning (WCHW, 1{,}392 items), intent-driven resource allocation (WCNS, 1{,}000 items), and proactive multi-step decisions under mobility (WCMSA, 1{,}000 items). Moreover, \wb{} is established on three design principles: \emph{(i)}~tolerance-aware scoring with catastrophic-error detection; \emph{(ii)}~tool-necessary tasks requiring a 3GPP-compliant ray-tracing query for channel quality; and \emph{(iii)}~Chain-of-Thought (CoT)-traceable items, where every benchmark item ships with a complete CoT trajectory enabling fine-grained diagnosis of where in the reasoning chain an agent fails. Our numerical results show that the direct-prompting model (GPT-4o) scores $68\%$, trailing a tool-integrated agent ($84.64\%$) by $16.64$\,pp; $23\%$ of errors are catastrophic failures invisible to exact-match metrics. More importantly, the hierarchy decomposes errors into four actionable diagnostic categories that flat evaluation cannot reveal. Code and data: https://wirelessbench.github.io/.

The robustness of angle of arrival (AoA) as a physical layer authentication (PLA) feature under spoofing attacks is studied, assuming a digital uniform linear array verifier. The verifier estimates the AoA assuming a legitimate user's single source model, whereas the received signal is generated by a multi antenna adversary at a different angle, leading to a model mismatch. Closed form expressions are derived for the misspecified Cramer Rao bound, the PLA decision threshold, the spoofing detection, false alarm and misdetection probabilities. Simulation results validate the theoretical findings and highlight the impact of the signal to noise ratio, array geometries, spoofing precoding and number of snapshots on authentication robustness.

Information-bearing reconfigurable intelligent surfaces (IB-RIS) provide a promising solution to self-sustainable and green communications by harvesting ambient radio frequency energy while embedding information via passive reflection. This paper investigates a self-sustainable IB-RIS (SIB-RIS)-assisted non-orthogonal multiple access (NOMA) network operating in an underlay cognitive radio (CR) system. Specifically, a multi-antenna primary transmitter (PT) serves a primary user (PU) and concurrently illuminates the secondary nodes, which enables each SIB-RIS to perform simultaneous energy harvesting and backscatter-based information embedding at each RIS. Based on this model, a weighted sum spectral efficiency (WSSE) maximization problem is formulated for the secondary network by jointly optimizing the PT transmit beamforming vector, the SIB-RIS reflection coefficients, and the power-splitting ratios. To tackle the intricately-coupled non-convex problem, an efficient block coordinate descent (BCD) optimization framework is developed, which leverages fractional programming via Lagrangian dual and quadratic transforms together with a difference-of-convex programming approach. Numerical results demonstrate that the proposed SIB-RIS-assisted NOMA CR system yields substantial WSSE gains over both orthogonal multiple access (OMA)-based and active antenna schemes. Moreover, a 2-bit discrete-phase SIB-RIS implementation achieves competitive to which WSSE performance, confirming the practicality of the low-resolution architecture.

We study a shift invariant space on an undirected graphs $G$ having $N$ vertices. We obtain a characterization theorem for a system of generalized translates $\{T_{i}g : 1\leq i\leq N\}$, for $g\in C^N$, to form an orthonormal basis. Moreover, we find a necessary and sufficient condition for the system $\{T_{i}g : 1\leq i\leq m\}$, $m\leq N$, to form a linearly independent set and an orthonormal set. Further, we obtain a characterization result for a system of generalized translates which is generated by multiple generators $g_{1},...,g_{M}$ to form a frame for $C^N$. In particular, we deduce similar results for the system $\{T_{i}M_{s}g : 1\leq i,s\leq N\}$ with modulation $M_{s}$ and the spectral graph wavelet system. We also provide an illustration for the spectral graph wavelet system.

We theoretically explain how multipath interference resizes PAM-4 constellation in the coherent regime and thus increases bit error rate fluctuation in intra data centers for the first time.

A unified beamforming and channel estimation framework relying on Bayesian learning is conceived. Recognizing the limitations imposed by low-resolution analog-to-digital converter (ADCs) and frequency-dependent propagation effects occurring in the Terahertz (THz) band, we formulate a dual-wideband channel model incorporating root raised cosine (RRC) pulse shaping. To address the non-linear distortions introduced by low-resolution ADCs, Bussgang decomposition is employed, leading to a tractable linearized inference process. By leveraging the shared sparsity inherent in a multi-user (MU) scenario of THz systems, we propose a Hierarchical Bayesian Group-sparse Regression (HBG-SR) based channel learning technique that exploits the group-sparse structure of THz band channels. The estimated dominant angle-of-arrival/ angle-of-departure (AoA/AoD) indices are then exploited for appropriately configuring the true-time-delay (TTD) elements in the hybrid transceiver, enabling precise beam alignment across subcarriers and the effective compensation of the beam-squint effect occurring in wideband THz systems. Extensive simulation results validate the efficiency of the proposed channel estimator and the TTD-aided beamforming architecture, highlighting their robustness and performance gains under practical wideband THz system constraints.

This paper investigates new efficient transmission architectures for multi-satellite massive multiple-input multiple-output (MIMO). We study the weighted sum-rate maximization problem in a multi-satellite system where multiple satellites transmit independent data streams to multi-antenna user terminals, thereby achieving higher throughput. We first adopt a multi-satellite weighted minimum mean square error (WMMSE) formulation under statistical channel state information (CSI), which yields closed-form updates for the precoding and receive vectors. To overcome the high complexity of optimization, we propose a learning-based WMMSE design that integrates tensor equivariance with closed-form recovery, enabling inference with near-optimal performance without iterative updates. Moreover, to reduce inter-satellite signaling overhead incurred by exchanging CSI and precoding vectors in centralized coordination, we develop a decentralized multi-satellite transmission scheme in which each satellite locally infers its precoders rather than receiving from the central satellite. The proposed decentralized scheme leverages periodically available satellite state information, such as orbital positions and satellite attitude, which is inherently accessible in satellite networks, and employs a dual-branch tensor-equivariant network to predict the precoders at each satellite locally. Numerical results demonstrate that the proposed multi-satellite transmission significantly outperforms single-satellite systems in sum rate; the decentralized scheme achieves sum-rate performance close to the centralized schemes while substantially reducing computational complexity and inter-satellite overhead; and the learning-based schemes exhibit strong robustness and scalability across different scenarios.

This paper develops a comprehensive analytical framework for the outage probability of fluid antenna system (FAS)-aided communications by modeling the antenna as a continuous aperture and approximating the Jakes (Bessel) spatial correlation with a Gaussian kernel $ρ_G(δ) = e^{-π^2δ^2}$. Three complementary analytical strategies are pursued. First, the Karhunen--Loève (KL) expansion under the Gaussian kernel is derived, yielding closed-form outage expressions for the rank-1 and rank-2 truncations and a Gauss--Hermite formula for arbitrary rank~$K$, with effective degrees of freedom $K_{\mathrm{eff}}^G \approx π\sqrt{2}\, W$. Second, rigorous two-sided outage bounds are established via Slepian's inequality and the Gaussian comparison theorem: by sandwiching the true correlation between equi-correlated models with $ρ_{\min}$ and $ρ_{\max}$, closed-form upper and lower bounds that avoid the optimistic bias of block-correlation models are obtained. Third, a continuous-aperture extreme value theory is developed using the Adler--Taylor expected Euler characteristic method and Piterbarg's theorem. The resulting outage expression $P_{\mathrm{out}} \approx 1 - e^{-x}(1 + π\sqrt{2}\, W\, x)$ depends only on the aperture~$W$ and threshold~$x$, is independent of the port count~$N$, and is identical for the Jakes and Gaussian models since both share the second spectral moment $λ_2 = 2π^2$. A Pickands-constant refinement for the deep-outage regime and a threshold-dependent effective diversity $N_{\mathrm{eff}} \approx 1 + π\sqrt{2}\, W\, x$ are further derived. Numerical results confirm that the Gaussian approximation incurs less than 10\% relative outage error for $W \leq 2$ and that the continuous-aperture formula converges with as few as $N \approx 10W$ ports.

Fluid antenna systems (FAS) achieve spatial diversity by dynamically switching among $N$ densely packed ports, but the resulting spatially correlated Rayleigh channels render exact outage analysis intractable. Existing block-correlation models (BCM) impose structural approximations on the channel covariance matrix that can introduce optimistic performance bias. This paper proposes a principled Karhunen-Loève (KL) expansion framework that decomposes the $N$-dimensional correlated FAS channel into independent eigenmodes and performs a controlled rank-$K$ truncation, reducing the outage analysis to a $K$-dimensional integration with $K \ll N$. Closed-form outage expressions are derived for the rank-1 and rank-2 cases, and a general Gauss-Hermite quadrature formula is provided for arbitrary $K$. On the theoretical front, it is proved via Anderson's inequality that the KL approximation \emph{always} overestimates the outage probability, providing a conservative guarantee essential for secure system design. Leveraging the Slepian--Landau--Pollak concentration theorem, it is established that only $K^* = 2\lceil W \rceil + 1$ eigenmodes are needed regardless of $N$, where $W$ is the normalized aperture. It is further shown that the KL truncation achieves the Gaussian rate-distortion bound, certifying it as the information-theoretically optimal channel compression. Extensive numerical results confirm that (i) theoretical predictions match Monte Carlo simulations, (ii) the entropy fraction converges faster than the power fraction, (iii) the KL framework uniformly outperforms BCM in approximation accuracy while avoiding the optimistic bias inherent in block-diagonal models, and (iv) the effective degrees of freedom scale with the aperture rather than the number of ports.

Due to the directive property of each antenna element, the received signal power can be severely attenuated when the emitter deviates from the array boresight, which will lead to a severe degradation in sensing performance along the corresponding direction. Although existing rotatable array sensing methods such as recursive rotation (RR-Root-MUSIC) can mitigate this issue by iteratively rotating and sensing, several mechanical rotations and repeated eigendecomposition operations are required to yield a high computational complexity and low time-efficiency. To address this problem, a pre-rotation initialization with recieve power as a rule is proposed to signifcantly reduce the computational complexity and improve the time-efficiency. Using this idea, a low-complexity enhanced direction-sensing framework with pre-rotation initialization and iterative greedy spatial-spectrum search (PRI-IGSS) is develped with three stages: (1) the normal vector of array is rotated to a set of candidates to find the opimal direction with the maximum sensing energy with the corresponding DOA value computed by the Root-MUSIC algorithm; (2) the array is mechanically rotated to the initial estimated direction and kept fixed; (3) an iterative greedy spatial-spectrum search or recieving beamforming method, moviated by reinforcement learning, is designed with a reduced search range and making a summation of all previous sampling variance matrices and the current one is adopted to provide an increasiong performance gain as the iteration process continues. To assess the performance of the proposed method, the corresponding CRLB is derived with a simplified rotation model. Simulation results demonstrate that the proposed PRI-IGSS method performs much better than RR-Root-MUSIC and achieves the CRLB in term of mean squared error due to the fact there is no sample accumulation for the latter.

Extremely Large Antenna Arrays (ELAA) operating at sub-terahertz frequencies introduce a regime where near-field Fresnel propagation and high-mobility carrier Doppler interact simultaneously, creating a four-dimensional signal space that existing schemes exploit only partially. This paper proposes \textbf{4D Fresnel Space-Time Modulation (4D-FSM)}, a unified framework encoding information jointly across angle, depth, synthetic velocity, and QAM amplitude through a structured symbol manifold $\mathcal{S}$. Synthetic velocity is introduced via Space-Time Modulation (STM): a linear phase ramp $u(ξ,t) = \exp(j[Ωt + g_kξ])$ induces a Doppler-equivalent shift without physical motion, creating velocity-orthogonal bubbles that resolve co-located users. We derive the joint orthogonality surface governing simultaneous user separability in depth and velocity, revealing that users separated in depth require strictly less velocity separation to remain orthogonal -- a multiplexing gain with no counterpart in OTFS or LDMA. The Discrete Fresnel Transform (DFnT) factorization $\mathbf{H} = \mathbf{F}_D \mathbf{C}(z) \mathbf{P}$ reduces precoder complexity from $\mathcal{O}(N^3)$ to $\mathcal{O}(N\log N)$, completing within \SI{500}{\nano\second} against a \SI{5.4}{\micro\second} coherence window. Monte Carlo evaluation at $f_c = \SI{140}{\giga\hertz}$, $N = 4096$ confirms $ρ\approx 0.998$ across the full velocity range, \SI{6.16}{\bit\per\second\per\hertz} spectral efficiency where all baselines collapse, and $K_{\max} = 64$ orthogonal users -- a $248\times$ sum-rate advantage over TTD at $K = 50$.

Current bias evaluations in Instruction Text-to-Speech (ITTS) often rely on univariate testing, overlooking the compositional structure of social cues. In this work, we investigate gender bias by modeling prompts as combinations of Social Status, Career stereotypes, and Persona descriptors. Analyzing open-source ITTS models, we uncover systematic interaction effects where social dimensions modulate one another, creating complex bias patterns missed by univariate baselines. Crucially, our findings indicate that these biases extend beyond surface-level artifacts, demonstrating strong associations with the semantic priors of pre-trained text encoders and the skewed distributions inherent in training data. We further demonstrate that generic diversity prompting is insufficient to override these entrenched patterns, underscoring the need for compositional analysis to diagnose latent risks in generative speech.

We present a statistical analysis of a variant of the periodogram method that forms power spectral density estimates by cross-correlating the discrete Fourier transforms of adjacent time windows. The proposed estimator is closely related to cross-power spectral methods and to a technique introduced by Nelson, which has been observed empirically to improve detection of sinusoidal components in noise. We show that, under a white Gaussian noise model, the expected contribution of noise to the proposed estimator is zero and that the estimator is unbiased under certain window alignment conditions. This contrasts with classical estimators where averaging reduces variance but not expected noise. Moreover, we derive closed-form expressions for the variance and prove an upper bound on the expected magnitude of the estimator that decreases as the number of windows increases. This establishes that the proposed method achieves a noise floor that decays with averaging, unlike standard nonparametric spectral estimators. We further analyze the effect of taking the absolute value to enforce nonnegativity, providing bounds on the resulting bias, and show that this bias also decreases with the number of windows. Theoretical results are validated through numerical simulations. We demonstrate the potential sensitivity to phase misalignment and methods of realignment. We also provide empirical evidence that the estimator is robust to other types of noise.

This paper investigates narrowband coordinated user scheduling in multi-cell massive multiple-input multiple-output (MIMO) systems. We formulate the problem under a spectral-efficiency maximization criterion, revealing inherent challenges in computational complexity and signaling overhead. To address these, we develop a user-scheduling-oriented CKM (US-CKM) and a US-CKM-driven two-stage coordinated scheduling framework. By exploiting the mapping between location information and statistical channel state information (SCSI), the system enables rapid SCSI retrieval and persistent reuse, substantially reducing CSI acquisition overhead. Embedding statistical channel correlation into the CKM further characterizes interuser interference patterns. The framework designs an intra-cell active-user selection scheme for the first stage and an inter-cell coordinated scheduling scheme for the second, both based on US-CKM entries. The first stage identifies users with favorable channel gains and low intra-cell interference, reducing the candidate set with marginal sum-rate loss. The second stage suppresses inter-cell interference (ICI) by exploiting cross-cell channel correlations. To enhance robustness against imperfect SCSI in dynamic scattering environments, we augment the framework with a reliability-guided mechanism. Instead of uniform treatment, we evaluate entry stability using a grid reliability metric quantifying channel measurement variance at sampling locations. Low-reliability grids are identified, and their instantaneous CSI is acquired in real time to integrate with existing SCSI. This process refines channel gain and spatial correlation characteristics, ensuring robust performance under imperfect conditions.

This paper investigates secure Directional Modulation (DM) design enhanced by a rotatable active Reconfigurable Intelligent Surface (RIS). In conventional RIS-assisted DM networks, the security performance gain is limited due to the multiplicative path loss introduced by the RIS reflection path. To address this challenge, a Secrecy Rate (SR) maximization problem is formulated, subject to constraints including the eavesdropper's Direction Of Arrival (DOA) estimation performance, transmit power, rotatable range, and maximum reflection amplitude of the RIS elements. To solve this non-convex optimization problem, three algorithms are proposed: a multi-stream null-space projection and leakage-based method, an enhanced leakage-based method, and an optimization scheme based on the Distributed Soft Actor-Critic with Three refinements (DSAC-T). Simulation results validate the effectiveness of the proposed algorithms. A performance trade-off is observed between eavesdropper's DOA estimation accuracy and the achievable SR. The security enhancement provided by the RIS is more significant in systems equipped with a small number of antennas. By optimizing the orientation of the RIS, a 52.6\% improvement in SR performance can be achieved.

We study the problem of implementing a fully-connected layer of a neural network using wireless over-the-air computing. We assume a multi hop system with a multi-antenna transmitter and receiver, along with a number of multi-hop amplify-and-forward relay devices in between. We formulate an optimization problem that optimizes the transmitter precoder, receiver combiner and amplify-and-forward gains, subject to relay device power constraint and transmitter power constraint. We propose an alternating optimization framework that optimizes the imitation accuracy. Simulation study results reveal that multi-hop relaying achieves an almost perfect classification accuracy when used in a neural network.

Covariance intersection (CI) methods provide a principled approach to fusing estimates with unknown cross-correlations by minimizing a worst-case measure of uncertainty that is consistent with the available information. This paper introduces a generalized CI framework, called overlapping covariance intersection (OCI), which unifies several existing CI formulations within a single optimization-based framework. This unification enables the characterization of family-optimal solutions for multiple CI variants, including standard CI and split covariance intersection (SCI), as solutions to a semidefinite program, for which efficient off-the-shelf solvers are available. When specialized to the corresponding settings, the proposed family-optimal solutions recover the state-of-the-art family-optimal solutions previously reported for CI and SCI. The resulting formulation facilitates the systematic design and real-time implementation of CI-based fusion methods in large-scale distributed estimation problems, such as cooperative localization.

Local constraint ordered statistics decoding (LC-OSD) provides strong soft decision performance for short block length linear codes, but its practical cost is dominated by the number of tested error patterns (TEPs). This paper proposes a neural early stopping (NES) protocol for LC-OSD with explicit cost control through one trade-off parameter balancing frame error risk and search effort. The proposed approach is trained with frame error rate (FER)-aligned supervision at predefined checkpoints, and learns if additional search is still likely to improve the current best candidate. Later, stopping is decided by comparing predicted continuation need with a cost measured in TEPs. Experimental results across multiple code families show that the proposed protocol significantly reduces average TEP count with only marginal FER degradation, using a single global model for the range of all operating signal-to-noise ratios (SNRs).

A Scene, represented visually using different formats such as RGB-D, LiDAR scan, keypoints, rectangular, spherical, multi-views, etc., contains information implicitly embedded relevant to applications such as scene indexing, vision-based navigation. Thus, these representations may not be efficient for such applications. This paper proposes a novel 360° saliency graph representation of the scenes. This rich representation explicitly encodes the relevant visual, contextual, semantic, and geometric information of the scene as nodes, edges, edge weights, and angular position in the 360° graph. Also, this representation is robust against scene view change and addresses challenges of indoor environments such as varied illumination, occlusions, and shadows as in the case of existing traditional methods. We have utilized this rich and efficient representation for vision-based navigation and compared it with existing navigation methods using 360° scenes. However, these existing methods suffer from limitations of poor scene representation, lacking scene-specific information. This work utilizes the proposed representation first to localize the query scene in the given topological map, and then facilitate 2D navigation by estimating the next required movement directions towards the target destination in the topological map by using the embedded geometric information in the 360° saliency graph. Experimental results demonstrate the efficacy of the proposed 360° saliency graph representation in enhancing both scene localization and vision-based indoor navigation.

Multipaths, reflections, diffractions, and material interactions complicate indoor wireless propagation modelling. More than 80% of wireless data is consumed indoors; hence, planning successful deployments and maximizing network performance depends on accurate propagation modelling of indoor environments. This work explains a complete framework for indoor wireless propagation modelling via ray tracing simulation in a step-by-step manner. The ray tracing simulations are conducted with Wireless Insite, a proprietary electromagnetic propagation software, whereas SketchUp is used at the input side for layout construction from the field measurements, and MATLAB is used at the output side for portraying channel model parameters such as power delay profile (PDP). A whole floor of the authors' department is modelled, and different transmitter-receiver locations were tested for possible use cases such as coverage hole prediction.

We investigate the behavior of scalar quantum fields in cosmological backgrounds under modified dispersion relations, specifically focusing on how ultraviolet asymptotics influence field quantization. We establish the conditions for both the validity of the adiabatic approximation and the unitary equivalence between quantizations defined via different time variables. Our analysis reveals that while superluminal modified dispersion relations consistently yield unitarily equivalent quantizations, asymptotically subluminal behaviors can lead to inequivalent physical descriptions. By applying adiabatic regularization to the two-point correlation function, we demonstrate that the ultraviolet scaling of the frequency uniquely dictates the required subtraction order. These results are illustrated through applications to standard, superluminal Corley--Jacobson, and Unruh dispersion relations.

Theoretical analysis of a prototypical two-qubit effective non-Hermitian system characterized by asymmetric Heisenberg $XY$ interactions in the absence of external magnetic fields demonstrates that maximal bipartite entanglement and quantum phase transitions can be induced exclusively through non-Hermiticity. At thermal equilibrium as $T\rightarrow 0$, the system attains maximal entanglement ${C}=1$ for values of the non-Hermiticity parameter greater than a critical value $γ>γ_c=J\sqrt{(1-δ^2)}$, where $J$ denotes the exchange interaction and $δ$ represents the anisotropy of the system; conversely, for $γ< γ_c$, entanglement is nonmaximal and given by ${C} = \sqrt{(1 - (γ/J)^2)}$. The entanglement undergoes a discontinuous transition to zero precisely at $γ= γ_c$. This phase transition originates from the closing of the energy gap at a non-Hermiticity-driven ground state degeneracy, which is fundamentally different from an exceptional point. This work suggests the use of singular-value-decomposition generalized density matrix for the computation of entanglement in bi-orthogonal systems.

Inspired by the MacDowell-Mansouri formulation of four-dimensional General Relativity, we study a class of four-dimensional gauge-theoretic functionals obtained from the Pontryagin density of a G-connection by inserting, under the trace, a matrix that breaks the gauge group G to a subgroup H. Concretely, we study the model with the pair (G,H) given by (SU(3), U(2)). We show that the critical points of the resulting functional are constant scalar curvature almost-Kahler 4-manifolds. On compact 4-manifolds, a stronger conclusion holds under the additional assumption that the scalar curvature is non-negative and the first Chern class is such that an Einstein metric can exist. In this case, results in the literature imply that the critical points are Kahler-Einstein 4-manifolds.

We investigate $f\left( Q\right) $-gravity with a matter-gravity coupling as a geometric dark energy candidate for the description of the late-time cosmic acceleration within a spatially flat Friedmann--Lemaître-Robertson-Walker geometry. We select a noncoincident connection that naturally follows from the general framework of cosmological models with nonzero spatial curvature. We present observational constraints for the simplest $f\left( Q\right) =f_{0}Q^{n}$ model using data from Supernovae, Baryon Acoustic Oscillations and Cosmic Chronometers. For different data combinations we found consistent constraints, with a best-fit value for the power-law index $n\simeq2$. A comparison with the $Λ$CDM model shows that the $f\left( Q\right) $-gravity leads to larger values for the likelihood, while Akaike's Information Criterion suggests statistical equivalence between the two models for most data combinations.

We introduce a classical estimator for the post-processing of quantum phase estimation data generated either by quantum-Fourier-transform-based or quantum-signal-processing-based methods. We focus on the estimation of a single target phase promised to be within an interval where no other phases are present, which is typical of e.g. ground state energy estimation of gapped quantum systems. This allows us to perform phase estimation by filtering the signal within the promise region and recovering the phase through a moment-projection estimator. We show that our methods are robust in the presence of both additional phases outside the promise region and global depolarizing noise. In the noiseless case our estimator can achieve an exponential suppression of bias with respect to a naive mean estimator. In the presence of global depolarizing noise our estimator achieves a bias exponentially small in the circuit depth $t$ at fixed circuit fidelity $F$, and a variance proportional to $t^{-2}$, improving by a factor of $t^2$ over the naive shifted-and-rescaled-mean approach. To mitigate realistic circuit-level noise, we combine our method with the explicit unbiasing scheme described in [Dutkiewicz et al., 2025]. As an illustrative example, we implement these estimators on a small-scale simulation of the Ising model, validating our theoretical results and finding better-than-expected performance for a global depolarizing noise approximation. The robustness of the moment-projection estimator in the presence of both multiple eigenvalues and realistic noise makes phase estimation with limited depth practical for early fault tolerant quantum experiments.

The canonically consistent quantum master equation (CCQME) method to treat system-bath dynamics is used to describe intramolecular proton transfer in the thioacetylacetone molecule (TAA, C$_4$H$_6$OS), modeled as an $N$-level quantum system coupled to a solvent. The solvent is represented as a harmonic bath (a continuum of oscillators) characterized by an Ohmic-Drude spectral density. We benchmark CCQME against numerically exact hierarchical equations of motion (HEOM) theory and compare to Redfield theory. Our results reveal that Redfield dynamics deviates increasingly from the HEOM reference as the system-bath coupling strength grows. In contrast, the recently proposed CCQME remains consistent with HEOM at intermediate coupling.

We study entanglement degradation near the horizons of regular, Reissner-Nordström, and Schwarzschild-de Sitter black holes, considering the Bardeen, Hayward, and generalized Hayward metrics as regular black holes. To this end, we compute the entanglement negativity, $\mathcal{N}$, for two Unruh-like modes of a scalar field shared by Alice, who is inertial, and Rob, who hovers at a fractional offset $ρ$ outside the horizon of the backgrounds under consideration. For each geometry, we locally approximate the metric by a Rindler patch characterized by Rob's proper acceleration $a_0$. Because this Rindler approximation breaks down near the extremal limit, we also compute a near-extremal cutoff. Tracing over the inaccessible Rindler wedge yields a mixed Alice-Rob state, from which we evaluate $\mathcal{N}$ as a function of the mode frequency $ω$ and the acceleration $a_0$. In all geometries considered, except for one, $\mathcal{N}$ increases monotonically with the parameter distinguishing that geometry form the Schwarzschild one. The exception is the Reissner-Nordström metric, for which $\mathcal{N}$ exhibits a shallow local minimum at a particular value of the charge. We also find that the Reissner-Nordström metric is the only background for which the negativity falls below that of the Schwarzschild case. Among all cases studied, the Schwarzschild-de Sitter spacetime provides the strongest protection of entanglement. Finally, across all backgrounds, high-frequency modes undergo less degradation than low-frequency modes. These results suggest that entanglement may serve as a useful probe for distinguishing Schwarzschild spacetime from other geometries.

Satellite based quantum key distribution (QKD) aims to establish secure key exchange over long distances despite significant technological challenges. To alleviate some of these challenges, Ghalaii et al. [PRX Quantum 4, 040320 (2023)] proposed that any airborne eavesdropper up to a certain size can be detected by classical monitoring techniques, limiting the transmission efficiencies of any undetected Eve. This creates a new QKD scenario in which some of the transmitted signal from Alice to Bob bypasses Eve entirely. In this manuscript, we develop a general framework for computing key rates in this "bypass" scenario for discrete variable protocols. We first numerically support a conjecture that the performance of BB84 with single photons does not improve under bypass constraints, and go on to find new regimes that do. Specifically, we find improvements when the receiver's detectors have an efficiency mismatch and when BB84 is implemented using weak coherent pulses under certain squashing assumptions. Technically, our framework is realized by including marginal constraints on the source to account for bypass effects, combined with existing numerical approaches for minimizing the key rate and squashing and dimension reduction techniques to handle photonic states of unbounded dimension.

We study the Dirac equation in 3+1 dimensions with a general combination of scalar, vector and tensor interactions with arbitrary strengths, all of them described by central Coulomb potentials acting on a particular plane of motion. For the tensor coupling a constant term is also included, since this gives rise to an effective Coulomb potential, which is necessary for the formation of bound states in a pure tensor coupling configuration. The exact bound-state solutions for this generalized Coulomb problem are computed by exploiting the freedom in choosing the coefficients of the \textit{Ansätze} for the radial functions, which leads to wave functions in terms of generalized Laguerre polynomials. From the quantization condition, the exact energy spectrum is also determined and its dependence on the parameters of the potentials is discussed. We show that similar features of the equations for the problem in the plane and the spherically symmetric problem allow a simple and direct mapping between them, thereby providing the solution to the spherical Coulomb problem. Our results are validated by showing that the solutions correctly encompass several previous solutions available in the literature for particular cases of this problem, for which we further develop the analysis of the parameters. We also derive two new particular cases not yet reported in the literature: the case of breaking of spin and pseudospin symmetries by the addition of a Coulomb plus constant tensor potential and the problem of a scalar plus tensor Coulomb potentials.

Integrating the scalability of individually addressable arrays of optical-tweezer-trapped single atoms with the efficient light-matter interface provided by nanophotonic waveguides has been a long-standing challenge in quantum technologies based on atoms and photons. Here we realize a quantum interface between photons guided in an optical nanofiber with a diameter of 310 nm and an array of on average 155 individually addressable atoms. Using a spatial light modulator and an objective lens with NA = 0.45, single cesium atoms are trapped in a one-dimensional array of 200 optical tweezer spots with micrometer-scale trap sizes on the nanofiber. Individual atoms are addressed by spatially scanning an excitation laser beam, focused to a spot size comparable to that of the traps through the same objective lens, along the nanofiber. We confirm the single-atom nature of the individual trapping sites through photon-correlation measurements of the guided fluorescence, observing strong photon antibunching with $g^{(2)}(0) \approx 0.26$. We measure trap lifetimes of a few hundred milliseconds, with a maximum value of 460 ms, at an atom-surface separation of 670 nm without active cooling, representing an order-of-magnitude improvement over previous nanofiber traps. This platform opens a new regime for atom-photon interfaces, paving the way for scalable distributed quantum computing and quantum networks, as well as for the exploration of collective radiative effects in waveguide QED with individually addressable atoms.

We show that, given an evolving quantum system and the quasiprobability distribution generated by the spatiotemporal generalization of the Born rule in pseudo density-matrices (PDMs), this distribution deviates from the sequential measurements probability distribution, given by the Lüders von-Neumann distribution, if and only if the non-signaling in time (NSIT) is violated; equivalently, if and only if the macroscopic realism (MR) is violated. Furthermore, we propose a definition of temporal entanglement according to the structure of the PDMs that is analogous to the definition of spatial entanglement in density matrices, showing that temporal entanglement is necessary for the violation of temporal Bell inequalities and the violation of MR. We employ our results to study the relationship between the negativity of the PDM, temporal entanglement, violation of temporal Bell inequalities, and MR.

We investigate the influence of boundary terms in gravitational field theories, by considering that in the Einstein-Hilbert action the boundary can be described by a non-metric Weyl-type geometry. The gravitational action and the the field equations, are thus generalized to include new geometrical terms, coming from the non-metric nature of the boundary, and depending on the Weyl vector, and its covariant derivatives. The field equations obtained within this framework generalize the standard Einstein equations by including in their mathematical structure the Weyl vector, and its covariant derivatives. As an applications of the general formalism we investigate the cosmological evolution in a flat FLRW geometry. We obtain the generalized Friedmann equations, which contain extra terms depending on the Weyl vector and its derivatives, arising due to the presence of the Weylian boundary, and which describe an effective, time dependent dark energy. By imposing to the dark energy an equation of state parameter of the Barboza-Alcaniz type, the Friedmann equations can be solved numerically. We compare the predictions of the Weylian boundary gravitational theory with late-time observational data and the predictions of the $Λ$CDM paradigm. Our results show that the Weylian boundary cosmological models give a good description of the observational data, and they can reproduce almost exactly the predictions of the $Λ$CDM paradigm. Hence, the extension of gravitational theories through the addition of Weylian boundary terms, in which dark energy has a purely geometric origin, emerges as a viable alternative to standard general relativity.

In the absence of error correction, noisy intermediate-scale quantum devices are operated by training parametrized quantum circuits (PQCs) so as to minimize a suitable loss function. Finding the optimal parameters of those circuits is a hard optimization problem, where global guarantees are known only for highly structured cases of limited practical relevance, and first-order methods can fail to find even local minima due to the presence of barren plateaus. In this work, we study the training of practical classes of PQCs, namely polynomial-depth circuits with a constant number of trainable parameters. This captures widely used PQC families, including fixed-depth QAOA, hardware-efficient ansätze, and Fixed Parameter Count QAOA. Our main technical result is a fully polynomial randomized approximation scheme (FPRAS), which, for every $ε>0$, returns an $ε$-approximate solution to the problem's global optimum with high probability, and has runtime and query complexity polynomial in $1/ε$ and the number of qubits. Unlike the standard hybrid quantum-classical training loop in variational algorithms, where the quantum device is queried repeatedly throughout the training, our approach separates the computation into two distinct stages: (1) an initial quantum data-acquisition phase, followed by (2) a classical global-optimization phase based on the trigonometric moment/sum-of-squares hierarchies. Under a standard flat-extension condition, which can be checked numerically, the method also supports the extraction of optimal circuit parameters. The existence of an FPRAS implies that the promise problem associated with the optimization of poly-depth constant-parameter PQC is in BQP. This imposes a limitation on the expressive power of the class, namely, it cannot encode combinatorial optimization problems whose objective values are separated by an inverse-polynomial gap.

Quantum error correction (QEC) is critical for scalable fault-tolerant quantum computing. Topological codes, such as the toric code, offer hardware-efficient architectures but their Tanner graphs contain many girth-4 cycles that degrade the performance of belief-propagation (BP) decoding. For this reason, BP decoding is typically followed by a more complex second stage decoder such as minimum-weight perfect matching. These combined decoders achieve a remarkable performance, albeit at the cost of increased complexity. In this paper we propose two key improvements for the decoding of toric code. The first one is replacing the BP decoder by a neural BP decoder, giving rise to the neural belief-matching decoder which substantially decreases the average decoding complexity. The main drawback of this approach is the high cost associated with the training of the neural BP decoder. To address this issue, we impose a convolutional architecture on the neural BP decoder, enabling weight sharing across the spatially homogeneous structure of the code's factor graph. This design allows a model trained on a modest-size topological code to be directly transferred to much larger instances, preserving decoding quality while dramatically lowering the training burden. Our numerical experiments on toric-code lattices of various sizes demonstrate that this technique does not result in a noticeable loss in performance.

Reliable quantum memory is essential for scalable quantum networks and fault-tolerant photonic quantum computing. We present a quantitative analysis of an all-optical quantum memory architecture in which a Gottesman-Kitaev-Preskill (GKP) encoded qubit is stored in a fibre loop and periodically stabilized using teleportation-based error correction. By modelling fibre propagation as a pure-loss channel and representing each correction round as an effective logical map acting on the Bloch vector, we obtain a compact description of the full multi-round memory channel. We show that syndrome decoder optimization plays a crucial role in the experimentally relevant finite-squeezing regime. The optimal decoder deviates from standard square-grid GKP decoder in both tile-size and tile-shape, leading to significant improved logical performance. Using this optimized decoding strategy, we identify a squeezing-dependent optimal spacing between correction nodes that maximizes the memory lifetime. Remarkably, this optimal segment length is largely independent of the desired storage time, providing a simple and practical design rule for fibre-loop quantum memory. We further find a squeezing threshold of approximately 6.7 dB below which intermediate error correction becomes counterproductive, while above threshold the achievable storage time increases approximately exponentially with squeezing. For example, at 17 dB squeezing, storage times exceeding 400 ms can be achieved with logical infidelity below 1%. These results establish clear performance benchmarks and reveal the fundamental trade-off between photon loss, squeezing, and correction frequency in continuous-variable architectures. Our findings provide actionable design principles for near-term photonic quantum memory and clarify the path toward scalable all-optical fault-tolerant quantum storage.

High-fidelity control of quantum systems is essential for scalable quantum technologies. We introduce a shooting-based quantum optimal control algorithm for systems of finite-dimensional Hilbert spaces, and demonstrate its performances through numerical simulations on systems inspired from single molecule magnets. Our method efficiently decomposes quantum gates into selective electromagnetic pulses, outperforming the standard GRAPE method against which it is benchmarked, especially in larger Hilbert spaces

This note provides an intrinsic construction of the Carrollian superplane $Π\mathbb{S}\simeq \mathbb{R}^{2|4}$ as a supermanifold generalisation of the Carrollian plane. Moving away from the $c\rightarrow 0$ limit of relativistic spinors, we define Carroll spinors as sections of a degenerate Clifford module. We show that the Carrollian superplane is a principal $\mathbb{R}^{1|2}$-bundle. Once clock forms and a complementary basic one-form are specified, there is a pair of odd vector fields that generate novel $N =2$ Carrollian supersymmetry transformations, not all of which come from an Inönü--Wigner contraction of a Poincaré superalgebra

We investigate the correspondence between unimodular diffusion cosmology and interacting dark sector models at the background and linear perturbation levels. In the diffusion framework, the effective cosmological constant becomes time dependent, $Λ(t)$, sourced by a diffusion current. We show that at background level this framework can be mapped onto interacting dark energy models with $w=-1$ and energy transfer $Q$. Using two common parameterizations, $Q = ξH ρ_{\rm de}$ and $Q = ξH ρ_{\rm dm}$, and data from supernovae, DESI BAO, cosmic chronometers, and CMB distance priors, we find $ξ= -0.0197 \pm 0.0076$ for the vacuum coupled case, while the matter coupled case gives a best fit $ξ= 0.0018$ with comparable fit. At linear perturbations, the diffusion framework is perturbatively equivalent only to interacting vacuum models with homogeneous energy transfer ($Q \propto ρ_{\rm de}$, $δQ=0$). Including redshift space distortion data, we obtain $ξ= -0.0147 \pm 0.0075$, consistent with $Λ$CDM ($ξ=0$) at $2σ$. The inferred clustering amplitude is $S_8 = 0.782 \pm 0.026$ for the diffusion model, compared to $S_8 = 0.77 \pm 0.025$ for $Λ$CDM under the same dataset, showing a modest impact on structure growth.

We review existing classical simulation methods for performing fermionic Gaussian operations and develop new methods to address the gap by adhering to the fundamental theoretical framework established by Bravyi [Quantum Info. Comput. 5, 216 (2005)] for the most general fermionic Gaussian processes. Throughout this attempt, the focus remains on the unified approach that can be applied to generic fermionic Gaussian operations. This is beneficial since the selection of simulation methods has often been based on an ad hoc choice, heavily influenced by the specific model and circumstances, rather than on a systematic approach.

Neutral atom arrays have seen tremendous progress in quantum simulation, quantum metrology, and fault-tolerant quantum computing. However, hardware constraints such as atom loss and heating remain significant challenges. In this work, we introduce a comprehensive ancilla-based toolbox for optical tweezer experiments that utilizes high-fidelity Rydberg entangling gates and ancilla atoms to mitigate these physical limitations. First, we demonstrate repeated ancilla-based atom readout, achieving improved detection fidelity over multiple rounds with minimal perturbation to data atoms. Second, leveraging the quantized motional states in tweezer-trapped strontium atoms, we transduce quantum information from the electronic to the motional manifold. This enables us to perform mid-circuit ancilla-based atom loss detection in a coherence-preserving fashion. Finally, we demonstrate algorithmic cooling, a circuit-based sequence that deterministically cools data atoms by transferring their motional entropy to the electronic states of ancilla atoms. We observe a marked reduction in the atomic temperature of data atoms. These tools offer a pathway to continuous operation in tweezer clocks and complement recent developments in continuous reloading experiments.

The potential of quantum computing is fundamentally constrained by the inherent susceptibility of qubits to noise and crosstalk, particularly during multi-qubit gate operations. Existing strategies, such as hardware isolation and dynamical decoupling, face limitations in scalability, experimental feasibility, and robustness against complex noise sources. In this manuscript, we propose a physics-guided neural control (PGNC) framework to generate robust control pulses for superconducting transmon qubit systems, specifically targeting crosstalk mitigation. By combining a hardware aware parameterization with a Hamiltonian-informed objective that accounts for condition-dependent crosstalk distortions, PGNC steers the search toward smooth and physically realizable pulses while efficiently exploring high dimensional control landscapes. Numerical simulations for the CZ gate demonstrate superior fidelity and pulse smoothness compared to a Krotov baseline under matched constraints. Taken together, the results show consistent and practically meaningful improvements in both nominal and perturbed conditions, with pronounced gains in worst-case fidelity, supporting PGNC as a viable route to robust control on near-term transmon devices.

In continuous-variable quantum computing, autonomous quantum error correction (QEC) can dissipatively steer a noisy quantum state into a target state or manifold, enabling robust quantum information processing without explicit syndrome measurements and feedback. Here, we propose a nullifier-based digital autonomous QEC enabled by conditional Gaussian operations. By designing jump operators for target nullifiers and compiling the resulting Lindbladian into a Trotterized sequence of elementary conditional Gaussian operations, we demonstrate two use cases: (i) deterministic preparation of non-Gaussian resource states for universal computation, including finitely squeezed cubic phase states and approximate trisqueezed states, and (ii) autonomous suppression of dephasing error for cat and squeezed cat states. We provide explicit gate decompositions for the required conditional Gaussian operations and numerically evaluate the performance under realistic imperfections, including photon loss in the bosonic mode and ancillary-qubit decoherence. Our results clarify the resource requirements and trade-offs, such as circuit depth, time-step choices, and the required set of conditional Gaussian operations, for scalable, gate-level implementations of autonomous state preparation and error suppression.

Estimating thermal expectation values of quantum many-body systems is a central challenge in physics, chemistry, and materials science. Standard quantum Gibbs sampling protocols address this task by preparing the Gibbs state from scratch after every measurement, incurring a full mixing-time cost at each step. Recent advances in single-trajectory Gibbs sampling \cite{jiang2026} substantially reduce this overhead: once stationarity is reached, measurements can be collected along a single trajectory without re-thermalizing, provided the measurement channel preserves the Gibbs ensemble. However, explicit constructions of such non-destructive measurements have been limited primarily to observables that commute with the Hamiltonian. In this work, we fundamentally extend the single-trajectory framework to arbitrary, non-commuting observables. We provide two measurement constructions that extract measurement information without fully destroying the Gibbs state, thereby eliminating the need for full re-mixing between samples. First, we construct a measurement that satisfies exact detailed balance. This ensures the system remains in equilibrium throughout the trajectory, allowing measurement outcomes to decorrelate in an autocorrelation time that could be significantly shorter than the global mixing time. Second, assuming the underlying quantum Gibbs sampler has a positive spectral gap, we design a simplified measurement scheme that ensures the post-selected state serves as a warm start for rapid re-mixing. This approach successfully decouples the resampling cost from the global mixing time. Both measurement schemes admit efficient quantum circuit implementations, requiring only polylogarithmic Hamiltonian simulation time.

Quantum Neural Networks (QNNs) are a promising class of quantum machine learning models with potential quantum advantages when implemented on scalable, error-corrected quantum computers. However, as system sizes increase, deploying QNNs becomes challenging. Similar to their classical counterparts, a key obstacle to their practical applications is that large-scale QNNs may not be easily deployed on smaller systems that have limited resources. Here, we tackle this challenge by compressing QNNs via knowledge distillation. We demonstrate how well-trained QNNs on large systems can be distilled into smaller architectures with similar configurations. We numerically show that knowledge distillation helps reduce the training cost of QNNs in terms of the number of qubits and circuit depth. Additionally, we find that a self-knowledge-distillation approach can accelerate training convergence. We believe our results offer new strategies for the efficient compression and practical deployment of QNNs.

Fixed-frequency transmon qubits, characterized by simple architectures and long coherence times, are promising platforms for large-scale quantum computing. However, the rapidly increasing frequency collisions, which directly reduce the fabrication yield, hinder scaling, especially in cross-resonance (CR) gate-based architectures, wherein the restricted drive frequency severely limits the available design space. We investigate the Stark-induced ZZ by level excursions (siZZle) gate, which relaxes this limitation by allowing arbitrary drive-frequency choices. Extensive numerical analyses across a broad parameter range -- including the far-detuned regime that has received negligible prior attention -- reveal wide operating windows that support controlled-Z (CZ) fidelities >99.6%. Leveraging these windows, we design lattice architectures containing >1000 qubits, showing that even under 0.25% fabrication-induced frequency dispersion, the zero-collision yields in square and heavy-hexagonal lattices reach 80% and 100%, respectively. Thus, the siZZle-CZ gate is a scalable and collision-robust alternative to the CR gate, offering a viable route toward high-yield fixed-frequency transmon quantum processors.

We consider hidden symmetries arising from U-duality in the dimensional reduction of non-maximal higher-derivative supergravities to three dimensions. In particular, we consider the $G_{2(2)}$ symmetry of minimal five-dimensional supergravity and the $O(d+p+1,d+1)$ symmetry of bosonic and heterotic string theory on $T^d$. Using a group theory argument, we show that the higher-derivative corrections explicitly break all hidden symmetry enhancements. As special cases, this also implies that higher-derivative corrections prevent the symmetry enhancement to $SL(3,\mathbb R)$ in pure five-dimensional gravity and $O(4,4)$ in the STU model.

Grinberg-Maldacena showed that, for AdS/CFT, the thermal one-point function of a heavy boundary operator in the dual conformal field theory encodes the complex geodesic length from the boundary insertion to the black hole singularity. We show that for an appropriate choice of bulk-boundary kernel, an analogous result holds for three-dimensional Schwarzschild-de Sitter black hole. We prove the result for the case where SdS$_3$ has a finite orbifold group.

Various stacking combinations of the two-dimensional (2D) boron nitride (BN) honeycomb lattice can significantly modify the properties of the resulting 2D BN crystal. Here, we demonstrate through first-principles calculations that the brightness of the negatively charged boron-vacancy center (V$_\text{B}^{-}$) is enhanced by at least one order of magnitude in rhombohedral BN (rBN) compared to hexagonal BN (hBN), while the spin properties remain either comparable or even improved. This enhancement arises from the reduced symmetry of the crystal field in rBN. Our results suggest that room-temperature single-spin coherent control of V$_\text{B}^{-}$ is feasible in rBN, enabling its application as a single-spin quantum sensor in this 2D host. These findings demonstrate that engineered stacking of BN layers provides a powerful means to tailor the properties of embedded quantum defects.

We consider a non-relativistic quantum particle in $\mathbb{R}^d$, $d=2$ or $d = 3$, interacting with singular zero-range potentials concentrated on a large collection of points. We analyze the homogenization regime where the intensities of the singular potentials and the distances between the points simultaneously go to zero as their number grows, while the total interaction strength remains finite. Assuming that the singular potentials have negative scattering lengths and are uniformly distributed, we prove the strong resolvent convergence as $N \to \infty$ of the family of operators to a Schrödinger operator with a regular electrostatic potential. The result is obtained via $Γ$-converge of the associated quadratic forms. Moreover, in presence of an external trapping potential, the convergence is lifted to uniform resolvent sense.

We investigate the rotational response of a confined, two-dimensional quantum droplet, which emerges in an attractive binary Bose mixture that is stabilized against collapse by beyond-mean-field effects. We consider both a harmonic and an anharmonic form for the external confining potential. We go beyond the widely employed ``phase-locked" single-order-parameter model, maintaining two separate order parameters for the two components, and calculating the lowest-energy state for various values of the angular momentum. For a population-balanced quantum droplet and sufficiently tight confinement, we find that near certain half-integer values of the angular momentum the droplet is excited in a ``heterosymmetric" manner, with the two components carrying different vorticities. This mode is naturally missed by the single-order-parameter model. We additionally investigate the effects of a small population imbalance in the droplet. Apart from an energy increase associated with the population difference, the imbalance also lifts the double degeneracy of the heterosymmetric states, which characterizes the $\mathbb{Z}_2$-symmetric balanced droplet. The heterosymmetric mode is found to be favored by the energy term which captures the beyond-mean-field effects in the mixture.

Multimode Gaussian states are a versatile resource for quantum information technologies and have been realized across a wide range of physical platforms. Recent progress in the large-scale generation of such states provides a key ingredient for scalable quantum technologies. Despite the importance of accurately characterizing these states, conventional tomography methods are often impractical because they require large sample sizes and can yield unphysical states. Here we present a reliable and efficient tomography method for multimode Gaussian states based on maximum-likelihood estimation. By directly operating on covariance matrices, the method avoids the exponential overhead associated with density-matrix reconstruction. We consider two commonly used detection schemes--single and joint homodyne detection--and systematically analyze the reconstruction performance. Our method outperforms conventional approaches by ensuring physical covariance matrices and achieving better agreement with the true states. To demonstrate the experimental applicability of the method, we experimentally generate various multipartite entangled states--six-mode graph states with different connectivity, a six-mode GHZ state, and a fully connected ten-mode graph state--and reconstruct their covariance matrices. Using the reconstructed covariance matrices, we quantify fidelities, detect entanglement, and reveal the multimode structure of squeezing and noise. Our technique offers a practical diagnostic tool for developing scalable quantum technologies.

Quantum reservoir computing (QRC) harnesses driven quantum dynamics for time-series processing, yet the mechanisms behind the differing performance levels across its many implementations remain unclear. We show that apparently unrelated approaches-including memory restriction, weak measurements, operation near the edge of quantum chaos, and dissipative dynamics-are in fact governed by the same underlying principle, namely a tunable balance between memory retention and nonlinear response. Using the information processing capacity, a dynamical measure from nonlinear systems theory, we place these behaviors in a unified framework and identify the regimes in which quantum reservoirs surpass the standard protocol. Our results reveal a fundamental connection between memory and nonlinear response. This provides a general design principle for enhanced information processing and enables systematic analysis and optimization inspired by classical dynamical quantifiers.

We investigate whether meridional screening of the kinematically irrotational Hawking-Ellis Type I warp-drive background can be accommodated within a restricted 6-variable reduced weakly birefringent deformation. The problem is posed as a perturbative admissibility question: whether the weakly birefringent area-metric closure framework of Schneider, Schüller, Stritzelberger, and Wolz can absorb the residual mixed-sector transport generated by that background while remaining within the locally hyperbolic weakly birefringent regime. Within the restricted ansatz class studied here, a naive rank-one uniaxial constitutive deformation fails as a kinematic screening mechanism. The minimal successful container is instead a symmetric meridional $3\times 3$ bivector block, whose sparse image in the 17-variable Schneider-Schüller-Stritzelberger-Wolz parametrization occupies six weakly birefringent variables. This yields a derived reduced working model coupled to the exact background transport system. Exploratory integration, using the irrotational profile together with a shear-based lift envelope, is reported at the benchmark angles $θ=0,π/4,π/2$. The mixed-sector tilt vanishes on the symmetry axis. In the representative coupled case $θ=π/4$, it is approximately cubic only for mild velocities $v\lesssim 1$; at higher velocity it departs strongly from that trend and the signed variable $φ_{17}$ changes sign between $v=2$ and $v=3$. At $θ=π/2$, by contrast, the benchmark values follow the decoupled reduced-model cubic law. The resulting evidence disfavors fast walls within the reduced perturbative model, while smaller subluminal velocities are less strained by the reduced admissibility conditions. Throughout, the reduced weakly birefringent model is treated as a derived exploratory specialization rather than a theorem-level no-go or existence proof.

A long-standing problem in the foundations of quantum mechanics is to identify a physical principle that explains why algebraically maximal violations of Bell inequalities can generally not be achieved in Nature. One recently proposed approach considers iterated Bell tests, where a Bell test is performed on a state that has undergone several rounds of entanglement swapping. Obtaining large violations in this scenario is more demanding, because it requires a theory to have both highly entangled states and highly entangled measurements. It has been conjectured that the maximal quantum mechanical Clauser-Horne-Shimony-Holt (CHSH)-value of $2\sqrt2$ might be optimal for any probabilistic theory which, like quantum mechanics, maintains its CHSH-value after an arbitrary number of rounds of entanglement swapping. However, in a previous paper, we have exhibited a first example of a probabilistic theory that can sustain a CHSH value of $4$ in this setting. In this work, further investigating this property, we give a classification of all general probabilistic theories (GPTs) whose CHSH value is stable in the above sense. The problem reduces to a representation-theoretic condition that allows for exactly seven solutions. The GPT from our previous work showed some counter-intuitive features, e.g. that the local state space had a higher dimension than seemed necessary to realize CHSH tests. The classification shows that this is necessarily so. Along the way, we generalize the concept of self-testing to GPTs.

Advances in the study of relativistic quantum phase space have established the set of Linear Canonical Transformations (LCTs) as a candidate for the fundamental symmetry group associated with relativistic quantum physics. In this framework, for a spacetime of signature $(N_+,N_-)$, the symmetry of the relativistic quantum phase space is described by the LCT group, isomorphic to the symplectic Lie group $Sp(2N_+,2N_-)$, which preserves the canonical commutation relations (CCRs) and treats spacetime coordinates and momenta operators on an equal footing. In this work, we investigate the contraction structure of the Lie algebra associated with the LCT group for signature $(1,4)$, clarifying how familiar spacetime symmetry groups emerge from this more fundamental quantum phase space symmetry.Using the Inönü-Wigner group contraction formalism, we examine each limit case corresponding to the possible combinations of asymptotic values of two fundamental length scale parameters associated with the theory, namely a minimum length $\ell$ and a maximum length $L$, which may be identified respectively with the Planck length and the de Sitter radius. We explicitly analyze how contractions of the LCT Lie algebra lead to the physically relevant de Sitter algebra $\mathfrak{so}(1,4)$ and, in the flat-curvature limit, to the Poincaré algebra $\mathfrak{iso}(1,3)$ of four-dimensional spacetime. This provides an explicit mechanism through which relativistic spacetime symmetry can emerge from a deeper quantum symplectic structure of phase space.

We discuss the connection between logotropes and polytropes in astrophysics and cosmology. The logotropic equation of state $P=A\ln(ρ/ρ_P)$ may be seen as a degenerate form of the polytropic equation of state $P=Kρ^γ$ in the limit $γ\rightarrow 0$, $K\rightarrow\infty$ with $A=Kγ$ fixed. The logotropic distribution function corresponds to the polytropic distribution function of index $γ=0$ for which the density is finite but the pressure diverges logarithmically. We show that the polytropic and logotropic distribution functions can be obtained in the nondegenerate limit of the Lynden-Bell theory of violent relaxation for a particular distribution of phase levels given by the $χ$-squared distribution. This provides a justification of the Tsallis entropy from the Lynden-Bell entropy. The logotropic distribution function presents a power-law energy tail decreasing as $ε^{-5/2}$. Interestingly, this ``universal'' power-law tail is predicted by recent kinetic theories of collisionless relaxation based on the coarse-grained Vlasov equation and on the secular dressed diffusion equation. When coupled to gravity, the associated density profile decreases as $r^{-1}$. This may explain the universal surface density of dark matter halos, or account for an effective NFW density cusp. This also accounts for the universal gravitational acceleration felt by a test particle and for the Tully-Fisher relation. The logotropic model can thus provide an alternative to the modification of Newtonian dynamics (MOND) theory. We recall how the logotropic model leads to a very accurate expression of the cosmological constant $Λ={G^2m_e^6}/{α^6\hbar^4}=1.36\times 10^{-52}\, {\rm m^{-2}}$ in terms of the mass of the electron and the fundamental constants of physics.

We demonstrate precision laser spectroscopy of a trapped $^{173}$Yb$^+$ ion that is not directly laser cooled by coupling it to ultracold atoms. The atomic bath continuously cools the internal degrees of freedom of the ion to its hyperfine ground state via spin-exchange collisions. Successful laser excitation is detected via state-selective charge transfer and subsequent ion loss. We probe the $6^2S_{1/2}\rightarrow 6^2P_{3/2}$ transition at 329 nm and measure the magnetic and electric hyperfine interaction constants for the $6^2P_{3/2}$ state to be $A=-241(1)$ MHz and $B=1460(8)$ MHz, respectively. Our results are in agreement with a previous measurement obtained in a hollow-cathode discharge experiment but are a factor of 6-9 more precise. The techniques demonstrated in this work may be extended to perform precision spectroscopy on other ions with complex level structures.

We present a compact quantum encoding of the Traveling Salesperson Problem (TSP) based on a time-register representation of tours. A candidate route is represented as a sequence of $n$ city labels over discrete time steps, with one fixed start city and the remaining cities encoded in binary registers. We describe three ingredients of the construction: uniform route generation over the route register, a reversible oracle for marking valid tours, and a phase oracle that encodes the total tour cost. The validity oracle distinguishes permutations of the non-start cities from invalid assignments, while the cost oracle accumulates the contribution of the start edge, intermediate transitions, and return edge into a tour-dependent phase. This yields a coherent superposition of candidate routes with feasibility and tour-length information embedded directly in the quantum state. The number of qubits required is $\Order{n\log_2(n)}$ and the circuit depth scales quadratically in $n$. The encoding is compatible with amplitude amplification or spectral filtering techniques such as the quantum singular value transform (QSVT) or Grover's algorithm. However, due to the exponentially small fraction of valid tours, the overall complexity remains exponential even when combined with amplitude amplification.

The description of states and dynamics in non-Hermitian systems is fundamentally linked to the choice of an appropriate theoretical framework--a point of ongoing debate in the field. This work addresses this issue by proposing a consistent formulation that reconciles existing controversies and establishes a unified theoretical understanding. Our approach rests on a foundational premise: The dynamics of both left- and right-vectors of a non-Hermitian system must satisfy the Schrödinger equation. Building on this physically motivated assumption, we refine the biorthogonal framework, leading to a consistent reformulation of non-Hermitian quantum theory. This refined framework can naturally reduce to standard quantum mechanics in the Hermitian limit. As a concrete application, we analyze the dynamical phase transition in a one-dimensional Su-Schrieffer-Heeger (SSH) model within this refined framework. Notably, our formulation naturally generalizes the known condition for such transitions in Hermitian two-band systems, namely, $\mathbf{d}_{k}^i\cdot\mathbf{d}_{k}^f=0$, to the non-Hermitian case, where it takes the form $\mathrm{Re}\Bigl[\frac{\mathbf{d}_{k}^i}{d_{k}^i}\cdot\frac{\mathbf{d}_{k}^f}{d_{k}^f}\Bigr]=0$. Furthermore, we identify entirely new dynamical phase transitions that cannot be characterized by the winding number. We hope that this refined framework will find broad applications in the study of non-Hermitian systems.

While the quasi-local thermodynamics of spherically symmetric black holes is well described by pressure and volume, extending this framework to rotating spacetimes poses a significant challenge. Rotation induces an oblate deformation of the horizon, breaking the direct functional dependence between geometric volume and area. In this work, we resolve this difficulty by establishing a quasi-local thermodynamic framework for Kerr-Newman black holes. We demonstrate that accommodating this kinematic deformation requires extending the thermodynamic phase space to include a geometric eccentricity parameter $Y$ and its conjugate, a thermodynamic shear tension $X$. Consequently, the rotational contribution is incorporated into the first law with a shear work term $X dY$. We derive the generalized first laws and Smarr formulas (Euler relations) for both the internal energy and enthalpy representations, showing that these thermodynamic potentials can be obtained through Legendre transformations that isolate the quasi-local energy from the rotational energy. Thus, this framework provides a novel perspective on the thermodynamics of rotating black holes, integrating the geometric deformation of the horizon into a quasi-local description.

We study the cosmological and thermodynamic implications of holographic dark energy derived from the Kaniadakis deformation of the Bekenstein-Hawking entropy. Within a spatially flat FLRW background, the generalized entropy leads to an effective dark energy density containing an infrared correction proportional to $H^{-2}$, modifying the dynamics of the apparent horizon. Using the Hayward Kodama formalism, we obtain a geometric equation of state and perform a criticality analysis, revealing a Van der Waals type structure with an inverted first order phase transition and a non physical swallowtail behavior in the Gibbs free energy, indicative of unstable thermodynamic branches. We further examine a dynamical extension including a $\dot{H}$ contribution and show that the unconventional critical behavior persists. The phenomenological viability of the model is tested through a joint statistical analysis with cosmic chronometers, PantheonPlus Type Ia supernovae, and DESI baryon acoustic oscillation data. These results establish Kaniadakis holographic cosmology as a consistent framework linking generalized entropy, gravitational thermodynamics, and observationally viable dark energy dynamics.

This paper revisits the classical construction of initial data using the conformal method, as originally proposed by Holst, Nagy, and Tsogtgerel and later refined by Maxwell. We demonstrate that the existence of the solution can be proven using the Banach fixed point theorem, whereas the original proof relied on the Schauder fixed point theorem. This new approach has two main advantages: it guarantees the uniqueness of the solution to the equations of the conformal method as soon as one imposes a bound on the physical volume of it and it provides an explicit construction of the solution.

We analyze the causal structure of horizonless compact objects via the light-cone geometry and conformal compactification of the Joshi-Malafarina-Narayan (JMN-1) and Janis-Newman-Winicour (JNW) spacetimes. Penrose diagrams reveal that JMN-1 undergoes a transition from timelike $(0<M_0<2/3)$ to null $(2/3<M_0<4/5)$ singularities, while JNW remains timelike throughout, in contrast to the spacelike singularity of the Schwarzschild spacetime. We show that photon spheres exist in Schwarzschild and JNW, but arise in JMN-1 only in the null singularity phase, establishing a direct link between causal character and null geodesic trapping. We further demonstrate that radial timelike geodesics develop turning points for certain parameter regimes in both JMN-1 and JNW spacetimes, indicating the emergence of effective repulsive behavior in the strong field region. These features lead to distinct strong field lensing and shadow signatures, potentially testable by very long baseline interferometric observations such as those of the Event Horizon Telescope.

A central issue in the thermodynamics of nonsingular black holes is the apparent violation of the first law. In this work, we use 2D dilaton gravity as a simple theoretical setting to study this issue. We systematically construct a broad class of nonsingular black hole solutions with metric function $A(x)=f(x)+c$, through a procedure that is considerably simpler than in higher-dimensional theories. Using the Iyer-Wald covariant phase space formalism, we derive the correct energy formula and establish a consistent first law for this entire class of solutions. The apparent first-law violation in a previous work is caused by an incorrect choice of energy. Our energy formula agrees, up to a normalization factor for the asymptotic Killing vector, with the Casimir function in 2D dilaton gravity, confirming its interpretation as the physical black hole energy. Our results clarify the correct first law for 2D nonsingular black holes and may provide insights into the first law of nonsingular black holes in higher dimensions.

Ever since a new symmetry was found for the imperfect fluid with vorticity the question of the effect of perturbations on the symmetry itself has been raised. This new symmetry arose when realizing that local four-velocity gauge-like transformations would render the left hand side of the Einstein equations invariant. Because the metric tensor would be invariant under this new kind of local transformations. Then the point was raised about the invariance of such a kind of transformations of the stress-energy tensor on the right hand side of the Einstein equations in curved four-dimensional Lorentz spacetimes. It was verified that these invariances do not work with plain perfect fluids but they do work for imperfect fluids. The imperfect fluid stress-energy tensor will be invariant under local four-velocity gauge-like transformations when additional transformations are introduced for several variables included in the stress-energy tensor itself. This local invariance was also the criteria introduced in order to present a new stress-energy tensor for vorticity as well. New tetrads are at the core of the realization of the existence of this new symmetry because it is through these new tetrads that this new symmetry is realized. In this manuscript we will introduce local perturbations by external agents to the relevant objects in the imperfect fluid geometry. We will demonstrate a theorem that proves that the symmetries under four-velocity gauge-like transformations will be instantaneously broken but at the same time transformed into new symmetries. Because the local orthogonal planes determined by these new tetrads, which happen to be the local planes of symmetry will tilt under local perturbations. There will be a symmetry evolution under perturbations.

We present a model-independent reconstruction of the quintessence scalar field's dynamics-both its potential and kinetic energy-directly from the latest cosmological observations. Our analysis combines DESI DR2 baryon acoustic oscillation measurements with the Pantheon plus Type Ia supernova compilation, employing Gaussian process with four distinct covariance kernels to avoid theoretical priors on the potential's functional form. Key findings reveal a monotonically decreasing potential with redshift, consistent with thawing quintessence, and a kinetic energy that crosses zero near $z\sim 1$, marking the dark energy-matter equality epoch. Notably, while apparent negative kinetic energy values emerge at intermediate redshifts (0.5<z<1.0), these are statistical artifacts within uncertainties, arising from error amplification in derivative reconstruction rather than new physics. Our results demonstrate the power of non-parametric methods to constrain dynamical dark energy and show minimal dependence on the choice of cosmological priors, whether from local (SH0ES) or early-universe (Planck) measurements.

We develop a time-dependent, grid-based framework for simulating infrared spectra that is specifically designed for quantum computers. The proposed circuit employs a probabilistic strategy for applying the non-unitary dipole operator and an Split Operator-Quantum Fourier Transform time evolution scheme. Using a vibrational model of the water molecule as a test system, our classical emulation results demonstrate accurate determination of fundamental and overtone band positions and intensities via Fourier-transformed dipole-dipole autocorrelation functions. We also identify the optimal time parameters that minimise gate depths while maintaining high fidelity. For further resource reduction, we validate the feasibility of utilising harmonic oscillator approximations in state preparation and dipole operator truncations. With its scalability to higher-dimensional normal mode spaces, this wavefunction-based approach establishes a robust foundation for studying IR spectra on future quantum hardware.

Quantum federated learning (QFL) has recently emerged as a promising paradigm for privacy-preserving collaborative learning, yet most existing studies focus on horizontal federated learning and ignore the vertical federated learning (VFL), where parties hold complementary features of aligned samples. In this work, we propose Evidential Quantum Vertical Federated Learning (eviQVFL), a VFL-tailored QFL framework that employs a hybrid classical-quantum architecture for party-side feature processing, mapping local features into a quantum state. To preserve privacy and avoid information loss, party-side output states are directly transmitted to the server via quantum teleportation, and the server fuses the received quantum states with a non-parametric evidential fusion circuit grounded in evidence theory, followed by measurement-based inference. Extensive simulations on image classification and other real-world datasets demonstrate that eviQVFL consistently achieves higher classification accuracy than other classical and quantum baselines under comparable parameter budgets. Both empirical observations and theoretical analysis indicate that eviQVFL achieve less approximation error with limited quantum resources, while maintaining training stability and offering stronger feature privacy.

We examine the existence of an apparent horizon in the collapse of an isolated dust cloud using expansion functions. Our results indicate that in the region of spacetime far away from the gravitational singularity, the considered system has a horizon, in which case the singularity is covered. Using this method may have limited applicability in the neighbourhood of the gravitational singularity, where quantum effects are expected to be essential.

We consider a broad class of static, spherically symmetric generalized Schwarzschild-like solutions with multiple non-interacting anisotropic fluid sources and derive the coordinate transformation from Schwarzschild-like (curvature) to isotropic coordinates with conformally flat spatial slices. The isotropic form removes spatial-sector coordinate pathologies at the horizon, clarifies geometric quantities (e.g., ADM mass and curvature invariants), and enables the construction of well-posed initial data on t=const hypersurfaces, suitable for the Hamiltonian and conformal formulations of numerical relativity and for perturbation theory. The backgrounds in isotropic coordinates we develop make it straightforward to separate environmental effects from intrinsic strong-gravity signals and meet the growing interest in non-vacuum black hole phenomenology across scattering, lensing, and waveform modeling.

Symmetry-adapted variational quantum eigensolvers (VQE) based on the Unitary Coupled-Cluster ansatz (SymUCCSD) effectively reduce the parameter count for Abelian molecular point groups, yet they systematically fail for non-Abelian groups without a fully established theoretical explanation. In this work, we prove that the Abelian-subgroup restriction induces a spurious splitting of multidimensional irreducible representations, prematurely discarding cross-component excitations. At the Lie-algebraic level, this filter confines the Dynamical Lie Algebra (DLA) to the Abelian subalgebra $\mathfrak{u}(1)^{d_λ}$, restricting the reachable state manifold to a measure-zero torus $\mathbb{T}^{d_λ}$. However, completing the algebra is insufficient on its own due to a critical numerical trap: when standard molecular orbitals adapted solely to an Abelian subgroup are utilized, cross-component integrals vanish identically, creating an artificial zero-gradient plateau along non-Abelian algebraic directions. Numerical experiments on NH$_3$/STO-3G ($C_{3v}$, 16 qubits) confirm both the predicted DLA confinement and the gradient plateau, with SymUCCSD converging to an error of $21.8$~mHa above the FCI energy despite full optimizer convergence. Our analysis provides a rigorous algebraic and geometric diagnosis for the observed numerical breakdown, establishing that recovering full equivariant dynamics fundamentally necessitates both the inclusion of complete off-diagonal generators and the strategic parametrization of non-Abelian degrees of freedom.

We propose an erasure conversion scheme on the $|e\rangle-|f\rangle$ and $|g\rangle-|f\rangle$ qubits in integer fluxonium qubits (IFQs), which are both first-order insensitive to $1/f$ flux noise. The $|e\rangle-|f\rangle$ transition is identical to that of a usual fluxonium qubit and hence is expected to have excellent coherence time, while the $|g\rangle-|f\rangle$ transition is additionally protected from the energy relaxation by the parity symmetry. The dominant error in both qubits arises due to the energy relaxation: from $|e\rangle$ to $|g\rangle$ in the $e\text{--}f$ qubit and from $|f\rangle$ to $|e\rangle$ in the $g\text{--}f$ qubit. Such errors can be treated as erasure events, and their efficient detection improves the performance of quantum error-correcting codes. We consider a protocol for such erasure conversion based on the dispersive readout. Our main finding is that, with proper circuit parameter choice, carefully designed gate sets, and the integration of erasure conversion, IFQs promise high effective coherence times.

The purpose of this paper is to explore, in a space of four-dimensions, the possible forms that second-order, bi-scalar-vector-tensor field equations derivable from a variational principle can assume. In order to restrict this enormous class of field equations I shall first require that the equations governing the vector field (which will be identified with the vector potential of an electromagnetic field) be consistent with the notion of conservation of charge. Secondly I shall require that these vector equations reduce to Maxwell's equations in a flat space when the scalar fields are constant. Unfortunately even with these two powerful restrictions on the form of the field equations I have not been able to construct a Lagrangian which yields all possible field equations of this nature. This situation will lead to a discussion of other ways in which the field equations can be restricted to obtain viable bi-scalar-vector-tensor field equations. Lastly I shall make a few remarks on how the results obtained can be used to show that the Higgs field can generate electromagnetic fields in the early Universe, and that it is not going to be practical to couple bi-scalar fields to gauge-tensor fields to construct a bi-scalar-Yang-Mills-tensor field theory.

We investigate the detectability of Gravitational Wave (GW) modes (emitted by black-holes and neutron stars) by third generation, ground-based gravitational wave detectors planned to be operational in the next decade. Our analysis focuses on the Cosmic Explorer and Einstein Telescope projects, which are expected to have arm lengths of tens of kilometers and to experience the amplification of a gravitational wave signal at their Full-Spectral Range (FSR) frequencies. We find that both projects will also observe with good Signal-to-Noise ratio (SNR) the elusive {\it w-modes}, which are expected to be emitted at these frequencies by spinning neutron stars.

We obtain the spectra of codimension-2 horizon "edge" degrees of freedom for gravity and higher-spin gauge fields in de Sitter space and in the static Nariai spacetime, advancing previous Lorentzian and Euclidean analyses of one-loop thermodynamics. The edge spectra exhibit universal shift symmetries, revealing a novel symmetry-breaking structure in one-loop partition functions with positive cosmological constant. For the graviton, these modes admit a geometric interpretation as fluctuations of the cosmic horizon, which also persists in the Nariai case.

A system, comprised of a qubit pair coupled to a common cavity, is studied with the aim of establishing qubit entanglement. This study is the sequel of the paper Phys. Rev. A 111, 043705 (2025), where similar model was investigated for an initially vacuum cavity. In the present manuscript the cavity with finite initial occupancy is considered and the effect of asymmetric qubit cavity couplings is investigated. For a closed system scenario, the ratio of the qubit-cavity couplings shows a threshold beyond which no maximally-entangled qubit state is available. The threshold value is shown to depend on the excitation level of the cavity. For a driven-dissipative case steady state entanglement is shown to depend non-monotonically on the qubit drive. Intricate interplay of drive, dissipation, and coupling asymmetry is shown to be pivotal for steady-state entanglement generation.

Cascaded emission from the biexciton state of a quantum dot results in polarization entangled photon pairs. However, modelling this system becomes challenging when photon emission is cavity-mediated due to the large Hilbert space and non-Markovian nature of its phonon-induced decoherence. Here, we introduce an algorithm that reduces the computational cost of the numerically exact process tensor method for non-Markovian dynamics simulations when the environmental coupling operator has degenerate eigenvalues, making calculations of the non-Markovian dynamics of large systems feasible. We compute the degree of entanglement of photon pairs generated by pulsed two-photon resonant excitation and find surprisingly good agreement between the numerically exact results and those calculated using the approximate polaron master equation method, permitting an efficient exploration of trends across system parameters.

The quantum theory of the electromagnetic field enables the description of multiphoton states exhibiting nonclassical statistical properties, often reflected in non-Gaussian phase-space distributions. While non-Gaussianity alone does not fully characterize quantum states, several classifications have been proposed to hierarchize non-Gaussian states according to physically or informationally relevant resources. Here, we provide a physical interpretation of non-Gaussianity and connect it to a computational perspective by showing how a prominent classification-the stellar rank-emerges as a limiting case of the roots of polynomials that univocally represent bosonic states defined with a quantized phase reference, namely the Majorana polynomials. A direct consequence of our results is a revised interpretation of both the stellar rank and non-Gaussianity itself: when superselection rules are properly taken into account, quadrature non-Gaussianity - and nonzero stellar rank - act as witnesses of particle entanglement, rather than being linked with photon addition to Gaussian states as previously assumed. In addition, we show that because the stellar rank depends on a specific choice of coherent states, its relation to computational resources and potential quantum advantage is inherently basis-dependent, being naturally tied to quadrature eigenstates as the computational basis. Motivated by this observation, we generalize the notion of stellar rank to arbitrary computational bases, thereby establishing it as a genuine witness of bosonic resources that may enable quantum advantage.