arXiv Daily Digest - 2026-03-27

Vision-language-action models have reshaped autonomous driving to incorporate languages into the decision-making process. However, most existing pipelines only utilize the language modality for scene descriptions or reasoning and lack the flexibility to follow diverse user instructions for personalized driving. To address this, we first construct a large-scale driving dataset (InstructScene) containing around 100,000 scenes annotated with diverse driving instructions with the corresponding trajectories. We then propose a unified Vision-Language-World-Action model, Vega, for instruction-based generation and planning. We employ the autoregressive paradigm to process visual inputs (vision) and language instructions (language) and the diffusion paradigm to generate future predictions (world modeling) and trajectories (action). We perform joint attention to enable interactions between the modalities and use individual projection layers for different modalities for more capabilities. Extensive experiments demonstrate that our method not only achieves superior planning performance but also exhibits strong instruction-following abilities, paving the way for more intelligent and personalized driving systems.

Human driving behavior is inherently personal, which is shaped by long-term habits and influenced by short-term intentions. Individuals differ in how they accelerate, brake, merge, yield, and overtake across diverse situations. However, existing end-to-end autonomous driving systems either optimize for generic objectives or rely on fixed driving modes, lacking the ability to adapt to individual preferences or interpret natural language intent. To address this gap, we propose Drive My Way (DMW), a personalized Vision-Language-Action (VLA) driving framework that aligns with users' long-term driving habits and adapts to real-time user instructions. DMW learns a user embedding from our personalized driving dataset collected across multiple real drivers and conditions the policy on this embedding during planning, while natural language instructions provide additional short-term guidance. Closed-loop evaluation on the Bench2Drive benchmark demonstrates that DMW improves style instruction adaptation, and user studies show that its generated behaviors are recognizable as each driver's own style, highlighting personalization as a key capability for human-centered autonomous driving. Our data and code are available at https://dmw-cvpr.github.io/.

The knowledge base in a retrieval-augmented generation (RAG) system is typically assembled once and never revised, even though the facts a query requires are often fragmented across documents and buried in irrelevant content. We argue that the knowledge base should be treated as a trainable component and propose WriteBack-RAG, a framework that uses labeled examples to identify where retrieval succeeds, isolate the relevant documents, and distill them into compact knowledge units that are indexed alongside the original corpus. Because the method modifies only the corpus, it can be applied once as an offline preprocessing step and combined with any RAG pipeline. Across four RAG methods, six benchmarks, and two LLM backbones, WriteBack-RAG improves every evaluated setting, with gains averaging +2.14%. Cross-method transfer experiments further show that the distilled knowledge benefits RAG pipelines other than the one used to produce it, confirming that the improvement resides in the corpus itself.

Autoregressive video diffusion models have demonstrated remarkable progress, yet they remain bottlenecked by intractable linear KV-cache growth, temporal repetition, and compounding errors during long-video generation. To address these challenges, we present PackForcing, a unified framework that efficiently manages the generation history through a novel three-partition KV-cache strategy. Specifically, we categorize the historical context into three distinct types: (1) Sink tokens, which preserve early anchor frames at full resolution to maintain global semantics; (2) Mid tokens, which achieve a massive spatiotemporal compression (32x token reduction) via a dual-branch network fusing progressive 3D convolutions with low-resolution VAE re-encoding; and (3) Recent tokens, kept at full resolution to ensure local temporal coherence. To strictly bound the memory footprint without sacrificing quality, we introduce a dynamic top-$k$ context selection mechanism for the mid tokens, coupled with a continuous Temporal RoPE Adjustment that seamlessly re-aligns position gaps caused by dropped tokens with negligible overhead. Empowered by this principled hierarchical context compression, PackForcing can generate coherent 2-minute, 832x480 videos at 16 FPS on a single H200 GPU. It achieves a bounded KV cache of just 4 GB and enables a remarkable 24x temporal extrapolation (5s to 120s), operating effectively either zero-shot or trained on merely 5-second clips. Extensive results on VBench demonstrate state-of-the-art temporal consistency (26.07) and dynamic degree (56.25), proving that short-video supervision is sufficient for high-quality, long-video synthesis. https://github.com/ShandaAI/PackForcing

Fine-grained facial expression editing has long been limited by intrinsic semantic overlap. To address this, we construct the Flex Facial Expression (FFE) dataset with continuous affective annotations and establish FFE-Bench to evaluate structural confusion, editing accuracy, linear controllability, and the trade-off between expression editing and identity preservation. We propose PixelSmile, a diffusion framework that disentangles expression semantics via fully symmetric joint training. PixelSmile combines intensity supervision with contrastive learning to produce stronger and more distinguishable expressions, achieving precise and stable linear expression control through textual latent interpolation. Extensive experiments demonstrate that PixelSmile achieves superior disentanglement and robust identity preservation, confirming its effectiveness for continuous, controllable, and fine-grained expression editing, while naturally supporting smooth expression blending.

Automatic speech recognition (ASR) systems have achieved near-human accuracy on curated benchmarks, yet still fail in real-world voice agents under conditions that current evaluations do not systematically cover. Without diagnostic tools that isolate specific failure factors, practitioners cannot anticipate which conditions, in which languages, will cause what degree of degradation. We introduce WildASR, a multilingual (four-language) diagnostic benchmark sourced entirely from real human speech that factorizes ASR robustness along three axes: environmental degradation, demographic shift, and linguistic diversity. Evaluating seven widely used ASR systems, we find severe and uneven performance degradation, and model robustness does not transfer across languages or conditions. Critically, models often hallucinate plausible but unspoken content under partial or degraded inputs, creating concrete safety risks for downstream agent behavior. Our results demonstrate that targeted, factor-isolated evaluation is essential for understanding and improving ASR reliability in production systems. Besides the benchmark itself, we also present three analytical tools that practitioners can use to guide deployment decisions.

Agent performance increasingly depends on \emph{harness engineering}, yet harness design is usually buried in controller code and runtime-specific conventions, making it hard to transfer, compare, and study as a scientific object. We ask whether the high-level control logic of an agent harness can instead be externalized as a portable executable artifact. We introduce \textbf{Natural-Language Agent Harnesses} (NLAHs), which express harness behavior in editable natural language, and \textbf{Intelligent Harness Runtime} (IHR), a shared runtime that executes these harnesses through explicit contracts, durable artifacts, and lightweight adapters. Across coding and computer-use benchmarks, we conduct controlled evaluations of operational viability, module ablation, and code-to-text harness migration.

Contrastive vision-language (V&L) models remain a popular choice for various applications. However, several limitations have emerged, most notably the limited ability of V&L models to learn compositional representations. Prior methods often addressed this limitation by generating custom training data to obtain hard negative samples. Hard negatives have been shown to improve performance on compositionality tasks, but are often specific to a single benchmark, do not generalize, and can cause substantial degradation of basic V&L capabilities such as zero-shot or retrieval performance, rendering them impractical. In this work we follow a different approach. We identify two root causes that limit compositionality performance of V&Ls: 1) Long training captions do not require a compositional representation; and 2) The final global pooling in the text and image encoders lead to a complete loss of the necessary information to learn binding in the first place. As a remedy, we propose two simple solutions: 1) We obtain short concept centric caption parts using standard NLP software and align those with the image; and 2) We introduce a parameter-free cross-modal attention-pooling to obtain concept centric visual embeddings from the image encoder. With these two changes and simple auxiliary contrastive losses, we obtain SOTA performance on standard compositionality benchmarks, while maintaining or improving strong zero-shot and retrieval capabilities. This is achieved without increasing inference cost. We release the code for this work at https://github.com/SamsungLabs/concept_centric_clip.

Robust perception and reasoning require consistency across sensory modalities. Yet current multimodal models often violate this principle, yielding contradictory predictions for visual and textual representations of the same concept. Rather than masking these failures with standard voting mechanisms, which can amplify systematic biases, we show that cross-modal inconsistency provides a rich and natural signal for learning. We introduce RC2, a reinforcement learning framework that resolves internal conflicts by enforcing cross-modal cycle consistency. By requiring a model to perform backward inference, switch modalities, and reliably reconstruct the answer through forward inference, we obtain a dense, label-free reward. This cyclic constraint encourages the model to align its internal representations autonomously. Optimizing for this structure mitigates modality-specific errors and improves reasoning accuracy by up to 7.6 points. Our results suggest that advanced reasoning emerges not only from scaling data, but also from enforcing a structurally consistent understanding of the world.

We present an empirical study of how far general-purpose coding agents -- without hardware-specific training -- can optimize hardware designs from high-level algorithmic specifications. We introduce an agent factory, a two-stage pipeline that constructs and coordinates multiple autonomous optimization agents. In Stage~1, the pipeline decomposes a design into sub-kernels, independently optimizes each using pragma and code-level transformations, and formulates an Integer Linear Program (ILP) to assemble globally promising configurations under an area constraint. In Stage~2, it launches $N$ expert agents over the top ILP solutions, each exploring cross-function optimizations such as pragma recombination, loop fusion, and memory restructuring that are not captured by sub-kernel decomposition. We evaluate the approach on 12 kernels from HLS-Eval and Rodinia-HLS using Claude Code (Opus~4.5/4.6) with AMD Vitis HLS. Scaling from 1 to 10 agents yields a mean $8.27\times$ speedup over baseline, with larger gains on harder benchmarks: streamcluster exceeds $20\times$ and kmeans reaches approximately $10\times$. Across benchmarks, agents consistently rediscover known hardware optimization patterns without domain-specific training, and the best designs often do not originate from top-ranked ILP candidates, indicating that global optimization exposes improvements missed by sub-kernel search. These results establish agent scaling as a practical and effective axis for HLS optimization.

Video world models have shown immense potential in simulating the physical world, yet existing memory mechanisms primarily treat environments as static canvases. When dynamic subjects hide out of sight and later re-emerge, current methods often struggle, leading to frozen, distorted, or vanishing subjects. To address this, we introduce Hybrid Memory, a novel paradigm requiring models to simultaneously act as precise archivists for static backgrounds and vigilant trackers for dynamic subjects, ensuring motion continuity during out-of-view intervals. To facilitate research in this direction, we construct HM-World, the first large-scale video dataset dedicated to hybrid memory. It features 59K high-fidelity clips with decoupled camera and subject trajectories, encompassing 17 diverse scenes, 49 distinct subjects, and meticulously designed exit-entry events to rigorously evaluate hybrid coherence. Furthermore, we propose HyDRA, a specialized memory architecture that compresses memory into tokens and utilizes a spatiotemporal relevance-driven retrieval mechanism. By selectively attending to relevant motion cues, HyDRA effectively preserves the identity and motion of hidden subjects. Extensive experiments on HM-World demonstrate that our method significantly outperforms state-of-the-art approaches in both dynamic subject consistency and overall generation quality.

Block-diffusion language models offer a promising path toward faster-than-autoregressive generation by combining block-wise autoregressive decoding with within-block parallel denoising. However, in the few-step regime needed for practical acceleration, standard confidence-thresholded decoding is often brittle: aggressive thresholds hurt quality, while conservative thresholds require unnecessary denoising steps. Existing approaches that address this issue either require additional training or incur extra test-time compute. We present S2D2, a training-free self-speculative decoding framework for block-diffusion language models. Our key observation is that a block-diffusion model becomes autoregressive when the block size is reduced to one, allowing the same pretrained model to act as both drafter and verifier. S2D2 inserts a speculative verification step into standard block-diffusion decoding and uses lightweight routing policies to decide when verification is worth its cost. This yields a hybrid decoding trajectory in which diffusion proposes tokens in parallel, while the autoregressive mode acts as a local sequence-level critic. Across three mainstream block-diffusion families, S2D2 consistently improves the accuracy-speed tradeoff over strong confidence-thresholding baselines. On SDAR, we observe up to $4.7\times$ speedup over autoregressive decoding, and up to $1.57\times$ over a tuned dynamic decoding baseline while improving accuracy by up to $4.5$ points. On LLaDA2.1-Mini, S2D2 remains complementary to built-in self-correction, including a conservative setting where it is $4.4\times$ faster than the static baseline with slightly higher accuracy.

Transfer learning and knowledge distillation has recently gained a lot of attention in the deep learning community. One transfer approach, the student-teacher learning, has been shown to successfully create ``small'' student neural networks that mimic the performance of a much bigger and more complex ``teacher'' networks. In this paper, we investigate an extension to this approach and transfer from a non-neural-based machine learning pipeline as teacher to a neural network (NN) student, which would allow for joint optimization of the various pipeline components and a single unified inference engine for multiple ML tasks. In particular, we explore replacing the random forest classifier by transfer learning to a student NN. We experimented with various NN topologies on 100 OpenML tasks in which random forest has been one of the best solutions. Our results show that for the majority of the tasks, the student NN can indeed mimic the teacher if one can select the right NN hyper-parameters. We also investigated the use of random forest for selecting the right NN hyper-parameters.

Code production is now a commodity; the bottleneck is knowing what to build and proving it works. We present the Kitchen Loop, a framework for autonomous, self-evolving software built on a unified trust model: (1) a specification surface enumerating what the product claims to support; (2) 'As a User x 1000', where an LLM agent exercises that surface as a synthetic power user at 1,000x human cadence; (3) Unbeatable Tests, ground-truth verification the code author cannot fake; and (4) Drift Control, continuous quality measurement with automated pause gates. We validate across two production systems over 285+ iterations, producing 1,094+ merged pull requests with zero regressions detected by the regression oracle (methodology in Section 6.1). We observe emergent properties at scale: multi-iteration self-correction chains, autonomous infrastructure healing, and monotonically improving quality gates. The primitives are not new; our contribution is their composition into a production-tested system with the operational discipline that makes long-running autonomous evolution safe.

Trustworthy artificial intelligence increasingly relies on probabilistic computation to achieve robustness, interpretability, security and privacy. In practical systems, such workloads interleave deterministic data access with repeated stochastic sampling across models, data paths and system functions, shifting performance bottlenecks from arithmetic units to memory systems that must deliver both data and randomness. Here we present a unified data-access perspective in which deterministic access is treated as a limiting case of stochastic sampling, enabling both modes to be analyzed within a common framework. This view reveals that increasing stochastic demand reduces effective data-access efficiency and can drive systems into entropy-limited operation. Based on this insight, we define memory-level evaluation criteria, including unified operation, distribution programmability, efficiency, robustness to hardware non-idealities and parallel compatibility. Using these criteria, we analyze limitations of conventional architectures and examine emerging probabilistic compute-in-memory approaches that integrate sampling with memory access, outlining pathways toward scalable hardware for trustworthy AI.

Neural scaling laws, which in some domains can predict the performance of large neural networks as a function of model, data, and compute scale, are the cornerstone of building foundation models in Natural Language Processing and Computer Vision. We study neural scaling in Scientific Machine Learning, focusing on models for weather forecasting. To analyze scaling behavior in as simple a setting as possible, we adopt a minimal, scalable, general-purpose Swin Transformer architecture, and we use continual training with constant learning rates and periodic cooldowns as an efficient training strategy. We show that models trained in this minimalist way follow predictable scaling trends and even outperform standard cosine learning rate schedules. Cooldown phases can be re-purposed to improve downstream performance, e.g., enabling accurate multi-step rollouts over longer forecast horizons as well as sharper predictions through spectral loss adjustments. We also systematically explore a wide range of model and dataset sizes under various compute budgets to construct IsoFLOP curves, and we identify compute-optimal training regimes. Extrapolating these trends to larger scales highlights potential performance limits, demonstrating that neural scaling can serve as an important diagnostic for efficient resource allocation. We open-source our code for reproducibility.

Cross-view geo-localization (CVGL) estimates a camera's location by matching a street-view image to geo-referenced overhead imagery, enabling GPS-denied localization and navigation. Existing methods almost universally formulate CVGL as an image-retrieval problem in a contrastively trained embedding space. This ties performance to large batches and hard negative mining, and it ignores both the geometric structure of maps and the coverage mismatch between street-view and overhead imagery. In particular, salient landmarks visible from the street view can fall outside a fixed satellite crop, making retrieval targets ambiguous and limiting explicit spatial inference over the map. We propose Just Zoom In, an alternative formulation that performs CVGL via autoregressive zooming over a city-scale overhead map. Starting from a coarse satellite view, the model takes a short sequence of zoom-in decisions to select a terminal satellite cell at a target resolution, without contrastive losses or hard negative mining. We further introduce a realistic benchmark with crowd-sourced street views and high-resolution satellite imagery that reflects real capture conditions. On this benchmark, Just Zoom In achieves state-of-the-art performance, improving Recall@1 within 50 m by 5.5% and Recall@1 within 100 m by 9.6% over the strongest contrastive-retrieval baseline. These results demonstrate the effectiveness of sequential coarse-to-fine spatial reasoning for cross-view geo-localization.

As large language models (LLMs) continue to advance, improving them solely through human supervision is becoming increasingly costly and limited in scalability. As models approach human-level capabilities in certain domains, human feedback may no longer provide sufficiently informative signals for further improvement. At the same time, the growing ability of models to make autonomous decisions and execute complex actions naturally enables abstractions in which components of the model development process can be progressively automated. Together, these challenges and opportunities have driven increasing interest in self-improvement, where models autonomously generate data, evaluate outputs, and iteratively refine their own capabilities. In this paper, we present a system-level perspective on self-improving language models and introduce a unified framework that organizes existing techniques. We conceptualize the self-improvement system as a closed-loop lifecycle, consisting of four tightly coupled processes: data acquisition, data selection, model optimization, and inference refinement, along with an autonomous evaluation layer. Within this framework, the model itself plays a central role in driving each stage: collecting or generating data, selecting informative signals, updating its parameters, and refining outputs, while the autonomous evaluation layer continuously monitors progress and guides the improvement cycle across stages. Following this lifecycle perspective, we systematically review and analyze representative methods for each component from a technical standpoint. We further discuss current limitations and outline our vision for future research toward fully self-improving LLMs.

Automated systems have been widely adopted across the educational testing industry for open-response assessment and essay scoring. These systems commonly achieve performance levels comparable to or superior than trained human raters, but have frequently been demonstrated to be vulnerable to the influence of construct-irrelevant factors (i.e., features of responses that are unrelated to the construct assessed) and adversarial conditions. Given the rising usage of large language models in automated scoring systems, there is a renewed focus on ``hallucinations'' and the robustness of these LLM-based automated scoring approaches to construct-irrelevant factors. This study investigates the effects of construct-irrelevant factors on a dual-architecture LLM-based scoring system designed to score short essay-like open-response items in a situational judgment test. It was found that the scoring system was generally robust to padding responses with meaningless text, spelling errors, and writing sophistication. Duplicating large passages of text resulted in lower scores predicted by the system, on average, contradicting results from previous studies of non-LLM-based scoring systems, while off-topic responses were heavily penalized by the scoring system. These results provide encouraging support for the robustness of future LLM-based scoring systems when designed with construct relevance in mind.

Early detection of atypical cognitive-motor development is critical for timely intervention, yet traditional assessments rely heavily on subjective, static evaluations. The integration of digital devices offers an opportunity for continuous, objective monitoring through digital biomarkers. In this work, we propose an AI-driven longitudinal framework to model developmental trajectories in children aged 18 months to 8 years. Using a dataset of tablet-based interactions collected over multiple academic years, we analyzed six cognitive-motor tasks (e.g., fine motor control, reaction time). We applied dimensionality reduction (t-SNE) and unsupervised clustering (K-Means++) to identify distinct developmental phenotypes and tracked individual transitions between these profiles over time. Our analysis reveals three distinct profiles: low, medium, and high performance. Crucially, longitudinal tracking highlights a high stability in the low-performance cluster (>90% retention in early years), suggesting that early deficits tend to persist without intervention. Conversely, higher-performance clusters show greater variability, potentially reflecting engagement factors. This study validates the use of unsupervised learning on touchscreen data to uncover heterogeneous developmental paths. The identified profiles serve as scalable, data-driven proxies for cognitive growth, offering a foundation for early screening tools and personalized pediatric interventions.

As superconducting processors scale, understanding how physical layout shapes qubit interactions is essential for architectural reliability. Existing methods offer limited insight into how electromagnetic design choices translate into execution-level behavior. We present EPAR, an electromagnetic-to-architecture framework that predicts robustness early directly from physical design by reconstructing how design distortion modifies the effective Hamiltonian, reroutes mediated connectivity, and influences control-pulse response. Across all tested layouts, EPAR's structural scores show 100% agreement with two-qubit error trends yet reveal over 10X robustness differences among edges with identical calibrated error rates, going beyond conventional metrics to provide improved and actionable compiler guidance.

Safety monitoring is essential for Cyber-Physical Systems (CPSs). However, unsafe events are rare in real-world CPS operations, creating an extreme class imbalance that degrades safety predictors. Standard rebalancing techniques perform poorly on time-series CPS telemetry, either generating unrealistic synthetic samples or overfitting on the minority class. Meanwhile, behavioral uncertainty in CPS operations, defined as the degree of doubt or uncertainty in CPS decisions , is often correlated with safety outcomes but unexplored in safety monitoring. To that end, we propose U-Balance, a supervised approach that leverages behavioral uncertainty to rebalance imbalanced datasets prior to training a safety predictor. U-Balance first trains a GatedMLP-based uncertainty predictor that summarizes each telemetry window into distributional kinematic features and outputs an uncertainty score. It then applies an uncertainty-guided label rebalancing (uLNR) mechanism that probabilistically relabels \textit{safe}-labeled windows with unusually high uncertainty as \textit{unsafe}, thereby enriching the minority class with informative boundary samples without synthesizing new data. Finally, a safety predictor is trained on the rebalanced dataset for safety monitoring. We evaluate U-Balance on a large-scale UAV benchmark with a 46:1 safe-to-unsafe ratio. Results confirm a moderate but significant correlation between behavioral uncertainty and safety. We then identify uLNR as the most effective strategy to exploit uncertainty information, compared to direct early and late fusion. U-Balance achieves a 0.806 F1 score, outperforming the strongest baseline by 14.3 percentage points, while maintaining competitive inference efficiency. Ablation studies confirm that both the GatedMLP-based uncertainty predictor and the uLNR mechanism contribute significantly to U-Balance's effectiveness.

SHAPR (Solo Human-Centred and AI-Assisted Practice) is a framework for research software development that integrates human-centred decision-making with AI-assisted capabilities. While prior work introduced SHAPR as a conceptual framework, this paper focuses on its operationalisation as a structured, traceable, and knowledge-generating approach to AI-assisted research practice. We present a set of interconnected models describing how research activities are organised through iterative cycles (Explore-Build-Use-Evaluate-Learn), how artefacts evolve through development and use, and how empirical evidence is transformed into conceptual knowledge. Central to this process are Structured Knowledge Units (SKUs), which provide modular and reusable representations of insights derived from practice, supporting knowledge accumulation across cycles. The framework introduces evidence and traceability as a cross-cutting mechanism linking human decisions, AI-assisted development, and artefact evolution to enable transparency, reproducibility, and systematic refinement. SHAPR is also positioned as an AI-executable research framework, as its structured processes and documentation can be interpreted by generative AI systems to guide research workflows. Simultaneously, SHAPR supports a continuum of AI involvement, allowing researchers to balance control, learning, and automation across different contexts. Beyond individual workflows, SHAPR is conceptualised as an integrated research system combining LLM workspaces, development environments, cloud storage, and version control to support scalable, knowledge-centred research practices. Overall, SHAPR provides a practical and theoretically grounded foundation for conducting rigorous, transparent, and reproducible research in AI-assisted environments, contributing to the development of scalable and methodologically sound research practices.

This paper presents an experimental platform for studying intentional-state attribution toward a non-humanoid robot. The system combines a simulated robot, realistic task environments, and large language model-based explanatory layers that can express the same behavior in mentalistic, teleological, or mechanistic terms. By holding behavior constant while varying the explanatory frame, the platform provides a controlled way to investigate how language and framing shape the adoption of the intentional stance in robotics.

Accurate parsing of citations is necessary for machine-readable scholarly infrastructure. But, despite sustained interest in this problem, existing evaluation techniques are often not generalizable, based on synthetic data, or not publicly available. We introduce RenoBench, a public domain benchmark for citation parsing, sourced from PDFs released on four publishing ecosystems: SciELO, Redalyc, the Public Knowledge Project, and Open Research Europe. Starting from 161,000 annotated citations, we apply automated validation and feature-based sampling to produce a dataset of 10,000 citations spanning multiple languages, publication types, and platforms. We then evaluate a variety of citation parsing systems and report field-level precision and recall. Our results show strong performance from language models, particularly when fine-tuned. RenoBench enables reproducible, standardized evaluation of citation parsing systems, and provides a foundation for advancing automated citation parsing and metascientific research.

Through an analysis of arXiv papers, we report several shifts in word usage that are likely driven by large language models (LLMs) but have not previously received sufficient attention, such as the increased frequency of "beyond" and "via" in titles and the decreased frequency of "the" and "of" in abstracts. Due to the similarities among different LLMs, experiments show that current classifiers struggle to accurately determine which specific model generated a given text in multi-class classification tasks. Meanwhile, variations across LLMs also result in evolving patterns of word usage in academic papers. By adopting a direct and highly interpretable linear approach and accounting for differences between models and prompts, we quantitatively assess these effects and show that real-world LLM usage is heterogeneous and dynamic.

This paper introduces Conchordal, a bio-acoustic instrument for generative composition whose sonic agents are governed by artificial life dynamics within a psychoacoustic fitness landscape. The system is built on Direct Cognitive Coupling (DCC), a design principle requiring that generative dynamics operate directly within a landscape derived from psychoacoustic observables and read from that landscape without symbolic harmonic rules. The environment integrates roughness and harmonicity into a continuous consonance field without presupposing discrete scales or explicit harmonic rules. Agents adjust pitch through local proposal-and-accept dynamics under a crowding penalty, regulate survival via consonance-dependent metabolism, and entrain temporally through Kuramoto-style phase coupling. Four experiments are reported: (1) consonance search produces structured polyphony with enriched consonant intervals; (2) consonance-dependent metabolism yields survival differentials that vanish when recharge is disabled; (3) a minimal hereditary adaptation assay shows that parent-guided respawn plus metabolic selection can accumulate more structured polyphony without adult hill-climbing; and (4) a shared oscillatory scaffold organizes rhythmic timing under external forcing. A supplementary mechanism check reports one possible composer-configurable bridge by which spectral state can modulate temporal coupling. These findings show that a psychoacoustically derived landscape serves as an effective artificial-life terrain, yielding self-organization, selection, synchronization, and lineage-level accumulation in a non-traditional computational medium. At the level of the model, the same landscape therefore functions both as ecological terrain and as an internal proxy for musical coherence.

Air flow modeling at a local scale is essential for applications such as pollutant dispersion modeling or wind farm modeling. To circumvent costly Computational Fluid Dynamics (CFD) computations, deep learning surrogate models have recently emerged as promising alternatives. However, in the context of urban air flow, deep learning models struggle to adapt to the high variations of the urban geometry and to large mesh sizes. To tackle these challenges, we introduce Anchored Branched Steady-state WInd Flow Transformer (AB-SWIFT), a transformer-based model with an internal branched structure uniquely designed for atmospheric flow modeling. We train our model on a specially designed database of atmospheric simulations around randomised urban geometries and with a mixture of unstable, neutral, and stable atmospheric stratifications. Our model reaches the best accuracy on all predicted fields compared to state-of-the-art transformers and graph-based models. Our code and data is available at https://github.com/cerea-daml/abswift.

Large Language Models (LLMs) are increasingly used in math education not only as problem solvers but also as assessors of learners' reasoning. However, it remains unclear whether stronger math problem-solving ability is associated with stronger step-level assessment performance. This study examines that relationship using the GSM8K and MATH subsets of PROCESSBENCH, a human-annotated benchmark for identifying the earliest erroneous step in mathematical reasoning. We evaluate two LLM-based math tutor agent settings, instantiated with GPT-4 and GPT-5, in two independent tasks on the same math problems: solving the original problem and assessing a benchmark-provided solution by predicting the earliest erroneous step. Results show a consistent within-model pattern: assessment accuracy is substantially higher on math problem items the same model solved correctly than on items it solved incorrectly, with statistically significant associations across both models and datasets. At the same time, assessment remains more difficult than direct problem solving, especially on error-present solutions. These findings suggest that math problem-solving expertise supports stronger assessment performance, but reliable step-level diagnosis also requires additional capabilities such as step tracking, monitoring, and precise error localization. The results have implications for the design and evaluation of AI-supported Adaptive Instructional Systems (AISs) for formative assessment in math education.

While language reasoning models excel in many tasks, visual reasoning remains challenging for current large multimodal models (LMMs). As a result, most LMMs default to verbalizing perceptual content into text, a strong limitation for tasks requiring fine-grained spatial and visual understanding. While recent approaches take steps toward thinking with images by invoking tools or generating intermediate images, they either rely on external modules, or incur unnecessary computation by reasoning directly in pixel space. In this paper, we introduce LanteRn, a framework that enables LMMs to interleave language with compact latent visual representations, allowing visual reasoning to occur directly in latent space. LanteRn augments a vision-language transformer with the ability to generate and attend to continuous visual thought embeddings during inference. We train the model in two stages: supervised fine-tuning to ground visual features in latent states, followed by reinforcement learning to align latent reasoning with task-level utility. We evaluate LanteRn on three perception-centric benchmarks (VisCoT, V*, and Blink), observing consistent improvements in visual grounding and fine-grained reasoning. These results suggest that internal latent representations provide a promising direction for more efficient multimodal reasoning.

Explanations are central to improving transparency, trust, and user satisfaction in recommender systems (RS), yet it remains unclear how different explanation formats (visual vs. textual) are suited to users with different personal characteristics (PCs). To this end, we report a within-subject user study (n=54) comparing visual and textual explanations and examine how explanation format and PCs jointly influence perceived control, transparency, trust, and satisfaction in an educational recommender system (ERS). Using robust mixed-effects models, we analyze the moderating effects of a wide range of PCs, including Big Five traits, need for cognition, decision making style, visualization familiarity, and technical expertise. Our results show that a well-designed visual, simple, interactive, selective, easy to understand visualization that clearly and intuitively communicates how user preferences are linked to recommendations, fosters perceived control, transparency, appropriate trust, and satisfaction in the ERS for most users, independent of their PCs. Moreover, we derive a set of guidelines to support the effective design of explanations in ERSs.

We present an efficient algorithm for uniformly sampling from an arbitrary compact body $\mathcal{X} \subset \mathbb{R}^n$ from a warm start under isoperimetry and a natural volume growth condition. Our result provides a substantial common generalization of known results for convex bodies and star-shaped bodies. The complexity of the algorithm is polynomial in the dimension, the Poincaré constant of the uniform distribution on $\mathcal{X}$ and the volume growth constant of the set $\mathcal{X}$.

Large language model (LLM)-based persona agents are rapidly being adopted as scalable proxies for human participants across diverse domains. Yet there is no systematic method for verifying whether a persona agent's responses remain free of contradictions and factual inaccuracies throughout an interaction. A principle from interrogation methodology offers a lens: no matter how elaborate a fabricated identity, systematic interrogation will expose its contradictions. We apply this principle to propose PICon, an evaluation framework that probes persona agents through logically chained multi-turn questioning. PICon evaluates consistency along three core dimensions: internal consistency (freedom from self-contradiction), external consistency (alignment with real-world facts), and retest consistency (stability under repetition). Evaluating seven groups of persona agents alongside 63 real human participants, we find that even systems previously reported as highly consistent fail to meet the human baseline across all three dimensions, revealing contradictions and evasive responses under chained questioning. This work provides both a conceptual foundation and a practical methodology for evaluating persona agents before trusting them as substitutes for human participants. We provide the source code and an interactive demo at: https://kaist-edlab.github.io/picon/

Social learning highlights that learning agents improve not in isolation, but through interaction and structured knowledge exchange with others. When introduced into machine learning, this principle gives rise to social machine learning (SML), where multiple agents collaboratively learn by sharing abstracted knowledge. Federated learning (FL) provides a natural collaboration substrate for this paradigm, yet existing heterogeneous FL approaches often rely on sharing model parameters or intermediate representations, which may expose sensitive information and incur additional overhead. In this work, we propose SoHip (Social Hippocampus Memory Learning), a memory-centric social machine learning framework that enables collaboration among heterogeneous agents via memory sharing rather than model sharing. SoHip abstracts each agent's individual short-term memory from local representations, consolidates it into individual long-term memory through a hippocampus-inspired mechanism, and fuses it with collectively aggregated long-term memory to enhance local prediction. Throughout the process, raw data and local models remain on-device, while only lightweight memory are exchanged. We provide theoretical analysis on convergence and privacy preservation properties. Experiments on two benchmark datasets with seven baselines demonstrate that SoHip consistently outperforms existing methods, achieving up to 8.78% accuracy improvements.

Multimodal Large Language Models (MLLMs) have recently been explored as face verification systems that determine whether two face images are of the same person. Unlike dedicated face recognition systems, MLLMs approach this task through visual prompting and rely on general visual and reasoning abilities. However, the demographic fairness of these models remains largely unexplored. In this paper, we present a benchmarking study that evaluates nine open-source MLLMs from six model families, ranging from 2B to 8B parameters, on the IJB-C and RFW face verification protocols across four ethnicity groups and two gender groups. We measure verification accuracy with the Equal Error Rate and True Match Rate at multiple operating points per demographic group, and we quantify demographic disparity with four FMR-based fairness metrics. Our results show that FaceLLM-8B, the only face-specialised model in our study, substantially outperforms general-purpose MLLMs on both benchmarks. The bias patterns we observe differ from those commonly reported for traditional face recognition, with different groups being most affected depending on the benchmark and the model. We also note that the most accurate models are not necessarily the fairest and that models with poor overall accuracy can appear fair simply because they produce uniformly high error rates across all demographic groups.

The widespread adoption of CT has notably increased the number of detected lung nodules. However, current deep learning methods for classifying benign and malignant nodules often fail to comprehensively integrate global and local features, and most of them have not been validated through clinical trials. To address this, we developed DeepFAN, a transformer-based model trained on over 10K pathology-confirmed nodules and further conducted a multi-reader, multi-case clinical trial to evaluate its efficacy in assisting junior radiologists. DeepFAN achieved diagnostic area under the curve (AUC) of 0.939 (95% CI 0.930-0.948) on an internal test set and 0.954 (95% CI 0.934-0.973) on the clinical trial dataset involving 400 cases across three independent medical institutions. Explainability analysis indicated higher contributions from global than local features. Twelve readers' average performance significantly improved by 10.9% (95% CI 8.3%-13.5%) in AUC, 10.0% (95% CI 8.9%-11.1%) in accuracy, 7.6% (95% CI 6.1%-9.2%) in sensitivity, and 12.6% (95% CI 10.9%-14.3%) in specificity (P<0.001 for all). Nodule-level inter-reader diagnostic consistency improved from fair to moderate (overall k: 0.313 vs. 0.421; P=0.019). In conclusion, DeepFAN effectively assisted junior radiologists and may help homogenize diagnostic quality and reduce unnecessary follow-up of indeterminate pulmonary nodules. Chinese Clinical Trial Registry: ChiCTR2400084624.

Predicting high-dimensional dynamical systems with irregular time steps presents significant challenges for current data-driven algorithms. These irregularities arise from missing data, sparse observations, or adaptive computational techniques, reducing prediction accuracy. To address these limitations, we propose a novel method: a Physics-Spatiotemporal Masked Autoencoder. This method integrates convolutional autoencoders for spatial feature extraction with masked autoencoders optimised for irregular time series, leveraging attention mechanisms to reconstruct the entire physical sequence in a single prediction pass. The model avoids the need for data imputation while preserving physical integrity of the system. Here, 'physics' refers to high-dimensional fields generated by underlying dynamical systems, rather than the enforcement of explicit physical constraints or PDE residuals. We evaluate this approach on multiple simulated datasets and real-world ocean temperature data. The results demonstrate that our method achieves significant improvements in prediction accuracy, robustness to nonlinearities, and computational efficiency over traditional convolutional and recurrent network methods. The model shows potential for capturing complex spatiotemporal patterns without requiring domain-specific knowledge, with applications in climate modelling, fluid dynamics, ocean forecasting, environmental monitoring, and scientific computing.

Repairing faulty quantum circuits is challenging and requires automated solutions. We present QRep, an automated repair approach that iteratively identifies and repairs faults in a circuit. QRep uniformly applies patches across the circuit and assigns each gate a suspiciousness score, reflecting its likelihood of being faulty. It then narrows the search space by prioritising the most suspicious gates in subsequent iterations, increasing the repair efficiency. We evaluated QRep on 40 (real and synthetic) faulty circuits. QRep completely repaired 70% of them, and for the remaining circuits, the actual faulty gate was ranked within the top 44% most suspicious gates, demonstrating the effectiveness of QRep in fault localisation. Compared with two baseline approaches, QRep scales to larger and more complex circuits, up to 13 qubits.

A key capability of modern neural networks is their capacity to simultaneously learn underlying rules and memorize specific facts or exceptions. Yet, theoretical understanding of this dual capability remains limited. We introduce the Rules-and-Facts (RAF) model, a minimal solvable setting that enables precise characterization of this phenomenon by bridging two classical lines of work in the statistical physics of learning: the teacher-student framework for generalization and Gardner-style capacity analysis for memorization. In the RAF model, a fraction $1 - \varepsilon$ of training labels is generated by a structured teacher rule, while a fraction $\varepsilon$ consists of unstructured facts with random labels. We characterize when the learner can simultaneously recover the underlying rule - allowing generalization to new data - and memorize the unstructured examples. Our results quantify how overparameterization enables the simultaneous realization of these two objectives: sufficient excess capacity supports memorization, while regularization and the choice of kernel or nonlinearity control the allocation of capacity between rule learning and memorization. The RAF model provides a theoretical foundation for understanding how modern neural networks can infer structure while storing rare or non-compressible information.

Accurate biodiversity identification from large-scale field data is a foundational problem with direct impact on ecology, conservation, and environmental monitoring. In practice, the core task is taxonomic prediction - inferring order, family, genus, or species from imperfect inputs such as specimen images, DNA barcodes, or both. Existing multimodal methods often treat taxonomy as a flat label space and therefore fail to encode the hierarchical structure of biological classification, which is critical for robustness under noise and missing modalities. We present two end-to-end variants for hierarchy-aware multimodal learning: CLiBD-HiR, which introduces Hierarchical Information Regularization (HiR) to shape embedding geometry across taxonomic levels, yielding structured and noise-robust representations; and CLiBD-HiR-Fuse, which additionally trains a lightweight fusion predictor that supports image-only, DNA-only, or joint inference and is resilient to modality corruption. Across large-scale biodiversity benchmarks, our approach improves taxonomic classification accuracy by over 14 percent compared to strong multimodal baselines, with particularly large gains under partial and corrupted DNA conditions. These results highlight that explicitly encoding biological hierarchy, together with flexible fusion, is key for practical biodiversity foundation models.

The deployment of reconfigurable intelligent surfaces (RISs) introduces new challenges for resource allocation in multi-cell wireless networks, particularly when user loads are uneven across base stations. In this work, we consider RISs as shared infrastructure that must be dynamically assigned among competing base stations, and we address this problem using a simultaneous ascending auction mechanism. To mitigate performance imbalances between cells, we propose a fairness-aware collaborative multi-agent reinforcement learning approach in which base stations adapt their bidding strategies based on both expected utility gains and relative service quality. A centrally computed performance-dependent fairness indicator is incorporated into the agents' observations, enabling implicit coordination without direct inter-base-station communication. Simulation results show that the proposed framework effectively redistributes RIS resources toward weaker-performing cells, substantially improving the rates of the worst-served users while preserving overall throughput. The results demonstrate that fairness-oriented RIS allocation can be achieved through cooperative learning, providing a flexible tool for balancing efficiency and equity in future wireless networks.

With the emergence of AI techniques for depression diagnosis, the conflict between high demand and limited supply for depression screening has been significantly alleviated. Among various modal data, audio-based depression diagnosis has received increasing attention from both academia and industry since audio is the most common carrier of emotion transmission. Unfortunately, audio data also contains User-sensitive Identity Information (ID), which is extremely vulnerable and may be maliciously used during the smart diagnosis process. Among previous methods, the clarification between depression features and sensitive features has always serve as a barrier. It is also critical to the problem for introducing a safe encryption methodology that only encrypts the sensitive features and a powerful classifier that can correctly diagnose the depression. To track these challenges, by leveraging adversarial loss-based Subspace Decomposition, we propose a first practical framework \name presented for Trustable Audio Affective Computing, to perform automated depression detection through audio within a trustable environment. The key enablers of TAAC are Differentiating Features Subspace Decompositor (DFSD), Flexible Noise Encryptor (FNE) and Staged Training Paradigm, used for decomposition, ID encryption and performance enhancement, respectively. Extensive experiments with existing encryption methods demonstrate our framework's preeminent performance in depression detection, ID reservation and audio reconstruction. Meanwhile, the experiments across various setting demonstrates our model's stability under different encryption strengths. Thus proving our framework's excellence in Confidentiality, Accuracy, Traceability, and Adjustability.

Translating natural language to SQL for data retrieval has become more accessible thanks to code generation LLMs. But how hard is it to generate SQL code? While databases can become unbounded in complexity, the complexity of queries is bounded by real life utility and human needs. With a sample of 376 databases, we show that SQL queries, as translations of natural language questions are finite in practical complexity. There is no clear monotonic relationship between increases in database table count and increases in complexity of SQL queries. In their template forms, SQL queries follow a Power Law-like distribution of frequency where 70% of our tested queries can be covered with just 13% of all template types, indicating that the high majority of SQL queries are predictable. This suggests that while LLMs for code generation can be useful, in the domain of database access, they may be operating in a narrow, highly formulaic space where templates could be safer, cheaper, and auditable.

On-policy distillation (OPD) is appealing for large language model (LLM) post-training because it evaluates teacher feedback on student-generated rollouts rather than fixed teacher traces. In long-horizon settings, however, the common sampled-token variant is fragile: it reduces distribution matching to a one-token signal and becomes increasingly unreliable as rollouts drift away from prefixes the teacher commonly visits. We revisit OPD from the estimator and implementation sides. Theoretically, token-level OPD is biased relative to sequence-level reverse-KL, but it has a much tighter worst-case variance bound; our toy study shows the same tradeoff empirically, with stronger future-reward coupling producing higher gradient variance and less stable learning. Empirically, we identify three failure modes of sampled-token OPD: an imbalanced one-token signal, unreliable teacher guidance on student-generated prefixes, and distortions caused by tokenizer or special-token mismatch. We address these issues with teacher top-K local support matching, implemented as truncated reverse-KL with top-p rollout sampling and special-token masking. Across single-task math reasoning and multi-task agentic-plus-math training, this objective yields more stable optimization and better downstream performance than sampled-token OPD.

Saccharomyces cerevisiae is a cornerstone organism in industrial biotechnology, valued for its genetic tractability and robust fermentative capacity. Accurately predicting biomass flux across diverse environmental and genetic perturbations remains a significant challenge for rational strain design. We present a computational framework combining the Yeast9 genome-scale metabolic model with machine learning and optimization to predict, interpret, and enhance biomass flux. Flux balance analysis generated 2,000 flux profiles by varying glucose, oxygen, and ammonium uptake rates. Random Forest and XGBoost regressors achieved R2 of 0.99989 and 0.9990, respectively. A variational autoencoder revealed four distinct metabolic clusters, and SHAP analysis identified glycolysis, the TCA cycle, and lipid biosynthesis as key biomass determinants. In silico overexpression achieved a biomass flux of 0.979 gDW/hr, while Bayesian optimization of nutrient constraints produced a 12-fold increase (0.0858 to 1.041 gDW/hr). A generative adversarial network proposed stoichiometrically feasible novel flux configurations. This framework demonstrates how genome-scale simulation, interpretable ML, and generative modeling can advance yeast metabolic engineering.

Non-fixed flexible antenna architectures, such as fluid antenna system (FAS), movable antenna (MA), and pinching antenna, have garnered significant interest in recent years. Among them, rotatable antenna (RA) has emerged as a promising technology for enhancing wireless communication and sensing performance through flexible antenna orientation/boresight rotation. By enabling mechanical or electronic boresight adjustment without altering physical antenna positions, RA introduces additional spatial degrees of freedom (DoFs) beyond conventional beamforming. In this paper, we provide a comprehensive tutorial on the fundamentals, architectures, and applications of RA-empowered wireless networks. Specifically, we begin by reviewing the historical evolution of RA-related technologies and clarifying the distinctive role of RA among flexible antenna architectures. Then, we establish a unified mathematical framework for RA-enabled systems, including general antenna/array rotation models, as well as channel models that cover near- and far-field propagation characteristics, wideband frequency selectivity, and polarization effects. Building upon this foundation, we investigate antenna/array rotation optimization in representative communication and sensing scenarios. Furthermore, we examine RA channel estimation/acquisition strategies encompassing orientation scheduling mechanisms and signal processing methods that exploit multi-view channel observations. Beyond theoretical modeling and algorithmic design, we discuss practical RA configurations and deployment strategies. We also present recent RA prototypes and experimental results that validate the practical performance gains enabled by antenna rotation. Finally, we highlight promising extensions of RA to emerging wireless paradigms and outline open challenges to inspire future research.

We introduce Voxtral TTS, an expressive multilingual text-to-speech model that generates natural speech from as little as 3 seconds of reference audio. Voxtral TTS adopts a hybrid architecture that combines auto-regressive generation of semantic speech tokens with flow-matching for acoustic tokens. These tokens are encoded and decoded with Voxtral Codec, a speech tokenizer trained from scratch with a hybrid VQ-FSQ quantization scheme. In human evaluations conducted by native speakers, Voxtral TTS is preferred for multilingual voice cloning due to its naturalness and expressivity, achieving a 68.4\% win rate over ElevenLabs Flash v2.5. We release the model weights under a CC BY-NC license.

Automated incident management is critical for microservice reliability. While recent unified frameworks leverage multimodal data for joint optimization, they unrealistically assume perfect data completeness. In practice, network fluctuations and agent failures frequently cause missing modalities. Existing approaches relying on static placeholders introduce imputation noise that masks anomalies and degrades performance. To address this, we propose ARMOR, a robust self-supervised framework designed for missing modality scenarios. ARMOR features: (i) a modality-specific asymmetric encoder that isolates distribution disparities among metrics, logs, and traces; and (ii) a missing-aware gated fusion mechanism utilizing learnable placeholders and dynamic bias compensation to prevent cross-modal interference from incomplete inputs. By employing self-supervised auto-regression with mask-guided reconstruction, ARMOR jointly optimizes anomaly detection (AD), failure triage (FT), and root cause localization (RCL). AD and RCL require no fault labels, while FT relies solely on failure-type annotations for the downstream classifier. Extensive experiments demonstrate that ARMOR achieves state-of-the-art performance under complete data conditions and maintains robust diagnostic accuracy even with severe modality loss.

We study narrative coherence in visually grounded stories by comparing human-written narratives with those generated by vision-language models (VLMs) on the Visual Writing Prompts corpus. Using a set of metrics that capture different aspects of narrative coherence, including coreference, discourse relation types, topic continuity, character persistence, and multimodal character grounding, we compute a narrative coherence score. We find that VLMs show broadly similar coherence profiles that differ systematically from those of humans. In addition, differences for individual measures are often subtle, but they become clearer when considered jointly. Overall, our results indicate that, despite human-like surface fluency, model narratives exhibit systematic differences from those of humans in how they organise discourse across a visually grounded story. Our code is available at https://github.com/GU-CLASP/coherence-driven-humans.

To date, there is little research on how to design back marks to best support individual-level monitoring of uniform looking species like pigs. With the recent surge of machine learning-based monitoring solutions, there is a particular need for guidelines on the design of marks that can be effectively recognised by such algorithms. This study provides valuable insights on effective back mark design, based on the analysis of a machine learning model, trained to distinguish pigs via their back marks. Specifically, a neural network of type ResNet-50 was trained to classify ten pigs with unique back marks. The analysis of the model's predictions highlights the significance of certain design choices, even in controlled settings. Most importantly, the set of back marks must be designed such that each mark remains unambiguous under conditions of motion blur, diverse view angles and occlusions, caused by animal behaviour. Further, the back mark design must consider data augmentation strategies commonly employed during model training, like colour, flip and crop augmentations. The generated insights can support individual-level monitoring in future studies and real-world applications by optimizing back mark design.

Many Cyber Physical System (CPS) work in a safety-critical environment, where correct execution, reliability and trustworthiness are essential. Signal Temporal Logic (STL) provides a formal framework for checking safety-critical CPS. However, static verification of STL is undecidable in general, except when we want to verify using run-time-based methods, which have limitations. We propose Synchronous Signal Temporal Logic (SSTL), a decidable fragment of STL, which admits static safety and liveness property verification. In SSTL, we assume that a signal is sampled at fixed discrete steps, called ticks, and then propose a hypothesis, called the Signal Invariance Hypothesis (SIH), which is inspired by a similar hypothesis for synchronous programs. We define the syntax and semantics of SSTL and show that SIH is a necessary and sufficient condition for equivalence between an STL formula and its SSTL counterpart. By translating SSTL to LTL_P (LTL defined over predicates), we enable decidable model checking using the SPIN model checker. We demonstrate the approach on a 33-node human heart model and other case studies.

Long-term behavioral monitoring of individual animals is crucial for studying behavioral changes that occur over different time scales, especially for conservation and evolutionary biology. Computer vision methods have proven to benefit biodiversity monitoring, but automated behavior monitoring in wild populations remains challenging. This stems from the lack of datasets that cover a range of computer vision tasks necessary to extract biologically meaningful measurements of individual animals. Here, we introduce such a dataset (CHIRP) with a new method (CORVID) for individual re-identification of wild birds. The CHIRP (Combining beHaviour, Individual Re-identification and Postures) dataset is curated from a long-term population of wild Siberian jays studied in Swedish Lapland, supporting re-identification (re-id), action recognition, 2D keypoint estimation, object detection, and instance segmentation. In addition to traditional task-specific benchmarking, we introduce application-specific benchmarking with biologically relevant metrics (feeding rates, co-occurrence rates) to evaluate the performance of models in real-world use cases. Finally, we present CORVID (COlouR-based Video re-ID), a novel pipeline for individual identification of birds based on the segmentation and classification of colored leg rings, a widespread approach for visual identification of individual birds. CORVID offers a probability-based id tracking method by matching the detected combination of color rings with a database. We use application-specific benchmarking to show that CORVID outperforms state-of-the-art re-id methods. We hope this work offers the community a blueprint for curating real-world datasets from ethically approved biological studies to bridge the gap between computer vision research and biological applications.

Neuroevolution automates the complex task of neural network design but often ignores the inherent adversarial fragility of evolved models which is a barrier to adoption in safety-critical scenarios. While robust training methods have received significant attention, the design of architectures exhibiting intrinsic robustness remains largely unexplored. In this paper, we propose NERO-Net, a neuroevolutionary approach to design convolutional neural networks better equipped to resist adversarial attacks. Our search strategy isolates architectural influence on robustness by avoiding adversarial training during the evolutionary loop. As such, our fitness function promotes candidates that, even trained with standard (non-robust) methods, achieve high post-attack accuracy without sacrificing the accuracy on clean samples. We assess NERO-Net on CIFAR-10 with a specific focus on $L_\infty$-robustness. In particular, the fittest individual emerged from evolutionary search with 33% accuracy against FGSM, used as an efficient estimator for robustness during the search phase, while maintaining 87% clean accuracy. Further standard training of this individual boosted these metrics to 47% adversarial and 93% clean accuracy, suggesting inherent architectural robustness. Adversarial training brings the overall accuracy of the model up to 40% against AutoAttack.

The use of hyperspectral imaging (HSI) in autonomous driving (AD), while promising, faces many challenges related to the specifics and requirements of this application domain. On the one hand, non-controlled and variable lighting conditions, the wide depth-of-field ranges, and dynamic scenes with fast-moving objects. On the other hand, the requirements for real-time operation and the limited computational resources of embedded platforms. The combination of these factors determines both the criteria for selecting appropriate HSI technologies and the development of custom vision algorithms that leverage the spectral and spatial information obtained from the sensors. In this article, we analyse several techniques explored in the research of HSI-based vision systems with application to AD, using as an example results obtained from experiments using data from the most recent version of the HSI-Drive dataset.

We propose a method for constructing distribution-free prediction intervals in nonparametric instrumental variable regression (NPIV), with finite-sample coverage guarantees. Building on the conditional guarantee framework in conformal inference, we reformulate conditional coverage as marginal coverage over a class of IV shifts $\mathcal{F}$. Our method can be combined with any NPIV estimator, including sieve 2SLS and other machine-learning-based NPIV methods such as neural networks minimax approaches. Our theoretical analysis establishes distribution-free, finite-sample coverage over a practitioner-chosen class of IV shifts.

Accurate Network Traffic Classification (NTC) is increasingly constrained by limited labeled data and strict privacy requirements. While Network Traffic Generation (NTG) provides an effective means to mitigate data scarcity, conventional generative methods struggle to model the complex temporal dynamics of modern traffic or/and often incur significant computational cost. In this article, we address the NTG task using lightweight Generative Artificial Intelligence (GenAI) architectures, including transformer-based, state-space, and diffusion models designed for practical deployment. We conduct a systematic evaluation along four axes: (i) (synthetic) traffic fidelity, (ii) synthetic-only training, (iii) data augmentation under low-data regimes, and (iv) computational efficiency. Experiments on two heterogeneous datasets show that lightweight GenAI models preserve both static and temporal traffic characteristics, with transformer and state-space models closely matching real distributions across a complete set of fidelity metrics. Classifiers trained solely on synthetic traffic achieve up to 87% F1-score on real data. In low-data settings, GenAI-driven augmentation improves NTC performance by up to +40%, substantially reducing the gap with full-data training. Overall, transformer-based models provide the best trade-off between fidelity and efficiency, enabling high-quality, privacy-aware traffic synthesis with modest computational overhead.

In recent years, fake news detection has received increasing attention in public debate and scientific research. Despite advances in detection techniques, the production and spread of false information have become more sophisticated, driven by Large Language Models (LLMs) and the amplification power of social media. We present a critical assessment of 12 representative fake news detection approaches, spanning traditional machine learning, deep learning, transformers, and specialized cross-domain architectures. We evaluate these methods on 10 publicly available datasets differing in genre, source, topic, and labeling rationale. We address text-only English fake news detection as a binary classification task by harmonizing labels into "Real" and "Fake" to ensure a consistent evaluation protocol. We acknowledge that label semantics vary across datasets and that harmonization inevitably removes such semantic nuances. Each dataset is treated as a distinct domain. We conduct in-domain, multi-domain and cross-domain experiments to simulate real-world scenarios involving domain shift and out-of-distribution data. Fine-tuned models perform well in-domain but struggle to generalize. Cross-domain architectures can reduce this gap but are data-hungry, while LLMs offer a promising alternative through zero- and few-shot learning. Given inherent dataset confounds and possible pre-training exposure, results should be interpreted as robustness evaluations within this English, text-only protocol.

Object detectors deployed in safety-critical environments can fail silently, e.g. missing pedestrians, workers, or other safety-critical objects without emitting any warning. Traditional Out Of Distribution (OOD) detection methods focus on identifying unfamiliar inputs, but do not directly predict functional failures of the detector itself. We introduce Knowledge Guided Failure Prediction (KGFP), a representation-based monitoring framework that treats missed safety-critical detections as anomalies to be detected at runtime. KGFP measures semantic misalignment between internal object detector features and visual foundation model embeddings using a dual-encoder architecture with an angular distance metric. A key property is that when either the detector is operating outside its competence or the visual foundation model itself encounters novel inputs, the two embeddings diverge, producing a high-angle signal that reliably flags unsafe images. We compare our novel KGFS method to baseline OOD detection methods. On COCO person detection, applying KGFP as a selective-prediction gate raises person recall among accepted images from 64.3% to 84.5% at 5% False Positive Rate (FPR), and maintains strong performance across six COCO-O visual domains, outperforming OOD baselines by large margins. Our code, models, and features are published at https://gitlab.cc-asp.fraunhofer.de/iosb_public/KGFP.

As the Web transitions from static retrieval to generative interaction, the escalating environmental footprint of Large Language Models (LLMs) presents a critical sustainability challenge. Current paradigms indiscriminately apply computation-intensive strategies like Chain-of-Thought (CoT) to billions of daily queries, causing LLM overthinking, a redundancy that amplifies carbon emissions and operational barriers. This inefficiency directly undermines UN Sustainable Development Goals 13 (Climate Action) and 10 (Reduced Inequalities) by hindering equitable AI access in resource-constrained regions. To address this, we introduce EcoThink, an energy-aware adaptive inference framework designed to reconcile high-performance AI intelligence with environmental responsibility. EcoThink employs a lightweight, distillation-based router to dynamically assess query complexity, skipping unnecessary reasoning for factoid retrieval while reserving deep computation for complex logic. Extensive evaluations across 9 diverse benchmarks demonstrate that EcoThink reduces inference energy by 40.4% on average (up to 81.9% for web knowledge retrieval) without statistically significant performance loss. By mitigating algorithmic waste, EcoThink offers a scalable path toward a sustainable, inclusive, and energy-efficient generative AI Agent.

Accurate short-term air-quality forecasting is essential for public health protection and urban management, yet many recent forecasting frameworks rely on complex, data-intensive, and computationally demanding models. This study investigates whether lightweight and interpretable forecasting approaches can provide competitive performance for hourly PM2.5 prediction in Beijing, China. Using multi-year pollutant and meteorological time-series data, we developed a leakage-aware forecasting workflow that combined chronological data partitioning, preprocessing, feature selection, and exogenous-driver modeling under the Perfect Prognosis setting. Three forecasting families were evaluated: SARIMAX, Facebook Prophet, and NeuralProphet. To assess practical deployment behavior, the models were tested under two adaptive regimes: weekly walk-forward refitting and frozen forecasting with online residual correction. Results showed clear differences in both predictive accuracy and computational efficiency. Under walk-forward refitting, Facebook Prophet achieved the strongest completed performance, with an MAE of $37.61$ and an RMSE of $50.10$, while also requiring substantially less execution time than NeuralProphet. In the frozen-model regime, online residual correction improved Facebook Prophet and SARIMAX, with corrected SARIMAX yielding the lowest overall error (MAE $32.50$; RMSE $46.85$). NeuralProphet remained less accurate and less stable across both regimes, and residual correction did not improve its forecasts. Notably, corrected Facebook Prophet reached nearly the same error as its walk-forward counterpart while reducing runtime from $15$ min $21.91$ sec to $46.60$ sec. These findings show that lightweight additive forecasting strategies can remain highly competitive for urban air-quality prediction, offering a practical balance between accuracy, interpretability, ...

Recent strategies for low-resource machine translation rely on LLMs to generate synthetic data from higher-resource languages. We find that this method fails for Romansh, because LLMs tend to confuse its 6 distinct language varieties. Our experiments show that instead, the direction of data augmentation should be aligned with the resource gradient between source and target language. This approach surpasses Gemini 3 Pro in the lowest-resource variety of Romansh by 23 BLEU. A human evaluation confirms that our experiments yield the first model that generates fluent translations in the individual Romansh varieties.

Migrating stateful microservices in Kubernetes requires careful state management because in-memory state is lost when a container restarts. For StatefulSet-managed workloads, the problem is compounded by identity constraints that prohibit two pods with the same ordinal from running simultaneously, forcing a sequential stop-recreate cycle with a median 38.5s of service downtime. This paper presents SHADOW Seamless Handoff And Zero-Downtime Orchestrated Workload Migration, a Kubernetes-native framework that implements the Message-based Stateful Microservice Migration (MS2M) approach as a Kubernetes Operator. SHADOW introduces the ShadowPod strategy, where a shadow pod is created from a CRIU checkpoint image on the target node while the source pod continues serving traffic, allowing concurrent operation during message replay. For StatefulSet workloads, an identity swap procedure with the ExchangeFence mechanism re-checkpoints the shadow pod, creates a StatefulSet-owned replacement, and drains both message queues to guarantee zero message loss during the handoff. An evaluation on a bare-metal Kubernetes cluster with 280 migration runs across four configurations and seven message rates (10--120msg/s) shows that, compared to the sequential baseline on the same StatefulSet workload, the ShadowPod strategy reduces the restore phase by up to 92%, eliminates service downtime entirely, and reduces total migration time by up to 77%, with zero message loss across all 280 runs.

Model retraining is usually treated as an ongoing maintenance task. But as Harrison Katz now argues, retraining can be better understood as approximate Bayesian inference under computational constraints. The gap between a continuously updated belief state and your frozen deployed model is "learning debt," and the retraining decision is a cost minimization problem with a threshold that falls out of your loss function. In this article Katz provides a decision-theoretic framework for retraining policies. The result is evidence-based triggers that replace calendar schedules and make governance auditable. For readers less familiar with the Bayesian and decision-theoretic language, key terms are defined in a glossary at the end of the article.

Transfer learning is a crucial concept within deep learning that allows artificial neural networks to benefit from a large pre-training data basis when confronted with a task of limited data. Despite its ubiquitous use and clear benefits, there are still many open questions regarding the inner workings of transfer learning and, in particular, regarding the understanding of when and how well it works. To that extent, we perform a rigorous study focusing on audio-to-audio transfer learning, in which we pre-train various model states on (ontology-based) subsets of AudioSet and fine-tune them on three computer audition tasks, namely acoustic scene recognition, bird activity recognition, and speech command recognition. We report that increasing the number of samples and classes in the pre-training data both have a positive impact on transfer learning. This is, however, generally surpassed by similarity between pre-training and the downstream task, which can lead the model to learn comparable features.

Understanding directed temporal interactions in multivariate time series is essential for interpreting complex dynamical systems and the predictive models trained on them. We present Causal-INSIGHT, a model-agnostic, post-hoc interpretation framework for extracting model-implied (predictor-dependent), directed, time-lagged influence structure from trained temporal predictors. Rather than inferring causal structure at the level of the data-generating process, Causal-INSIGHT analyzes how a fixed, pre-trained predictor responds to systematic, intervention-inspired input clamping applied at inference time. From these responses, we construct directed temporal influence signals that reflect the dependencies the predictor relies on for prediction, and introduce Qbic, a sparsity-aware graph selection criterion that balances predictive fidelity and structural complexity without requiring ground-truth graph labels. Experiments across synthetic, simulated, and realistic benchmarks show that Causal-INSIGHT generalizes across diverse backbone architectures, maintains competitive structural accuracy, and yields significant improvements in temporal delay localization when applied to existing predictors.

A growing body of literature has focused on predicting wildfire occurrence using machine learning methods, capitalizing on high-resolution data and fire predictors that canonical process-based frameworks largely ignore. Standard evaluation metrics for an ML classifier, while important, provide a potentially limited measure of the model's operational performance for the Fire Danger Index (FDI) forecast. Furthermore, model evaluation is frequently conducted without adequately accounting for false positive rates, despite their critical relevance in operational contexts. In this paper, we revisit the daily FDI model evaluation paradigm and propose a novel method for evaluating a forest fire forecasting model that is aligned with real-world decision-making. Furthermore, we systematically assess performance in accurately predicting fire activity and the false positives (false alarms). We further demonstrate that an ensemble of ML models improves both fire identification and reduces false positives.

We study statistical estimation in a student--teacher setting, where predictions from a pre-trained teacher are used to guide a student model. A standard approach is to train the student to directly match the teacher's outputs, which we refer to as student soft matching (SM). This approach directly propagates any systematic bias or mis-specification present in the teacher, thereby degrading the student's predictions. We propose and analyze an alternative scheme, known as residual-as-teacher (RaT), in which the teacher is used to estimate residuals in the student's predictions. Our analysis shows how the student can thereby emulate a proximal gradient scheme for solving an oracle optimization problem, and this provably reduces the effect of teacher bias. For general student--teacher pairs, we establish non-asymptotic excess risk bounds for any RaT fixed point, along with convergence guarantees for the student-teacher iterative scheme. For kernel-based student--teacher pairs, we prove a sharp separation: the RaT method achieves the minimax-optimal rate, while the SM method incurs constant prediction error for any sample size. Experiments on both synthetic data and ImageNette classification under covariate shift corroborate our theoretical findings.

Zero-shot reinforcement learning (RL) algorithms aim to learn a family of policies from a reward-free dataset, and recover optimal policies for any reward function directly at test time. Naturally, the quality of the pretraining dataset determines the performance of the recovered policies across tasks. However, pre-collecting a relevant, diverse dataset without prior knowledge of the downstream tasks of interest remains a challenge. In this work, we study $\textit{online}$ zero-shot RL for quadrupedal control on real robotic systems, building upon the Forward-Backward (FB) algorithm. We observe that undirected exploration yields low-diversity data, leading to poor downstream performance and rendering policies impractical for direct hardware deployment. Therefore, we introduce FB-MEBE, an online zero-shot RL algorithm that combines an unsupervised behavior exploration strategy with a regularization critic. FB-MEBE promotes exploration by maximizing the entropy of the achieved behavior distribution. Additionally, a regularization critic shapes the recovered policies toward more natural and physically plausible behaviors. We empirically demonstrate that FB-MEBE achieves and improved performance compared to other exploration strategies in a range of simulated downstream tasks, and that it renders natural policies that can be seamlessly deployed to hardware without further finetuning. Videos and code available on our website.

Motion planning in dynamic urban environments requires balancing immediate safety with long-term goals. While diffusion models effectively capture multi-modal decision-making, existing approaches treat trajectories as monolithic entities, overlooking heterogeneous temporal dependencies where near-term plans are constrained by instantaneous dynamics and far-term plans by navigational goals. To address this, we propose Temporally Decoupled Diffusion Model (TDDM), which reformulates trajectory generation via a noise-as-mask paradigm. By partitioning trajectories into segments with independent noise levels, we implicitly treat high noise as information voids and weak noise as contextual cues. This compels the model to reconstruct corrupted near-term states by leveraging internal correlations with better-preserved temporal contexts. Architecturally, we introduce a Temporally Decoupled Adaptive Layer Normalization (TD-AdaLN) to inject segment-specific timesteps. During inference, our Asymmetric Temporal Classifier-Free Guidance utilizes weakly noised far-term priors to guide immediate path generation. Evaluations on the nuPlan benchmark show TDDM approaches or exceeds state-of-the-art baselines, particularly excelling in the challenging Test14-hard subset.

Detecting when a language model is wrong without ground truth labels is a fundamental challenge for safe deployment. Existing approaches rely on a model's own uncertainty -- such as token entropy or confidence scores -- but these signals fail critically on the most dangerous failure mode: confident errors, where a model is wrong but certain. In this work we introduce cross-model disagreement as a correctness indicator -- a simple, training-free signal that can be dropped into existing production systems, routing pipelines, and deployment monitoring infrastructure without modification. Given a model's generated answer, cross-model disagreement computes how surprised or uncertain a second verifier model is when reading that answer via a single forward pass. No generation from the verifying model is required, and no correctness labels are needed. We instantiate this principle as Cross-Model Perplexity (CMP), which measures the verifying model's surprise at the generating model's answer tokens, and Cross-Model Entropy (CME), which measures the verifying model's uncertainty at those positions. Both CMP and CME outperform within-model uncertainty baselines across benchmarks spanning reasoning, retrieval, and mathematical problem solving (MMLU, TriviaQA, and GSM8K). On MMLU, CMP achieves a mean AUROC of 0.75 against a within-model entropy baseline of 0.59. These results establish cross-model disagreement as a practical, training-free approach to label-free correctness estimation, with direct applications in deployment monitoring, model routing, selective prediction, data filtering, and scalable oversight of production language model systems.

We introduce and solve a teacher-student formulation of the symmetric binary Perceptron, turning a traditionally storage-oriented model into a planted inference problem with a guaranteed solution at any sample density. We adapt the formulation of the symmetric Perceptron which traditionally considers either the u-shaped potential or the rectangular one, by including labels in both regions. With this formulation, we analyze both the Bayes-optimal regime at for noise-less examples and the effect of thermal noise under two different potential/classification rules. Using annealed and quenched free-entropy calculations in the high-dimensional limit, we map the phase diagram in the three control parameters, namely the sample density $α$, the distance between the origin and one of the symmetric hyperplanes $κ$ and temperature $T$, and identify a robust scenario where learning is organized by a second-order instability that creates teacher-correlated suboptimal states, followed by a first-order transition to full alignment. We show how this structure depends on the choice of potential, the interplay between metastability of the suboptimal solution and its melting towards the planted configuration, which is relevant for Monte Carlo-based optimization algorithms.

The rise of micro-videos has reshaped how misinformation spreads, amplifying its speed, reach, and impact on public trust. Existing benchmarks typically focus on a single deception type, overlooking the diversity of real-world cases that involve multimodal manipulation, AI-generated content, cognitive bias, and out-of-context reuse. Meanwhile, most detection models lack fine-grained attribution, limiting interpretability and practical utility. To address these gaps, we introduce WildFakeBench, a large-scale benchmark of over 10,000 real-world micro-videos covering diverse misinformation types and sources, each annotated with expert-defined attribution labels. Building on this foundation, we develop FakeAgent, a Delphi-inspired multi-agent reasoning framework that integrates multimodal understanding with external evidence for attribution-grounded analysis. FakeAgent jointly analyzes content and retrieved evidence to identify manipulation, recognize cognitive and AI-generated patterns, and detect out-of-context misinformation. Extensive experiments show that FakeAgent consistently outperforms existing MLLMs across all misinformation types, while WildFakeBench provides a realistic and challenging testbed for advancing explainable micro-video misinformation detection. Data and code are available at: https://github.com/Aiyistan/FakeAgent.

Recent developments in text classification using Large Language Models (LLMs) in the social sciences suggest that costs can be cut significantly, while performance can sometimes rival existing computational methods. However, with a wide variance in performance in current tests, we move to the question of how to maximize performance. In this paper, we focus on prompt context as a possible avenue for increasing accuracy by systematically varying three aspects of prompt engineering: label descriptions, instructional nudges, and few shot examples. Across two different examples, our tests illustrate that a minimal increase in prompt context yields the highest increase in performance, while further increases in context only tend to yield marginal performance increases thereafter. Alarmingly, increasing prompt context sometimes decreases accuracy. Furthermore, our tests suggest substantial heterogeneity across models, tasks, and batch size, underlining the need for individual validation of each LLM coding task rather than reliance on general rules.

Large Language Models (LLMs) have demonstrated remarkable proficiency in English mathematical reasoning, yet a significant performance disparity persists in multilingual contexts, largely attributed to deficiencies in language understanding. To bridge this gap, we introduce Translation-Augmented Policy Optimization (TAPO), a novel reinforcement learning framework built upon GRPO. TAPO enforces an explicit alignment strategy where the model leverages English as a pivot and follows an understand-then-reason paradigm. Crucially, we employ a step-level relative advantage mechanism that decouples understanding from reasoning, allowing the integration of translation quality rewards without introducing optimization conflicts. Extensive experiments reveal that TAPO effectively synergizes language understanding with reasoning capabilities and is compatible with various models. It outperforms baseline methods in both multilingual mathematical reasoning and translation tasks, while generalizing well to unseen languages and out-of-domain tasks.

Semantic world models enable embodied agents to reason about objects, relations, and spatial context beyond purely geometric representations. In Organic Computing, such models are a key enabler for objective-driven self-adaptation under uncertainty and resource constraints. The core challenge is to acquire observations maximising model quality and downstream usefulness within a limited action budget. Semantic scene graphs (SSGs) provide a structured and compact representation for this purpose. However, constructing them within a finite action horizon requires exploration strategies that trade off information gain against navigation cost and decide when additional actions yield diminishing returns. This work presents a modular navigation component for Embodied Semantic Scene Graph Generation and modernises its decision-making by replacing the policy-optimisation method and revisiting the discrete action formulation. We study compact and finer-grained, larger discrete motion sets and compare a single-head policy over atomic actions with a factorised multi-head policy over action components. We evaluate curriculum learning and optional depth-based collision supervision, and assess SSG completeness, execution safety, and navigation behaviour. Results show that replacing the optimisation algorithm alone improves SSG completeness by 21\% relative to the baseline under identical reward shaping. Depth mainly affects execution safety (collision-free motion), while completeness remains largely unchanged. Combining modern optimisation with a finer-grained, factorised action representation yields the strongest overall completeness--efficiency trade-off.

A prevailing assumption in machine learning is that model correctness must be enforced after the fact. We observe that the properties determining whether an AI model is numerically stable, computationally correct, or consistent with a physical domain do not necessarily demand post hoc enforcement. They can be verified at design time, before training begins, at marginal computational cost, with particular relevance to models deployed in high-leverage decision support and scientifically constrained settings. These properties share a specific algebraic structure: they are expressible as constraints over finitely generated abelian groups $\mathbb{Z}^n$, where inference is decidable in polynomial time and the principal type is unique. A framework built on this observation composes three prior results (arXiv:2603.16437, arXiv:2603.17627, arXiv:2603.18104): a dimensional type system carrying arbitrary annotations as persistent codata through model elaboration; a program hypergraph that infers Clifford algebra grade and derives geometric product sparsity from type signatures alone; and an adaptive domain model architecture preserving both invariants through training via forward-mode coeffect analysis and exact posit accumulation. We believe this composition yields a novel information-theoretic result: Hindley-Milner unification over abelian groups computes the maximum a posteriori hypothesis under a computable restriction of Solomonoff's universal prior, placing the framework's type inference on the same formal ground as universal induction. We compare four contemporary approaches to AI reliability and show that each imposes overhead that can compound across deployments, layers, and inference requests. This framework eliminates that overhead by construction.

Large language models (LLMs) increasingly rely on explicit chain-of-thought (CoT) reasoning to solve complex tasks, yet the safety of the reasoning process itself remains largely unaddressed. Existing work on LLM safety focuses on content safety--detecting harmful, biased, or factually incorrect outputs -- and treats the reasoning chain as an opaque intermediate artifact. We identify reasoning safety as an orthogonal and equally critical security dimension: the requirement that a model's reasoning trajectory be logically consistent, computationally efficient, and resistant to adversarial manipulation. We make three contributions. First, we formally define reasoning safety and introduce a nine-category taxonomy of unsafe reasoning behaviors, covering input parsing errors, reasoning execution errors, and process management errors. Second, we conduct a large-scale prevalence study annotating 4111 reasoning chains from both natural reasoning benchmarks and four adversarial attack methods (reasoning hijacking and denial-of-service), confirming that all nine error types occur in practice and that each attack induces a mechanistically interpretable signature. Third, we propose a Reasoning Safety Monitor: an external LLM-based component that runs in parallel with the target model, inspects each reasoning step in real time via a taxonomy-embedded prompt, and dispatches an interrupt signal upon detecting unsafe behavior. Evaluation on a 450-chain static benchmark shows that our monitor achieves up to 84.88\% step-level localization accuracy and 85.37\% error-type classification accuracy, outperforming hallucination detectors and process reward model baselines by substantial margins. These results demonstrate that reasoning-level monitoring is both necessary and practically achievable, and establish reasoning safety as a foundational concern for the secure deployment of large reasoning models.

Long-horizon tabletop games pose a distinct systems challenge for robotics: small perceptual or execution errors can invalidate accumulated task state, propagate across decision-making modules, and ultimately derail interaction. This paper studies how to maintain internal state consistency in turn-based, multi-human robotic tabletop games through deliberate system design rather than isolated component improvement. Using Mahjong as a representative long-horizon setting, we present an integrated architecture that explicitly maintains perceptual, execution, and interaction state, partitions high-level semantic reasoning from time-critical perception and control, and incorporates verified action primitives with tactile-triggered recovery to prevent premature state corruption. We further introduce interaction-level monitoring mechanisms to detect turn violations and hidden-information breaches that threaten execution assumptions. Beyond demonstrating complete-game operation, we provide an empirical characterization of failure modes, recovery effectiveness, cross-module error propagation, and hardware-algorithm trade-offs observed during deployment. Our results show that explicit partitioning, monitored state transitions, and recovery mechanisms are critical for sustaining executable consistency over extended play, whereas monolithic or unverified pipelines lead to measurable degradation in end-to-end reliability. The proposed system serves as an empirical platform for studying system-level design principles in long-horizon, turn-based interaction.

On-device Vision-Language Models (VLMs) promise data privacy via local execution. However, we show that the architectural shift toward Dynamic High-Resolution preprocessing (e.g., AnyRes) introduces an inherent algorithmic side-channel. Unlike static models, dynamic preprocessing decomposes images into a variable number of patches based on their aspect ratio, creating workload-dependent inputs. We demonstrate a dual-layer attack framework against local VLMs. In Tier 1, an unprivileged attacker can exploit significant execution-time variations using standard unprivileged OS metrics to reliably fingerprint the input's geometry. In Tier 2, by profiling Last-Level Cache (LLC) contention, the attacker can resolve semantic ambiguity within identical geometries, distinguishing between visually dense (e.g., medical X-rays) and sparse (e.g., text documents) content. By evaluating state-of-the-art models such as LLaVA-NeXT and Qwen2-VL, we show that combining these signals enables reliable inference of privacy-sensitive contexts. Finally, we analyze the security engineering trade-offs of mitigating this vulnerability, reveal substantial performance overhead with constant-work padding, and propose practical design recommendations for secure Edge AI deployments.

In this applied paper, we address the difficult open problem of when to discharge patients from the Intensive Care Unit. This can be conceived as an optimal stopping scenario with three added challenges: 1) the evaluation of a stopping strategy from observational data is itself a complex causal inference problem, 2) the composite objective is to minimize the length of intervention and maximize the outcome, but the two cannot be collapsed to a single dimension, and 3) the recording of variables stops when the intervention is discontinued. Our contributions are two-fold. First, we generalize the implementation of the g-formula Python package, providing a framework to evaluate stopping strategies for problems with the aforementioned structure, including positivity and coverage checks. Second, with a fully open-source pipeline, we apply this approach to MIMIC-IV, a public ICU dataset, demonstrating the potential for strategies that improve upon current care.

Multi-robot systems can be extremely efficient for accomplishing team-wise tasks by acting concurrently and collaboratively. However, most existing methods either assume static task features or simply replan when environmental changes occur. This paper addresses the challenging problem of coordinating multi-robot systems for collaborative tasks involving dynamic and moving targets. We explicitly model the uncertainty in target motion prediction via Conformal Prediction(CP), while respecting the spatial-temporal constraints specified by Linear Temporal Logic (LTL). The proposed framework (UMBRELLA) combines the Monte Carlo Tree Search (MCTS) over partial plans with uncertainty-aware rollouts, and introduces a CP-based metric to guide and accelerate the search. The objective is to minimize the Conditional Value at Risk (CVaR) of the average makespan. For tasks released online, a receding-horizon planning scheme dynamically adjusts the assignments based on updated task specifications and motion predictions. Spatial and temporal constraints among the tasks are always ensured, and only partial synchronization is required for the collaborative tasks during online execution. Extensive large-scale simulations and hardware experiments demonstrate substantial reductions in both the average makespan and its variance by 23% and 71%, compared with static baselines.

Every software organization faces the onboarding challenge: helping newcomers navigate complex codebases, compensate for insufficient documentation, and comprehend code they did not author. Expert walkthroughs are among the most effective forms of support, yet they are expensive, repetitive, and do not scale. We present Lacy, a hybrid human-AI onboarding system that captures expert mentoring in reusable code tours-to our knowledge, the first hybrid approach combining AI-generated content with expert curation in code tours. Our design is grounded in requirements derived from 20+ meetings, surveys, and interviews across a year-long industry partnership with Beko. Supporting features include Voice-to-Tour capture, comprehension quizzes, podcasts, and a dashboard. We deployed Lacy on Beko's production environment and conducted a controlled study on a legacy finance system (30K+ LOC). Learners using expert-guided tours achieved 83% quiz scores versus 57% for AI-only tours, preferred tours over traditional self-study, and reported they would need fewer expert consultations. Experts found tour creation less burdensome than live walkthroughs. Beko has since adopted Lacy for organizational onboarding, and we release our code and study instruments as a replication package.

Quantization techniques such as BitsAndBytes, AWQ, and GPTQ are widely used as a standard method in deploying large language models but often degrades accuracy when using low-bit representations, e.g., 4 bits. Low-rank correction methods (e.g., LQER, QERA, ASER) has been proposed to mitigate this issue, however, they restore all layers and insert error-correction modules into every decoder block, which increases latency and memory overhead. To address this limitation, we propose GlowQ, a group-shared low-rank approximation for quantized LLMs that caches a single shared right factor per input-sharing group and restores only the groups or layers that yield the highest accuracy benefit. GlowQ computes the high-precision projection once per input-sharing group and reuses it across its modules, reducing parameter and memory overhead, and retaining the expressivity of layer-specific corrections. We also propose a selective variant, GlowQ-S, that applies the cached shared module only where it provides the largest benefit. Compared with strong baselines, our approach reduces TTFB by (5.6%) and increases throughput by (9.6%) on average, while reducing perplexity on WikiText-2 by (0.17%) and increasing downstream accuracy by 0.42 percentage points. The selective model GlowQ-S further reduces latency, cutting TTFB by (23.4%) and increasing throughput by (37.4%), while maintaining accuracy within 0.2 percentage points on average.

A faithful description of chemical processes requires exploring extended regions of the molecular potential energy surface (PES), which remains challenging for strongly correlated systems. Transferable deep-learning variational Monte Carlo (VMC) offers a promising route by efficiently solving the electronic Schrödinger equation jointly across molecular geometries at consistently high accuracy, yet its stochastic nature renders direct exploration of molecular configuration space nontrivial. Here, we present a framework for highly accurate ab initio exploration of PESs that combines transferable deep-learning VMC with a cost-effective estimation of energies, forces, and Hessians. By continuously sampling nuclear configurations during VMC optimization of electronic wave functions, we obtain transferable descriptions that achieve zero-shot chemical accuracy within chemically relevant distributions of molecular geometries. Throughout the subsequent characterization of molecular configuration space, the PES is evaluated only sparsely, with local approximations constructed by estimating VMC energies and forces at sampled geometries and aggregating the resulting noisy data using Gaussian process regression. Our method enables accurate and efficient exploration of complex PES landscapes, including structure relaxation, transition-state searches, and minimum-energy pathways, for both ground and excited states. This opens the door to studying bond breaking, formation, and large structural rearrangements in systems with pronounced multi-reference character.

Does structured intent representation generalize across languages and models? We study PPS (Prompt Protocol Specification), a 5W3H-based framework for structured intent representation in human-AI interaction, and extend prior Chinese-only evidence along three dimensions: two additional languages (English and Japanese), a fourth condition in which a user's simple prompt is automatically expanded into a full 5W3H specification by an AI-assisted authoring interface, and a new research question on cross-model output consistency. Across 2,160 model outputs (3 languages x 4 conditions x 3 LLMs x 60 tasks), we find that AI-expanded 5W3H prompts (Condition D) show no statistically significant difference in goal alignment from manually crafted 5W3H prompts (Condition C) across all three languages, while requiring only a single-sentence input from the user. Structured PPS conditions often reduce or reshape cross-model output variance, though this effect is not uniform across languages and metrics; the strongest evidence comes from identifying spurious low variance in unconstrained baselines. We also show that unstructured prompts exhibit a systematic dual-inflation bias: artificially high composite scores and artificially low apparent cross-model variance. These findings suggest that structured 5W3H representations can improve intent alignment and accessibility across languages and models, especially when AI-assisted authoring lowers the barrier for non-expert users.

Accurately forecasting GPU workloads is essential for AI infrastructure, enabling efficient scheduling, resource allocation, and power management. Modern workloads are highly volatile, multiple periodicity, and heterogeneous, making them challenging for traditional predictors. We propose PRISM, a primitive-based compositional forecasting framework combining dictionary-driven temporal decomposition with adaptive spectral refinement. This dual representation extracts stable, interpretable workload signatures across diverse GPU jobs. Evaluated on large-scale production traces, PRISM achieves state-of-the-art results. It significantly reduces burst-phase errors, providing a robust, architecture-aware foundation for dynamic resource management in GPU-powered AI platforms.

RAG typically assumes centralized access to documents, which breaks down when knowledge is distributed across private data silos. We propose a secure Federated RAG system built using Flower that performs local silo retrieval, while server-side aggregation and text generation run inside an attested, confidential compute environment, enabling confidential remote LLM inference even in the presence of honest-but-curious or compromised servers. We also propose a cascading inference approach that incorporates a non-confidential third-party model (e.g., Amazon Nova) as auxiliary context without weakening confidentiality.

Local curvature of potential energy surfaces is critical for predicting certain experimental observables of molecules and materials from first principles, yet it remains far beyond reach for complex systems. In this work, we introduce a Hessian-informed Machine Learning Interatomic Potential (Hi-MLIP) that captures such curvature reliably, thereby enabling accurate analysis of associated thermodynamic and kinetic phenomena. To make Hessian supervision practically viable, we develop a highly efficient training protocol, termed Hessian INformed Training (HINT), achieving two to four orders of magnitude reduction for the requirement of expensive Hessian labels. HINT integrates critical techniques, including Hessian pre-training, configuration sampling, curriculum learning and stochastic projection Hessian loss. Enabled by HINT, Hi-MLIP significantly improves transition-state search and brings Gibbs free-energy predictions close to chemical accuracy especially in data-scarce regimes. Our framework also enables accurate treatment of strongly anharmonic hydrides, reproducing phonon renormalization and superconducting critical temperatures in close agreement with experiment while bypassing the computational bottleneck of anharmonic calculations. These results establish a practical route to enhancing curvature awareness of machine learning interatomic potentials, bridging simulation and experimental observables across a wide range of systems.

Probabilistic forecasting provides a principled framework for uncertainty quantification in dynamical systems by representing predictions as probability distributions rather than deterministic trajectories. However, existing forecasting approaches, whether physics-based or neural-network-based, remain fundamentally trajectory-oriented: predictive distributions are usually accessed through ensembles or sampling, rather than evolved directly as dynamical objects. A distribution-to-distribution (D2D) neural probabilistic forecasting framework is developed to operate directly on predictive distributions. The framework introduces a distributional encoding and decoding structure around a replaceable neural forecasting module, using kernel mean embeddings to represent input distributions and mixture density networks to parameterise output predictive distributions. This design enables recursive propagation of predictive uncertainty within a unified end-to-end neural architecture, with model training and evaluation carried out directly in terms of probabilistic forecast skill. The framework is demonstrated on the Lorenz63 chaotic dynamical system. Results show that the D2D model captures nontrivial distributional evolution under nonlinear dynamics, produces skillful probabilistic forecasts without explicit ensemble simulation, and remains competitive with, and in some cases outperforms, a simplified perfect model benchmark. These findings point to a new paradigm for probabilistic forecasting, in which predictive distributions are learned and evolved directly rather than reconstructed indirectly through ensemble-based uncertainty propagation.

We investigate the \emph{minimum weight cycle (MWC)} problem in the $\mathsf{CONGEST}$ model of distributed computing. For undirected weighted graphs, we design a randomized algorithm that achieves a $(k+1)$-approximation, for any \emph{real} number $k \ge 1$. The round complexity of algorithm is \[ \tilde{O}\!\Big( n^{\frac{k+1}{2k+1}} + n^{\frac{1}{k}} + D\, n^{\frac{1}{2(2k+1)}} + D^{\frac{2}{5}} n^{\frac{2}{5}+\frac{1}{2(2k+1)}} \Big). \] where $n$ denotes the number of nodes and $D$ is the unweighted diameter of the graph. This result yields a smooth trade-off between approximation ratio and round complexity. In particular, when $k \geq 2$ and $D = \tilde{O}(n^{1/4})$, the bound simplifies to \[ \tilde{O}\!\left( n^{\frac{k+1}{2k+1}} \right) \] On the lower bound side, assuming the Erdős girth conjecture, we prove that for every \emph{integer} $k \ge 1$, any randomized $(k+1-ε)$-approximation algorithm for MWC requires \[ \tildeΩ\!\left( n^{\frac{k+1}{2k+1}} \right) \] rounds. This lower bound holds for both directed unweighted and undirected weighted graphs, and applies even to graphs with small diameter $D = Θ(\log n)$. Taken together, our upper and lower bounds \emph{match up to polylogarithmic factors} for graphs of sufficiently small diameter $D = \tilde{O}(n^{1/4})$ (when $k \geq 2$), yielding a nearly tight bound on the distributed complexity of the problem. Our results improve upon the previous state of the art: Manoharan and Ramachandran (PODC~2024) demonstrated a $(2+ε)$-approximation algorithm for undirected weighted graphs with round complexity $\tilde{O}(n^{2/3}+D)$, and proved that for any arbitrarily large number $α$, any $α$-approximation algorithm for directed unweighted or undirected weighted graphs requires $Ω(\sqrt{n}/\log n)$ rounds.

Autonomous object search is challenging for mobile robots operating in indoor environments due to partial observability, perceptual uncertainty, and the need to trade off exploration and navigation efficiency. Classical probabilistic approaches explicitly represent uncertainty but typically rely on handcrafted action-selection heuristics, while deep reinforcement learning enables adaptive policies but often suffers from slow convergence and limited interpretability. This paper proposes a hybrid object-search framework that integrates Bayesian inference with deep reinforcement learning. The method maintains a spatial belief map over target locations, updated online through Bayesian inference from calibrated object detections, and trains a reinforcement learning policy to select navigation actions directly from this probabilistic representation. The approach is evaluated in realistic indoor simulation using Habitat 3.0 and compared against developed baseline strategies. Across two indoor environments, the proposed method improves success rate while reducing search effort. Overall, the results support the value of combining Bayesian belief estimation with learned action selection to achieve more efficient and reliable objectsearch behavior under partial observability.

Arithmetic puzzle games provide a controlled setting for studying difficulty in mathematical reasoning tasks, a core challenge in adaptive learning systems. We investigate the structural determinants of difficulty in a class of integer arithmetic puzzles inspired by number games. We formalize the problem and develop an exact dynamic-programming solver that enumerates reachable targets, extracts minimal-operation witnesses, and enables large-scale labeling. Using this solver, we construct a dataset of over 3.4 million instances and define difficulty via the minimum number of operations required to reach a target. We analyze the relationship between difficulty and solver-derived features. While baseline machine learning models based on bag- and target-level statistics can partially predict solvability, they fail to reliably distinguish easy instances. In contrast, we show that difficulty is fully determined by a small set of interpretable structural attributes derived from exact witnesses. In particular, the number of input values used in a minimal construction serves as a minimal sufficient statistic for difficulty under this labeling. These results provide a transparent, computationally grounded account of puzzle difficulty that bridges symbolic reasoning and data-driven modeling. The framework supports explainable difficulty estimation and principled task sequencing, with direct implications for adaptive arithmetic learning and intelligent practice systems.

Automatic image rotation estimation is a key preprocessing step in many vision pipelines. This task is challenging because angles have circular topology, creating boundary discontinuities that hinder standard regression methods. We present a comprehensive study of five circular-aware methods for global orientation estimation: direct angle regression with circular loss, classification via angular binning, unit-vector regression, phase-shifting coder, and circular Gaussian distribution. Using transfer learning from ImageNet-pretrained models, we systematically evaluate these methods across sixteen modern architectures by adapting their output heads for rotation-specific predictions. Our results show that probabilistic methods, particularly the circular Gaussian distribution, are the most robust across architectures, while classification achieves the best accuracy on well-matched backbones but suffers training instabilities on others. The best configuration (classification with EfficientViT-B3) achieves a mean absolute error (MAE) of 1.23° (mean across five independent runs) on the DRC-D dataset, while the circular Gaussian distribution with MambaOut Base achieves a virtually identical 1.24° with greater robustness across backbones. Training and evaluating our top-performing method-architecture combinations on COCO 2014, the best configuration reaches 3.71° MAE, improving substantially over prior work, with further improvement to 2.84° on the larger COCO 2017 dataset.

Although deep research agents (DRAs) have emerged as a promising paradigm for complex information synthesis, their evaluation remains constrained by ad hoc empirical benchmarks. These heuristic approaches do not rigorously model agent behavior or adequately stress-test long-horizon synthesis and ambiguity resolution. To bridge this gap, we formalize DRA behavior through the lens of category theory, modeling deep research workflow as a composition of structure-preserving maps (functors). Grounded in this theoretical framework, we introduce a novel mechanism-aware benchmark with 296 questions designed to stress-test agents along four interpretable axes: traversing sequential connectivity chains, verifying intersections within V-structure pullbacks, imposing topological ordering on retrieved substructures, and performing ontological falsification via the Yoneda Probe. Our rigorous evaluation of 11 leading models establishes a persistently low baseline, with the state-of-the-art achieving only a 19.9\% average accuracy, exposing the difficulty of formal structural stress-testing. Furthermore, our findings reveal a stark dichotomy in the current AI capabilities. While advanced deep research pipelines successfully redefine dynamic topological re-ordering and exhibit robust ontological verification -- matching pure reasoning models in falsifying hallucinated premises -- they almost universally collapse on multi-hop structural synthesis. Crucially, massive performance variance across tasks exposes a lingering reliance on brittle heuristics rather than a systemic understanding. Ultimately, this work demonstrates that while top-tier autonomous agents can now organically unify search and reasoning, achieving a generalized mastery over complex structural information remains a formidable open challenge.\footnote{Our implementation will be available at https://github.com/tzq1999/CDR.

In this paper, we establish the novel insight that an off-the-shelf LLM can function as an excellent token compressor and decompressor. To demonstrate, we design a self-expressive autoencoding learning framework fine-tunes a pretrained LLM to translate long texts into a compact internal language of discrete, variable-length latent codes, termed Z-tokens, and to reconstruct the original text exactly from them. The resulting representation is content-adaptive: semantically dense segments receive more Z-tokens, while redundant or predictable regions are aggressively compressed, via lightweight LoRA-based adapter heads. Empirically, our method achieves up to 18 times token reduction on Wikipedia, CNN/DailyMail, HotpotQA, and Qulac-style long-query datasets, while preserving reconstruction fidelity and downstream performance. This simple yet effective design supports applications including prompt compression and autoregressive generation directly in the Z-token space, offering a potential pathway toward token-efficient long-context reasoning.

Distributed intelligence in industrial networks increasingly integrates sensing, communication, and computation across heterogeneous and resource constrained devices. Federated learning (FL) enables collaborative model training in such environments, but its reliability is affected by inconsistent client behaviour, noisy sensing conditions, and the presence of faulty or adversarial updates. Trust based mechanisms are commonly used to mitigate these effects, yet most remain statistical and heuristic, relying on fixed parameters or simple adaptive rules that struggle to accommodate changing operating conditions. This paper presents a lightweight agentic trust coordination approach for FL in sustainable and resilient industrial networks. The proposed Agentic Trust Control Layer operates as a server side control loop that observes trust related and system level signals, interprets their evolution over time, and applies targeted trust adjustments when instability is detected. The approach extends prior adaptive trust mechanisms by enabling context aware intervention decisions, rather than relying on fixed or purely reactive parameter updates. By explicitly separating observation, reasoning, and action, the proposed framework supports stable FL operation without modifying client side training or increasing communication overhead.

The effectiveness of Retrieval-Augmented Generation (RAG) is highly dependent on how documents are chunked, that is, segmented into smaller units for indexing and retrieval. Yet, commonly used "one-size-fits-all" approaches often fail to capture the nuanced structure and semantics of diverse texts. Despite its central role, chunking lacks a dedicated evaluation framework, making it difficult to assess and compare strategies independently of downstream performance. We challenge this paradigm by introducing Adaptive Chunking, a framework that selects the most suitable chunking strategy for each document based on a set of five novel intrinsic, document-based metrics: References Completeness (RC), Intrachunk Cohesion (ICC), Document Contextual Coherence (DCC), Block Integrity (BI), and Size Compliance (SC), which directly assess chunking quality across key dimensions. To support this framework, we also introduce two new chunkers, an LLM-regex splitter and a split-then-merge recursive splitter, alongside targeted post-processing techniques. On a diverse corpus spanning legal, technical, and social science domains, our metric-guided adaptive method significantly improves downstream RAG performance. Without changing models or prompts, our framework increases RAG outcomes, raising answers correctness to 72% (from 62-64%) and increasing the number of successfully answered questions by over 30% (65 vs. 49). These results demonstrate that adaptive, document-aware chunking, guided by a complementary suite of intrinsic metrics, offers a practical and effective path to more robust RAG systems. Code available at https://github.com/ekimetrics/adaptive-chunking.

As generative AI tools like ChatGPT enter classrooms, workplaces and everyday thinking, writing is at risk of becoming a formality -- outsourced, automated and stripped of its cognitive value. But writing is not just output; it is how we learn to think. This paper explores what is lost when we let machines write for us, drawing on cognitive psychology, educational theory and real classroom practices. We argue that the process of writing -- messy, slow, often frustrating -- is where a human deep learning happens. The paper also explores the current possibilities of AI-text detection, how educators can adapt through smarter pedagogy rather than bans, and why the ability to recognize machine-generated language may become a critical literacy of the 21st century. In a world where writing can be faked, learning can not.

Automated Vehicle (AV) control in mixed traffic, where AVs coexist with human-driven vehicles, poses significant challenges in balancing safety, efficiency, comfort, fuel efficiency, and compliance with traffic rules while capturing heterogeneous driver behavior. Traditional car-following models, such as the Intelligent Driver Model (IDM), often struggle to generalize across diverse traffic scenarios and typically do not account for fuel efficiency, motivating the use of learning-based approaches. Although Deep Reinforcement Learning (DRL) has shown strong microscopic performance in car-following conditions, its macroscopic traffic flow characteristics remain underexplored. This study focuses on analyzing the macroscopic traffic flow characteristics and fuel efficiency of DRL-based models in mixed traffic. A Twin Delayed Deep Deterministic Policy Gradient (TD3) algorithm is implemented for AVs' control and trained using the NGSIM highway dataset, enabling realistic interaction with human-driven vehicles. Traffic performance is evaluated using the Fundamental Diagram (FD) under varying driver heterogeneity, heterogeneous time-gap penetration levels, and different shares of RL-controlled vehicles. A macroscopic level comparison of fuel efficiency between the RL-based AV model and the IDM is also conducted. Results show that traffic performance is sensitive to the distribution of safe time gaps and the proportion of RL vehicles. Transitioning from fully human-driven to fully RL-controlled traffic can increase road capacity by approximately 7.52%. Further, RL-based AVs also improve average fuel efficiency by about 28.98% at higher speeds (above 50 km/h), and by 1.86% at lower speeds (below 50 km/h) compared to the IDM. Overall, the DRL framework enhances traffic capacity and fuel efficiency without compromising safety.

Interest in the concept of AI-driven harmful manipulation is growing, yet current approaches to evaluating it are limited. This paper introduces a framework for evaluating harmful AI manipulation via context-specific human-AI interaction studies. We illustrate the utility of this framework by assessing an AI model with 10,101 participants spanning interactions in three AI use domains (public policy, finance, and health) and three locales (US, UK, and India). Overall, we find that that the tested model can produce manipulative behaviours when prompted to do so and, in experimental settings, is able to induce belief and behaviour changes in study participants. We further find that context matters: AI manipulation differs between domains, suggesting that it needs to be evaluated in the high-stakes context(s) in which an AI system is likely to be used. We also identify significant differences across our tested geographies, suggesting that AI manipulation results from one geographic region may not generalise to others. Finally, we find that the frequency of manipulative behaviours (propensity) of an AI model is not consistently predictive of the likelihood of manipulative success (efficacy), underscoring the importance of studying these dimensions separately. To facilitate adoption of our evaluation framework, we detail our testing protocols and make relevant materials publicly available. We conclude by discussing open challenges in evaluating harmful manipulation by AI models.

Weight pruning is a standard technique for compressing large language models, yet its effect on learned internal representations remains poorly understood. We present the first systematic study of how unstructured pruning reshapes the feature geometry of language models, using Sparse Autoencoders (SAEs) as interpretability probes. Across three model families (Gemma 3 1B, Gemma 2 2B, Llama 3.2 1B), two pruning methods (magnitude and Wanda), and six sparsity levels (0--60%), we investigate five research questions spanning seed stability, feature survival, SAE transferability, feature fragility, and causal relevance. Our most striking finding is that rare SAE features--those with low firing rates--survive pruning far better than frequent ones, with within-condition Spearman correlations of rho = -1.0 in 11 of 17 experimental conditions. This counter-intuitive result suggests that pruning acts as implicit feature selection, preferentially destroying high-frequency generic features while preserving specialized rare ones. We further show that Wanda pruning preserves feature structure up to 3.7x better than magnitude pruning, that pre-trained SAEs remain viable on Wanda-pruned models up to 50% sparsity, and that geometric feature survival does not predict causal importance--a dissociation with implications for interpretability under compression.

Alzheimer's disease (AD) is a growing global health challenge as populations age, and timely, accurate diagnosis is essential to reduce individual and societal burden. However, real-world AD assessment is hampered by incomplete, heterogeneous multimodal data and variability across sites and patient demographics. Although large language models (LLMs) have shown promise in biomedicine, their use in AD has largely been confined to answering narrow, disease-specific questions rather than generating comprehensive diagnostic reports that support clinical decision-making. Here we expand LLM capabilities for clinical decision support by introducing AD-CARE, a modality-agnostic agent that performs guideline-grounded diagnostic assessment from incomplete, heterogeneous inputs without imputing missing modalities. By dynamically orchestrating specialized diagnostic tools and embedding clinical guidelines into LLM-driven reasoning, AD-CARE generates transparent, report-style outputs aligned with real-world clinical workflows. Across six cohorts comprising 10,303 cases, AD-CARE achieved 84.9% diagnostic accuracy, delivering 4.2%-13.7% relative improvements over baseline methods. Despite cohort-level differences, dataset-specific accuracies remain robust (80.4%-98.8%), and the agent consistently outperforms all baselines. AD-CARE reduced performance disparities across racial and age subgroups, decreasing the average dispersion of four metrics by 21%-68% and 28%-51%, respectively. In a controlled reader study, the agent improved neurologist and radiologist accuracy by 6%-11% and more than halved decision time. The framework yielded 2.29%-10.66% absolute gains over eight backbone LLMs and converges their performance. These results show that AD-CARE is a scalable, practically deployable framework that can be integrated into routine clinical workflows for multimodal decision support in AD.

Efficient global optimization (EGO) is one of the most widely used noise-free Bayesian optimization algorithms.It comprises the Gaussian process (GP) surrogate model and expected improvement (EI) acquisition function. In practice, when EGO is applied, a scalar matrix of a small positive value (also called a nugget or jitter) is usually added to the covariance matrix of the deterministic GP to improve numerical stability. We refer to this EGO with a positive nugget as the practical EGO. Despite its wide adoption and empirical success, to date, cumulative regret bounds for practical EGO have yet to be established. In this paper, we present for the first time the cumulative regret upper bound of practical EGO. In particular, we show that practical EGO has sublinear cumulative regret bounds and thus is a no-regret algorithm for commonly used kernels including the squared exponential (SE) and Matérn kernels ($ν>\frac{1}{2}$). Moreover, we analyze the effect of the nugget on the regret bound and discuss the theoretical implication on its choice. Numerical experiments are conducted to support and validate our findings.

Current Large Language Models (LLMs) mostly use BPE (Byte Pair Encoding) based tokenizers, which are very effective for simple structured Latin scripts such as English. However, standard BPE tokenizers struggle to process complex Abugida scripts due to their structural complexity. The problem is that these tokenizers break complex conjuncts, which are multi-codepoint grapheme clusters, into meaningless sub-character units. This degrades the LLM's reasoning efficiency by forcing it to learn basic orthographic structures at inference time and raises inference costs, resulting in a significant "Token Tax" for the Global South. We propose a new three-layer architecture, the WWHO (Where-What-How Often), and an algorithm named SGPE (Syllable-aware Grapheme Pair Encoding) that separates the linguistic rules of the script from the statistical compression process while enabling seamless multilingual tokenization. Using Sinhala and Devanagari (Hindi/Sanskrit) as highly complex Abugida scripts, we trained WWHO on a cleaned 30-million-sentence dataset and evaluated on a 1,499,950-sentence test set. For Sinhala, SGPE achieves a Token to Word Ratio (TWR) of 1.274 with 4.83 characters per token, representing a 61.7 percent reduction in tokens compared to OpenAI's o200k base. For Hindi, it achieves a TWR of 1.181 (27.0 percent reduction vs o200k). On the mixed-script (Sinhala, Devanagari, and English) dataset, SGPE achieves an overall TWR of 1.240, representing token reductions of 36.7 percent, 39.6 percent, and 60.2 percent relative to o200k base, Llama 4 Scout, and DeepSeek V3, respectively. This effectively extends the usable context window by up to 4.38 times for these Abugida languages while ensuring a Linguistic Zero-Breakage Guarantee, which ensures that no valid syllable is ever split across multiple tokens.

The Kelly or proportional allocation mechanism is a simple and efficient auction-based scheme that distributes an infinitely divisible resource proportionally to the agents bids. When agents are aware of the allocation rule, their interactions form a game extensively studied in the literature. This paper examines the less explored repeated Kelly game, focusing mainly on utilities that are logarithmic in the allocated resource fraction. We first derive this logarithmic form from fairness-throughput trade-offs in wireless network slicing, and then prove that the induced stage game admits a unique Nash equilibrium NE. For the repeated play, we prove convergence to this NE under three behavioral models: (i) all agents use Online Gradient Descent (OGD), (ii) all agents use Dual Averaging with a quadratic regularizer (DAQ) (a variant of the Follow-the-Regularized leader algorithm), and (iii) all agents play myopic best responses (BR). Our convergence results hold even when agents use personalized learning rates in OGD and DAQ (e.g., tuned to optimize individual regret bounds), and they extend to a broader class of utilities that meet a certain sufficient condition. Finally, we complement our theoretical results with extensive simulations of the repeated Kelly game under several behavioral models, comparing them in terms of convergence speed to the NE, and per-agent time-average utility. The results suggest that BR achieves the fastest convergence and the highest time-average utility, and that convergence to the stage-game NE may fail under heterogeneous update rules.

Directed Acyclic Graphs (DAGs) are widely used to represent structured knowledge in scientific and technical domains. However, datasets for real-world DAGs remain scarce because constructing them typically requires expert interpretation of domain documents. We study Doc2SemDAG construction: recovering a preferred semantic DAG from a document together with the cited evidence and context that explain it. This problem is challenging because a document may admit multiple plausible abstractions, the intended structure is often implicit, and the supporting evidence is scattered across prose, equations, captions, and figures. To address these challenges, we leverage scientific papers containing explicit DAG figures as a natural source of supervision. In this setting, the DAG figure provides the DAG structure, while the accompanying text provides context and explanation. We introduce DAGverse, a framework for constructing document-grounded semantic DAGs from online scientific papers. Its core component, DAGverse-Pipeline, is a semi-automatic system designed to produce high-precision semantic DAG examples through figure classification, graph reconstruction, semantic grounding, and validation. As a case study, we test the framework for causal DAGs and release DAGverse-1, a dataset of 108 expert-validated semantic DAGs with graph-level, node-level, and edge-level evidence. Experiments show that DAGverse-Pipeline outperforms existing Vision-Language Models on DAG classification and annotation. DAGverse provides a foundation for document-grounded DAG benchmarks and opens new directions for studying structured reasoning grounded in real-world evidence.

Federated Learning (FL) is a paradigm for training machine learning (ML) models in collaborative settings while preserving participants' privacy by keeping raw data local. A key requirement for the use of FL in production is reliability, as insufficient reliability can compromise the validity, stability, and reproducibility of learning outcomes. FL inherently operates as a distributed system and is therefore susceptible to crash failures, network partitioning, and other fault scenarios. Despite this, the impact of such failures on FL outcomes has not yet been studied systematically. In this paper, we address this gap by investigating the impact of missing participants in FL. To this end, we conduct extensive experiments on image, tabular, and time-series data and analyze how the absence of participants affects model performance, taking into account influencing factors such as data skewness, different availability patterns, and model architectures. Furthermore, we examine scenario-specific aspects, including the utility of the global model for missing participants. Our experiments provide detailed insights into the effects of various influencing factors. In particular, we show that data skewness has a strong impact, often leading to overly optimistic model evaluations and, in some cases, even altering the effects of other influencing factors.

Low-altitude communications can promote the integration of aerial and terrestrial wireless resources, expand network coverage, and enhance transmission quality, thereby empowering the development of sixth-generation (6G) mobile communications. As an enabler for low-altitude transmission, 3D channel fingerprints (3D-CF), also referred to as the 3D radio map or 3D channel knowledge map, are expected to enhance the understanding of communication environments and assist in the acquisition of channel state information (CSI), thereby avoiding repeated estimations and reducing computational complexity. In this paper, we propose a modularized multimodal framework to construct 3D-CF. Specifically, we first establish the 3D-CF model as a collection of CSI-tuples based on Rician fading channels, with each tuple comprising the low-altitude vehicle's (LAV) positions and its corresponding statistical CSI. In consideration of the heterogeneous structures of different prior data, we formulate the 3D-CF construction problem as a multimodal regression task, where the target channel information in the CSI-tuple can be estimated directly by its corresponding LAV positions, together with communication measurements and geographic environment maps. Then, a high-efficiency multimodal framework is proposed accordingly, which includes a correlation-based multimodal fusion (Corr-MMF) module, a multimodal representation (MMR) module, and a CSI regression (CSI-R) module. Numerical results show that our proposed framework can efficiently construct 3D-CF and achieve at least 27.5% higher accuracy than the state-of-the-art algorithms under different communication scenarios, demonstrating its competitive performance and excellent generalization ability. We also analyze the computational complexity and illustrate its superiority in terms of the inference time.

In this paper, we address post-training quantization (PTQ) for large language models (LLMs) from an overlooked perspective: given a pre-trained high-precision LLM, the predominant sequential quantization framework treats different layers equally, but this may be not optimal in challenging bit-width settings. We empirically study the quantization impact of different layers on model accuracy, and observe that: (1) shallow/deep layers are usually more sensitive to quantization than intermediate layers; (2) among shallow/deep layers, the most sensitive one is the first/last layer, which exhibits significantly larger quantization error than others. These empirical observations imply that the quantization design for different layers of LLMs is required on multiple levels instead of a single level shared to all layers. Motivated by this, we propose a new PTQ framework termed Sliding-layer Quantization (SliderQuant) that relies on a simple adaptive sliding quantization concept facilitated by few learnable parameters. The base component of SliderQuant is called inter-layer sliding quantization, which incorporates three types of novel sliding window designs tailored for addressing the varying quantization sensitivity of shallow, intermediate and deep layers. The other component is called intra-layer sliding quantization that leverages an incremental strategy to quantize each window. As a result, SliderQuant has a strong ability to reduce quantization errors across layers. Extensive experiments on basic language generation, zero-shot commonsense reasoning and challenging math and code tasks with various LLMs, including Llama/Llama2/Llama3/Qwen2.5 model families, DeepSeek-R1 distilled models and large MoE models, show that our method outperforms existing PTQ methods (including the latest PTQ methods using rotation transformations) for both weight-only quantization and weight-activation quantization.

Gait is increasingly recognized as a vital sign, yet current approaches treat it as a symptom of specific pathologies rather than a systemic biomarker. We developed a gait foundation model for 3D skeletal motion from 3,414 deeply phenotyped adults, recorded via a depth camera during five motor tasks. Learned embeddings outperformed engineered features, predicting age (Pearson r = 0.69), BMI (r = 0.90), and visceral adipose tissue area (r = 0.82). Embeddings significantly predicted 1,980 of 3,210 phenotypic targets; after adjustment for age, BMI, VAT, and height, gait provided independent gains in all 18 body systems in males and 17 of 18 in females, and improved prediction of clinical diagnoses and medication use. Anatomical ablation revealed that legs dominated metabolic and frailty predictions while torso encoded sleep and lifestyle phenotypes. These findings establish gait as an independent multi-system biosignal, motivating translation to consumer-grade video and its integration as a scalable, passive vital sign.

The probabilistic abstract interpretation framework of neural network analysis analyzes a neural network by analyzing its density distribution flow of all possible inputs. The grids approximation is one of abstract domains the framework uses which abstracts concrete space into grids. In this paper, we introduce two novel approximation methods: distribution approximation and clusters approximation. We show how these two methods work in theory with corresponding abstract transformers with help of illustrations of some simple examples.

Hateful content online is often expressed using fact-like, not necessarily correct information, especially in coordinated online harassment campaigns and extremist propaganda. Failing to jointly address hate speech (HS) and misinformation can deepen prejudice, reinforce harmful stereotypes, and expose bystanders to psychological distress, while polluting public debate. Moreover, these messages require more effort from content moderators because they must assess both harmfulness and veracity, i.e., fact-check them. To address this challenge, we release WSF-ARG+, the first dataset which combines hate speech with check-worthiness information. We also introduce a novel LLM-in-the-loop framework to facilitate the annotation of check-worthy claims. We run our framework, testing it with 12 open-weight LLMs of different sizes and architectures. We validate it through extensive human evaluation, and show that our LLM-in-the-loop framework reduces human effort without compromising the annotation quality of the data. Finally, we show that HS messages with check-worthy claims show significantly higher harassment and hate, and that incorporating check-worthiness labels improves LLM-based HS detection up to 0.213 macro-F1 and to 0.154 macro-F1 on average for large models.

We introduce CRAFT, a multi-agent benchmark for evaluating pragmatic communication in large language models under strict partial information. In this setting, multiple agents with complementary but incomplete views must coordinate through natural language to construct a shared 3D structure that no single agent can fully observe. We formalize this problem as a multi-sender pragmatic reasoning task and provide a diagnostic framework that decomposes failures into spatial grounding, belief modeling and pragmatic communication errors, including a taxonomy of behavioral failure profiles in both frontier and open-weight models. Across a diverse set of models, including 8 open-weight and 7 frontier including reasoning models, we find that stronger reasoning ability does not reliably translate to better coordination: smaller open-weight models often match or outperform frontier systems, and improved individual communication does not guarantee successful collaboration. These results suggest that multi-agent coordination remains a fundamentally unsolved challenge for current language models. Our code can be found at https://github.com/csu-signal/CRAFT

Probabilistic abstract interpretation is a theory used to extract particular properties of a computer program when it is infeasible to test every single inputs. In this paper we apply the theory on neural networks for the same purpose: to analyse density distribution flow of all possible inputs of a neural network when a network has uncountably many or countable but infinitely many inputs. We show how this theoretical framework works in neural networks and then discuss different abstract domains and corresponding Moore-Penrose pseudo-inverses together with abstract transformers used in the framework. We also present experimental examples to show how this framework helps to analyse real world problems.

This paper investigates the susceptibility to model integrity attacks that overload virtual machines assigned by the k-means algorithm used for resource provisioning in fog networks. The considered k-means algorithm runs two phases iteratively: offline clustering to form clusters of requested workload and online classification of new incoming requests into offline-created clusters. First, we consider an evasion attack against the classifier in the online phase. A threat actor launches an exploratory attack using query-based reverse engineering to discover the Machine Learning (ML) model (the clustering scheme). Then, a passive causative (evasion) attack is triggered in the offline phase. To defend the model, we suggest a proactive method using adversarial training to introduce attack robustness into the classifier. Our results show that our mitigation technique effectively maintains the stability of the resource provisioning system against attacks.

Trajectory anomaly detection underpins applications from fraud detection to urban mobility analysis. Dense GPS methods preserve fine-grained evidence such as abnormal speeds and short-duration events, but their quadratic cost makes multi-month analysis intractable; consequently, no existing approach detects anomalies over multi-month dense GPS trajectories. The field instead relies on scalable sparse stay-point methods that discard this evidence, forcing separate architectures for each regime and preventing knowledge transfer. We argue this bottleneck is unnecessary: human trajectories, dense or sparse, share a natural two-dimensional cyclic structure along within-day and across-day axes. We therefore propose TITAnD (Trajectory Image Transformer for Anomaly Detection), which reformulates trajectory anomaly detection as a vision problem by representing trajectories as a Hyperspectral Trajectory Image (HTI): a day x time-of-day grid whose channels encode spatial, semantic, temporal, and kinematic information from either modality, unifying both under a single representation. Under this formulation, agent-level detection reduces to image classification and temporal localization to semantic segmentation. To model this representation, we introduce the Cyclic Factorized Transformer (CFT), which factorizes attention along the two temporal axes, encoding the cyclic inductive bias of human routines, while reducing attention cost by orders of magnitude and enabling dense multi-month anomaly detection for the first time. Empirically, TITAnD achieves the best AUC-PR across sparse and dense benchmarks, surpassing vision models like UNet while being 11-75x faster than the Transformer with comparable memory, demonstrating that vision reformulation and structure-aware modeling are jointly essential. Code will be made public soon.

Large language models (LLMs) hold considerable potential for advancing scientific discovery, yet systematic assessment of their dynamic reasoning in real-world research remains limited. Current scientific evaluation benchmarks predominantly rely on static, single-turn Question Answering (QA) formats, which are inadequate for measuring model performance in complex scientific tasks that require multi-step iteration and experimental interaction. To address this gap, we introduce MolQuest, a novel agent-based evaluation framework for molecular structure elucidation built upon authentic chemical experimental data. Unlike existing datasets, MolQuest formalizes molecular structure elucidation as a multi-turn interactive task, requiring models to proactively plan experimental steps, integrate heterogeneous spectral sources (e.g., NMR, MS), and iteratively refine structural hypotheses. This framework systematically evaluates LLMs' abductive reasoning and strategic decision-making abilities within a vast and complex chemical space. Empirical results reveal that contemporary frontier models exhibit significant limitations in authentic scientific scenarios: notably, even state-of-the-art (SOTA) models achieve an accuracy of only approximately 50%, while the performance of most other models remains below the 30% threshold. This work provides a reproducible and extensible framework for science-oriented LLM evaluation, our findings highlight the critical gap in current LLMs' strategic scientific reasoning, setting a clear direction for future research toward AI that can actively participate in the scientific process.

Explainable AI (XAI) methods are commonly evaluated with functional metrics such as correctness, which computationally estimate how accurately an explanation reflects the model's reasoning. Higher correctness is assumed to produce better human understanding, but this link has not been tested experimentally with controlled levels. We conducted a user study (N=200) that manipulated explanation correctness at four levels (100%, 85%, 70%, 55%) in a time series classification task where participants could not rely on domain knowledge or visual intuition and instead predicted the AI's decisions based on explanations (forward simulation). Correctness affected understanding, but not at every level: performance dropped at 70% and 55% correctness relative to fully correct explanations, while further degradation below 70% produced no additional loss. Rather than shifting performance uniformly, lower correctness decreased the proportion of participants who learned the decision pattern. At the same time, even fully correct explanations did not guarantee understanding, as only a subset of participants achieved high accuracy. Exploratory analyses showed that self-reported ratings correlated with demonstrated performance only when explanations were fully correct and participants had learned the pattern. These findings show that not all differences in functional correctness translate to differences in human understanding, underscoring the need to validate functional metrics against human outcomes.

Out-of-distribution (OOD) detection aims to identify samples that deviate from in-distribution (ID). One popular pipeline addresses this by introducing negative labels distant from ID classes and detecting OOD based on their distance to these labels. However, such labels may present poor activation on OOD samples, failing to capture the OOD characteristics. To address this, we propose \underline{T}est-time \underline{A}ctivated \underline{N}egative \underline{L}abels (TANL) by dynamically evaluating activation levels across the corpus dataset and mining candidate labels with high activation responses during the testing process. Specifically, TANL identifies high-confidence test images online and accumulates their assignment probabilities over the corpus to construct a label activation metric. Such a metric leverages historical test samples to adaptively align with the test distribution, enabling the selection of distribution-adaptive activated negative labels. By further exploring the activation information within the current testing batch, we introduce a more fine-grained, batch-adaptive variant. To fully utilize label activation knowledge, we propose an activation-aware score function that emphasizes negative labels with stronger activations, boosting performance and enhancing its robustness to the label number. Our TANL is training-free, test-efficient, and grounded in theoretical justification. Experiments on diverse backbones and wide task settings validate its effectiveness. Notably, on the large-scale ImageNet benchmark, TANL significantly reduces the FPR95 from 17.5\% to 9.8\%. Codes are available at \href{https://github.com/YBZh/OpenOOD-VLM}{YBZh/OpenOOD-VLM}.

Spatial Transcriptomics (ST) provides spatially-resolved gene expression, offering crucial insights into tissue architecture and complex diseases. However, its prohibitive cost limits widespread adoption, leading to significant attention on inferring spatial gene expression from readily available whole slide images. While graph neural networks have been proposed to model interactions between tissue regions, their reliance on pre-defined sparse graphs prevents them from considering potentially interacting spot pairs, resulting in a structural limitation in capturing complex biological relationships. To address this, we propose FEAST (Fully connected Expressive Attention for Spatial Transcriptomics), an attention-based framework that models the tissue as a fully connected graph, enabling the consideration of all pairwise interactions. To better reflect biological interactions, we introduce negative-aware attention, which models both excitatory and inhibitory interactions, capturing essential negative relationships that standard attention often overlooks. Furthermore, to mitigate the information loss from truncated or ignored context in standard spot image extraction, we introduce an off-grid sampling strategy that gathers additional images from intermediate regions, allowing the model to capture a richer morphological context. Experiments on public ST datasets show that FEAST surpasses state-of-the-art methods in gene expression prediction while providing biologically plausible attention maps that clarify positive and negative interactions. Our code is available at https://github.com/starforTJ/ FEAST.

Large language models and autonomous agents are increasingly explored for EDA automation, but many existing integrations still rely on script-level or request-level interactions, which makes it difficult to preserve tool state and support iterative optimization in real production-oriented environments. In this work, we present FluxEDA, a unified and stateful infrastructure substrate for agentic EDA. FluxEDA introduces a managed gateway-based execution interface with structured request and response handling. It also maintains persistent backend instances. Together, these features allow upper-layer agents and programmable clients to interact with heterogeneous EDA tools through preserved runtime state, rather than through isolated shell invocations. We evaluate the framework using two representative commercial backend case studies: automated post-route timing ECO and standard-cell sub-library optimization. The results show that FluxEDA can support multi-step analysis and optimization over real tool contexts, including state reuse, rollback, and coordinated iterative execution. These findings suggest that a stateful and governed infrastructure layer is a practical foundation for agent-assisted EDA automation.

Combinatorial optimization problems like the Traveling Salesman Problem are critical in industry yet NP-hard. Neural Combinatorial Optimization has shown promise, but its reliance on online reinforcement learning (RL) hampers deployment and underutilizes decades of algorithmic knowledge. We address these limitations by applying the offline RL framework, Decision Transformer, to learn superior strategies directly from datasets of heuristic solutions; it aims to not only to imitate but to synthesize and outperform them. Concretely, we (i) integrate a Pointer Network to handle the instance-dependent, variable action space of node selection, and (ii) employ expectile regression for optimistic conditioning of Return-to-Go, which is crucial for instances with widely varying optimal values. Experiments show that our method consistently produces higher-quality tours than the four classical heuristics it is trained on, demonstrating the potential of offline RL to unlock and exceed the performance embedded in existing domain knowledge.

Skin diseases are a major public health concern worldwide, and their detection is often challenging without access to dermatological expertise. In countries like Bangladesh, which is highly populated, the number of qualified skin specialists and diagnostic instruments is insufficient to meet the demand. Due to the lack of proper detection and treatment of skin diseases, that may lead to severe health consequences including death. Common properties of skin diseases are, changing the color, texture, and pattern of skin and in this era of artificial intelligence and machine learning, we are able to detect skin diseases by using image processing and computer vision techniques. In response to this challenge, we develop a publicly available dataset focused on common skin disease detection using machine learning techniques. We focus on five prevalent skin diseases in Bangladesh: Contact Dermatitis, Vitiligo, Eczema, Scabies, and Tinea Ringworm. The dataset consists of 1612 images (of which, 250 are distinct while others are augmented), collected directly from patients at the outpatient department of Faridpur Medical College, Faridpur, Bangladesh. The data comprises of 302, 381, 301, 316, and 312 images of Dermatitis, Eczema, Scabies, Tinea Ringworm, and Vitiligo, respectively. Although the data are collected regionally, the selected diseases are common across many countries especially in South Asia, making the dataset potentially valuable for global applications in machine learning-based dermatology. We also apply several machine learning and deep learning models on the dataset and report classification performance. We expect that this research would garner attention from machine learning and deep learning researchers and practitioners working in the field of automated disease diagnosis.

This study compares the impact of natural and synthetic data on training and evaluating large language models (LLMs), using the case of passive verb alternation in French and Italian. We use Blackbird Language Matrices (BLMs), structured datasets designed to probe linguistic knowledge of underlying patterns across sentence sets. We compare structured templates instantiated with natural sentences extracted from Universal Dependencies to structured templates of synthetic sentences. Experiments show that while models achieve ceiling performance when trained and tested on synthetic datasets, they do not reliably generalize to natural sentences. In contrast, models trained on natural data exhibit robust performance across both natural and synthetic test suites, demonstrating their superior ability to capture abstract linguistic patterns. These results corroborate the value of natural data and of structured set ups in linguistic evaluation for probing LLMs' syntactic and semantic knowledge.

The emergence of Large Language Models (LLMs) has catalyzed a paradigm shift in programming, giving rise to "vibe coding", where users can build complete projects and even control computers using natural language instructions. This paradigm has driven automated webpage development, but it introduces a new requirement about how to automatically verify whether the web functionalities are reliably implemented. Existing works struggle to adapt, relying on static visual similarity or predefined checklists that constrain their utility in open-ended environments. Furthermore, they overlook a vital aspect of software quality, namely latent logical constraints. To address these gaps, we introduce WebTestBench, a benchmark for evaluating end-to-end automated web testing. WebTestBench encompasses comprehensive dimensions across diverse web application categories. We decompose the testing process into two cascaded sub-tasks, checklist generation and defect detection, and propose WebTester, a baseline framework for this task. Evaluating popular LLMs with WebTester reveals severe challenges, including insufficient test completeness, detection bottlenecks, and long-horizon interaction unreliability. These findings expose a substantial gap between current computer-use agent capabilities and industrial-grade deployment demands. We hope that WebTestBench provides valuable insights and guidance for advancing end-to-end automated web testing. Our dataset and code are available at https://github.com/friedrichor/WebTestBench.

Demographic parity (DP) is a widely used group fairness criterion requiring predictive distributions to be invariant across sensitive groups. While natural in classification, full distributional DP is often overly restrictive in regression and can lead to substantial accuracy loss. We propose a relaxation of DP tailored to regression, enforcing parity only at a finite set of quantile levels and/or score thresholds. Concretely, we introduce a novel (${\ell}$, Z)-fair predictor, which imposes groupwise CDF constraints of the form F f |S=s (z m ) = ${\ell}$ m for prescribed pairs (${\ell}$ m , z m ). For this setting, we derive closed-form characterizations of the optimal fair discretized predictor via a Lagrangian dual formulation and quantify the discretization cost, showing that the risk gap to the continuous optimum vanishes as the grid is refined. We further develop a model-agnostic post-processing algorithm based on two samples (labeled for learning a base regressor and unlabeled for calibration), and establish finite-sample guarantees on constraint violation and excess penalized risk. In addition, we introduce two alternative frameworks where we match group and marginal CDF values at selected score thresholds. In both settings, we provide closed-form solutions for the optimal fair discretized predictor. Experiments on synthetic and real datasets illustrate an interpretable fairness-accuracy trade-off, enabling targeted corrections at decision-relevant quantiles or thresholds while preserving predictive performance.

The landscape of extremely low-resource machine translation (MT) is characterized by perplexing variability in reported performance, often making results across different language pairs difficult to contextualize. For researchers focused on specific language groups -- such as ancient languages -- it is nearly impossible to determine if breakthroughs reported in other contexts (e.g., native African or American languages) result from superior methodologies or are merely artifacts of benchmark collection. To address this problem, we introduce the FRED Difficulty Metrics, which include the Fertility Ratio (F), Retrieval Proxy (R), Pre-training Exposure (E), and Corpus Diversity (D) and serve as dataset-intrinsic metrics to contextualize reported scores. These metrics reveal that a significant portion of result variability is explained by train-test overlap and pre-training exposure rather than model capability. Additionally, we identify that some languages -- particularly extinct and non-Latin indigenous languages -- suffer from poor tokenization coverage (high token fertility), highlighting a fundamental limitation of transferring models from high-resource languages that lack a shared vocabulary. By providing these indices alongside performance scores, we enable more transparent evaluation of cross-lingual transfer and provide a more reliable foundation for the XLR MT community.

Robust Support Vector Machines (R-SVMs) address feature noise by adopting a worst-case robust formulation that explicitly incorporates uncertainty sets into training. While this robustness improves reliability, it also leads to increased computational cost. In this work, we develop safe sample screening rules for R-SVMs that reduce the training complexity without affecting the optimal solution. To the best of our knowledge, this is the first study to apply safe screening techniques to worst-case robust models in supervised machine learning. Our approach safely identifies training samples whose uncertainty sets are guaranteed to lie entirely on either side of the margin hyperplane, thereby reducing the problem size and accelerating optimization. Owing to the nonstandard structure of R-SVMs, the proposed screening rules are derived from the Lagrangian duality rather than the Fenchel-Rockafellar duality commonly used in recent methods. Based on this analysis, we first establish an ideal screening rule, and then derive a practical rule by adapting GAP-based safe regions to the robust setting. Experiments demonstrate that the proposed method significantly reduces training time while preserving classification accuracy.

Integrating AI into the physical layer is a cornerstone of 6G networks. However, current data-driven approaches struggle to generalize across dynamic environments because they lack an intrinsic understanding of electromagnetic wave propagation. We introduce the Wireless World Model (WWM), a multi-modal foundation framework predicting the spatiotemporal evolution of wireless channels by internalizing the causal relationship between 3D geometry and signal dynamics. Pre-trained on a massive ray-traced multi-modal dataset, WWM overcomes the data authenticity gap, further validated under real-world measurement data. Using a joint-embedding predictive architecture with a multi-modal mixture-of-experts Transformer, WWM fuses channel state information, 3D point clouds, and user trajectories into a unified representation. Across the five key downstream tasks supported by WWM, it achieves remarkable performance in seen environments, unseen generalization scenarios, and real-world measurements, consistently outperforming SOTA uni-modal foundation models and task-specific models. This paves the way for physics-aware 6G intelligence that adapts to the physical world.

Transmitter localization in vessel-like molecular communication channels is a fundamental problem with potential applications in healthcare. Existing analytical solutions either assume knowledge of emission time or require multiple closely spaced receivers, which limits their applicability in realistic scenarios. In this letter, we propose a simple closed-form approximation that exploits the temporal variance of the received molecular signal to estimate the distance between the transmitter and the receiver without emission time information. The method is derived from a Gaussian approximation of the received signal near its peak and gives an explicit variance-distance relation. Simulation results in physically relevant capillary vessel scale show that the proposed method achieves distance prediction with error on the order of 1%.

Generating long videos using pre-trained video diffusion models, which are typically trained on short clips, presents a significant challenge. Directly applying these models for long-video inference often leads to a notable degradation in visual quality. This paper identifies that this issue primarily stems from two out-of-distribution (O.O.D) problems: frame-level relative position O.O.D and context-length O.O.D. To address these challenges, we propose FreeLOC, a novel training-free, layer-adaptive framework that introduces two core techniques: Video-based Relative Position Re-encoding (VRPR) for frame-level relative position O.O.D, a multi-granularity strategy that hierarchically re-encodes temporal relative positions to align with the model's pre-trained distribution, and Tiered Sparse Attention (TSA) for context-length O.O.D, which preserves both local detail and long-range dependencies by structuring attention density across different temporal scales. Crucially, we introduce a layer-adaptive probing mechanism that identifies the sensitivity of each transformer layer to these O.O.D issues, allowing for the selective and efficient application of our methods. Extensive experiments demonstrate that our approach significantly outperforms existing training-free methods, achieving state-of-the-art results in both temporal consistency and visual quality. Code is available at https://github.com/Westlake-AGI-Lab/FreeLOC.

Conditional density estimation (CDE) is a fundamental task in machine learning that aims to model the full conditional law $\mathbb{P}(\mathbf{y} \mid \mathbf{x})$, beyond mere point prediction (e.g., mean, mode). A core challenge is free-form density estimation, capturing distributions that exhibit multimodality, asymmetry, or topological complexity without restrictive assumptions. However, prevailing methods typically estimate the probability density function (PDF) directly, which is mathematically ill-posed: differentiating the empirical distribution amplifies random fluctuations inherent in finite datasets, necessitating strong inductive biases that limit expressivity and fail when violated. We propose a CDF-first framework that circumvents this issue by estimating the cumulative distribution function (CDF), a stable and well-posed target, and then recovering the PDF via differentiation of the learned smooth CDF. Parameterizing the CDF with a Smooth Min-Max (SMM) network, our framework guarantees valid PDFs by construction, enables tractable approximate likelihood training, and preserves complex distributional shapes. For multivariate outputs, we use an autoregressive decomposition with SMM factors. Experiments demonstrate our approach outperforms state-of-the-art density estimators on a range of univariate and multivariate tasks.

Multimodal misinformation poses an escalating challenge that often evades traditional detectors, which are opaque black boxes and fragile against new manipulation tactics. We present Probabilistic Concept Graph Reasoning (PCGR), an interpretable and evolvable framework that reframes multimodal misinformation detection (MMD) as structured and concept-based reasoning. PCGR follows a build-then-infer paradigm, which first constructs a graph of human-understandable concept nodes, including novel high-level concepts automatically discovered and validated by multimodal large language models (MLLMs), and then applies hierarchical attention over this concept graph to infer claim veracity. This design produces interpretable reasoning chains linking evidence to conclusions. Experiments demonstrate that PCGR achieves state-of-the-art MMD accuracy and robustness to emerging manipulation types, outperforming prior methods in both coarse detection and fine-grained manipulation recognition.

Recent research points toward LLMs being manipulated through adversarial and seemingly benign inputs, resulting in harmful, biased, or policy-violating outputs. In this paper, we study an underexplored issue concerning harmful and toxic mathematical word problems. We show that math questions, particularly those framed as natural language narratives, can serve as a subtle medium for propagating biased, unethical, or psychologically harmful content, with heightened risks in educational settings involving children. To support a systematic study of this phenomenon, we introduce ToxicGSM, a dataset of 1.9k arithmetic problems in which harmful or sensitive context is embedded while preserving mathematically well-defined reasoning tasks. Using this dataset, we audit the behaviour of existing LLMs and analyse the trade-offs between safety enforcement and mathematical correctness. We further propose SafeMath -- a safety alignment technique that reduces harmful outputs while maintaining, and in some cases improving, mathematical reasoning performance. Our results highlight the importance of disentangling linguistic harm from math reasoning and demonstrate that effective safety alignment need not come at the cost of accuracy. We release the source code and dataset at https://github.com/Swagnick99/SafeMath/tree/main.

As AI assistants become integrated into safety engineering workflows for Physical AI systems, a critical question emerges: does AI assistance improve safety analysis quality, or introduce systematic blind spots that surface only through post-deployment incidents? This paper develops a formal framework for AI assistance in safety analysis. We first establish why safety engineering resists benchmark-driven evaluation: safety competence is irreducibly multidimensional, constrained by context-dependent correctness, inherent incompleteness, and legitimate expert disagreement. We formalize this through a five-dimensional competence framework capturing domain knowledge, standards expertise, operational experience, contextual understanding, and judgment. We introduce the competence shadow: the systematic narrowing of human reasoning induced by AI-generated safety analysis. The shadow is not what the AI presents, but what it prevents from being considered. We formalize four canonical human-AI collaboration structures and derive closed-form performance bounds, demonstrating that the competence shadow compounds multiplicatively to produce degradation far exceeding naive additive estimates. The central finding is that AI assistance in safety engineering is a collaboration design problem, not a software procurement decision. The same tool degrades or improves analysis quality depending entirely on how it is used. We derive non-degradation conditions for shadow-resistant workflows and call for a shift from tool qualification toward workflow qualification for trustworthy Physical AI.

Clinical practice guidelines (CPGs) play a pivotal role in ensuring evidence-based decision-making and improving patient outcomes. While Large Language Models (LLMs) are increasingly deployed in healthcare scenarios, it is unclear to which extend LLMs could identify and adhere to CPGs during conversations. To address this gap, we introduce CPGBench, an automated framework benchmarking the clinical guideline detection and adherence capabilities of LLMs in multi-turn conversations. We collect 3,418 CPG documents from 9 countries/regions and 2 international organizations published in the last decade spanning across 24 specialties. From these documents, we extract 32,155 clinical recommendations with corresponding publication institute, date, country, specialty, recommendation strength, evidence level, etc. One multi-turn conversation is generated for each recommendation accordingly to evaluate the detection and adherence capabilities of 8 leading LLMs. We find that the 71.1%-89.6% recommendations can be correctly detected, while only 3.6%-29.7% corresponding titles can be correctly referenced, revealing the gap between knowing the guideline contents and where they come from. The adherence rates range from 21.8% to 63.2% in different models, indicating a large gap between knowing the guidelines and being able to apply them. To confirm the validity of our automatic analysis, we further conduct a comprehensive human evaluation involving 56 clinicians from different specialties. To our knowledge, CPGBench is the first benchmark systematically revealing which clinical recommendations LLMs fail to detect or adhere to during conversations. Given that each clinical recommendation may affect a large population and that clinical applications are inherently safety critical, addressing these gaps is crucial for the safe and responsible deployment of LLMs in real world clinical practice.

Public blockchains impose an inherent tension between regulatory compliance and user privacy. Existing on-chain identity solutions require centralized KYC attestors, specialized hardware, or Decentralized Identifier (DID) frameworks needing entirely new credential infrastructure. Meanwhile, over four billion active X.509 certificates constitute a globally deployed, government-grade trust infrastructure largely unexploited for decentralized identity. This paper presents zk-X509, a privacy-preserving identity system bridging legacy Public Key Infrastructure (PKI) with public ledgers via a RISC-V zero-knowledge virtual machine (zkVM). Users prove ownership of standard X.509 certificates without revealing private keys or personal identifiers. Crucially, the private key never enters the ZK circuit; ownership is proven via OS keychain signature delegation (e.g., macOS Secure Enclave, Windows TPM). The circuit verifies certificate chain validity, temporal validity, key ownership, trustless CRL revocation, blockchain address binding, and Sybil-resistant nullifier generation. It commits 13 public values, including a Certificate Authority (CA) Merkle root hiding the issuing CA, and four selective disclosure hashes. We formalize eight security properties under a Dolev-Yao adversary with game-based definitions and reductions to sEUF-CMA, SHA-256 collision resistance, and ZK soundness. Evaluated on the SP1 zkVM, the system achieves 11.8M cycles for ECDSA P-256 (17.4M for RSA-2048), with on-chain Groth16 verification costing ~300K gas. By leveraging certificates deployed at scale across jurisdictions, zk-X509 enables adoption without new trust establishment, complementing emerging DID-based systems.

Recent research on dialectal NLP has identified data scarcity as a primary limitation. To address this limitation, this paper presents a catalog of contemporary Basque dialectal data and resources, offering a systematic and comprehensive compilation of the dialectal data currently available in Basque. Two types of data sources have been distinguished: online data originally written in some dialect, and standard-to-dialect adapted data. The former includes all dialectal data that can be found online, such as news and radio sites, informal tweets, as well as online resources such as dictionaries, atlases, grammar rules, or videos. The latter consists of data that has been adapted from the standard variety to dialectal varieties, either manually or automatically. Regarding the manual adaptation, the test split of the XNLI Natural Language Inference dataset was manually adapted into three Basque dialects: Western, Central, and Navarrese-Lapurdian, yielding a high-quality parallel gold standard evaluation dataset. With respect to the automatic dialectal adaptation, the automatically adapted physical commonsense dataset (BasPhyCowest) underwent additional manual evaluation by native speakers to assess its quality and determine whether it could serve as a viable substitute for full manual adaptation (i.e., silver data creation).

Persona-driven large language models (LLMs) require consistent behavioral tendencies across interactions to simulate human-like personality traits, such as persistence or reliability. However, current LLMs often lack stable internal representations that anchor their responses over extended dialogues. This work explores whether LLMs can maintain "implicit consistency", defined as persistent adherence to an unstated goal in multi-turn interactions. We designed a 20-question-style riddle game paradigm where an LLM is tasked with secretly selecting a target and responding to users' guesses with "yes/no" answers. Through evaluations, we find that LLMs struggle to preserve latent consistency: their implicit "goals" shift across turns unless explicitly provided their selected target in context. These findings highlight critical limitations in the building of persona-driven LLMs and underscore the need for mechanisms that anchor implicit goals over time, which is a key to realistic personality modeling in interactive applications such as dialogue systems.

AI systems in healthcare research have shown potential to increase patient throughput and assist clinicians, yet progress is constrained by limited access to real patient data. To address this issue, we present a zero-shot, knowledge-guided framework for psychiatric tabular data in which large language models (LLMs) are steered via Retrieval-Augmented Generation using the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) and the International Classification of Diseases (ICD-10). We conducted experiments using different combinations of knowledge bases to generate privacy-preserving synthetic data. The resulting models were benchmarked against two state-of-the-art deep learning models for synthetic tabular data generation, namely CTGAN and TVAE, both of which rely on real data and therefore entail potential privacy risks. Evaluation was performed on six anxiety-related disorders: specific phobia, social anxiety disorder, agoraphobia, generalized anxiety disorder, separation anxiety disorder, and panic disorder. CTGAN typically achieves the best marginals and multivariate structure, while the knowledge-augmented LLM is competitive on pairwise structure and attains the lowest pairwise error in separation anxiety and social anxiety. An ablation study shows that clinical retrieval reliably improves univariate and pairwise fidelity over a no-retrieval LLM. Privacy analyses indicate that the real data-free LLM yields modest overlaps and a low average linkage risk comparable to CTGAN, whereas TVAE exhibits extensive duplication despite a low k-map score. Overall, grounding an LLM in clinical knowledge enables high-quality, privacy-preserving synthetic psychiatric data when real datasets are unavailable or cannot be shared.

Reinforcement learning (RL) has become essential for post-training large language models (LLMs) in reasoning tasks. While scaling rollouts can stabilize training and enhance performance, the computational overhead is a critical issue. In algorithms like GRPO, multiple rollouts per prompt incur prohibitive costs, as a large portion of prompts provide negligible gradients and are thus of low utility. To address this problem, we investigate how to select high-utility prompts before the rollout phase. Our experimental analysis reveals that sample utility is non-uniform and evolving: the strongest learning signals concentrate at the ``learning edge", the intersection of intermediate difficulty and high uncertainty, which shifts as training proceeds. Motivated by this, we propose HIVE (History-Informed and online-VErified prompt selection), a dual-stage framework for data-efficient RL. HIVE utilizes historical reward trajectories for coarse selection and employs prompt entropy as a real-time proxy to prune instances with stale utility. By evaluating HIVE across multiple math reasoning benchmarks and models, we show that HIVE yields significant rollout efficiency without compromising performance.

Context-aware machine translation (MT) leverages document-level information, yet it does not consistently outperform sentence-level MT, as contextual signals are unevenly beneficial across sentences. Existing training objectives do not explicitly model this variability, limiting a model's ability to adaptively exploit context. In this paper, we propose Cross-Preference Learning (CPL), a preference-based training framework that explicitly captures the complementary benefits of sentence-level and context-aware MT. CPL achieves this by integrating both intra- and cross-condition preferences into the preference optimization objective. The introduction of intra- and cross-condition preferences provides explicit supervision on when and how contextual information improves translation quality. We validate the proposed approach on several public context-aware MT tasks using multiple models, including Qwen3-4B, Qwen3-8B, and Llama-3-8B. Experimental results demonstrate consistent improvements in translation quality and robustness across both input conditions, achieved without any architectural modifications.

Text-to-motion generation holds significant potential for cross-linguistic applications, yet it is hindered by the lack of bilingual datasets and the poor cross-lingual semantic understanding of existing language models. To address these gaps, we introduce BiHumanML3D, the first bilingual text-to-motion benchmark, constructed via LLM-assisted annotation and rigorous manual correction. Furthermore, we propose a simple yet effective baseline, Bilingual Motion Diffusion (BiMD), featuring Cross-Lingual Alignment (CLA). CLA explicitly aligns semantic representations across languages, creating a robust conditional space that enables high-quality motion generation from bilingual inputs, including zero-shot code-switching scenarios. Extensive experiments demonstrate that BiMD with CLA achieves an FID of 0.045 vs. 0.169 and R@3 of 82.8\% vs. 80.8\%, significantly outperforms monolingual diffusion models and translation baselines on BiHumanML3D, underscoring the critical necessity and reliability of our dataset and the effectiveness of our alignment strategy for cross-lingual motion synthesis. The dataset and code are released at \href{https://wengwanjiang.github.io/BilingualT2M-page}{https://wengwanjiang.github.io/BilingualT2M-page}

Prompt attacks, including jailbreaks and prompt injections, pose a critical security risk to Large Language Model (LLM) systems. In production, guardrails must mitigate these attacks under strict low-latency constraints, resulting in a deployment gap in which lightweight classifiers and rule-based systems struggle to generalize under distribution shift, while high-capacity LLM-based judges remain too slow or costly for live enforcement. In this work, we examine whether lightweight, general-purpose LLMs can reliably serve as security judges under real-world production constraints. Through careful prompt and output design, lightweight LLMs are guided through a structured reasoning process involving explicit intent decomposition, safety-signal verification, harm assessment, and self-reflection. We evaluate our method on a curated dataset combining benign queries from real-world chatbots with adversarial prompts generated via automated red teaming (ART), covering diverse and evolving patterns. Our results show that general-purpose LLMs, such as gemini-2.0-flash-lite-001, can serve as effective low-latency judges for live guardrails. This configuration is currently deployed in production as a centralized guardrail service for public service chatbots in Singapore. We additionally evaluate a Mixture-of-Models (MoM) setting to assess whether aggregating multiple LLM judges improves prompt-attack detection performance relative to single-model judges, with only modest gains observed.

In complex environments, infrared object detection exhibits broad applicability and stability across diverse scenarios. However, infrared object detection is vulnerable to both common corruptions and adversarial examples, leading to potential security risks. To improve the robustness of infrared object detection, current methods mostly adopt a data-driven ideology, which only superficially drives the network to fit the training data without specifically considering the unique characteristics of infrared images, resulting in limited robustness. In this paper, we revisit infrared physical knowledge and find that relative thermal radiation relations between different classes can be regarded as a reliable knowledge source under the complex scenarios of adversarial examples and common corruptions. Thus, we theoretically model thermal radiation relations based on the rank order of gray values for different classes, and further quantify the stability of various inter-class thermal radiation relations. Based on the above theoretical framework, we propose Knowledge-Guided Adversarial Training (KGAT) for infrared object detection, in which infrared physical knowledge is embedded into the adversarial training process, and the predicted results are optimized to be consistent with the actual physical laws. Extensive experiments on three infrared datasets and six mainstream infrared object detection models demonstrate that KGAT effectively enhances both clean accuracy and robustness against adversarial attacks and common corruptions.

LLM performance is highly sensitive to prompt design, yet whether automatic prompt optimization can replace expert prompt engineering in linguistic tasks remains unexplored. We present the first systematic comparison of hand-crafted zero-shot expert prompts, base DSPy signatures, and GEPA-optimized DSPy signatures across translation, terminology insertion, and language quality assessment, evaluating five model configurations. Results are task-dependent. In terminology insertion, optimized and manual prompts produce mostly statistically indistinguishable quality. In translation, each approach wins on different models. In LQA, expert prompts achieve stronger error detection while optimization improves characterization. Across all tasks, GEPA elevates minimal DSPy signatures, and the majority of expert-optimized comparisons show no statistically significant difference. We note that the comparison is asymmetric: GEPA optimization searches programmatically over gold-standard splits, whereas expert prompts require in principle no labeled data, relying instead on domain expertise and iterative refinement.

Large Language Models (LLMs) have demonstrated remarkable performance across a wide range of applications. However, their practical deployment is often hindered by issues such as outdated knowledge and the tendency to generate hallucinations. To address these limitations, Retrieval-Augmented Generation (RAG) systems have been introduced, enhancing LLMs with external, up-to-date knowledge sources. Despite their advantages, RAG systems remain vulnerable to adversarial attacks, with data poisoning emerging as a prominent threat. Existing poisoning-based attacks typically require prior knowledge of the user's specific queries, limiting their flexibility and real-world applicability. In this work, we propose PIDP-Attack, a novel compound attack that integrates prompt injection with database poisoning in RAG. By appending malicious characters to queries at inference time and injecting a limited number of poisoned passages into the retrieval database, our method can effectively manipulate LLM response to arbitrary query without prior knowledge of the user's actual query. Experimental evaluations across three benchmark datasets (Natural Questions, HotpotQA, MS-MARCO) and eight LLMs demonstrate that PIDP-Attack consistently outperforms the original PoisonedRAG. Specifically, our method improves attack success rates by 4% to 16% on open-domain QA tasks while maintaining high retrieval precision, proving that the compound attack strategy is both necessary and highly effective.

Equipping Large Language Model (LLM) agents with domain-specific skills is critical for tackling complex tasks. Yet, manual authoring creates a severe scalability bottleneck. Conversely, automated skill generation often yields fragile or fragmented results because it either relies on shallow parametric knowledge or sequentially overfits to non-generalizable trajectory-local lessons. To overcome this, we introduce Trace2Skill, a framework that mirrors how human experts author skills: by holistically analyzing broad execution experience before distilling it into a single, comprehensive guide. Instead of reacting sequentially to individual trajectories, Trace2Skill dispatches a parallel fleet of sub-agents to analyze a diverse pool of executions. It extracts trajectory-specific lessons and hierarchically consolidates them into a unified, conflict-free skill directory via inductive reasoning. Trace2Skill supports both deepening existing human-written skills and creating new ones from scratch. Experiments in challenging domains, such as spreadsheet, VisionQA and math reasoning, show that Trace2Skill significantly improves upon strong baselines, including Anthropic's official xlsx skills. Crucially, this trajectory-grounded evolution does not merely memorize task instances or model-specific quirks: evolved skills transfer across LLM scales and generalize to OOD settings. For example, skills evolved by Qwen3.5-35B on its own trajectories improved a Qwen3.5-122B agent by up to 57.65 absolute percentage points on WikiTableQuestions. Ultimately, our results demonstrate that complex agent experience can be packaged into highly transferable, declarative skills -- requiring no parameter updates, no external retrieval modules, and utilizing open-source models as small as 35B parameters.

Recent vision and multimodal foundation backbones, such as Transformer families and state-space models like Mamba, have achieved remarkable progress, enabling unified modeling across images, text, and beyond. Despite their empirical success, these architectures remain far from the computational principles of the human brain, often demanding enormous amounts of training data while offering limited interpretability. In this work, we propose the Vision Hopfield Memory Network (V-HMN), a brain-inspired foundation backbone that integrates hierarchical memory mechanisms with iterative refinement updates. Specifically, V-HMN incorporates local Hopfield modules that provide associative memory dynamics at the image patch level, global Hopfield modules that function as episodic memory for contextual modulation, and a predictive-coding-inspired refinement rule for iterative error correction. By organizing these memory-based modules hierarchically, V-HMN captures both local and global dynamics in a unified framework. Memory retrieval exposes the relationship between inputs and stored patterns, making decisions more interpretable, while the reuse of stored patterns improves data efficiency. This brain-inspired design therefore enhances interpretability and data efficiency beyond existing self-attention- or state-space-based approaches. We conducted extensive experiments on public computer vision benchmarks, and V-HMN achieved competitive results against widely adopted backbone architectures, while offering better interpretability, higher data efficiency, and stronger biological plausibility. These findings highlight the potential of V-HMN to serve as a next-generation vision foundation model, while also providing a generalizable blueprint for multimodal backbones in domains such as text and audio, thereby bridging brain-inspired computation with large-scale machine learning.

Multimodal large language models are promising for clinical visual question answering tasks, but scaling to 3D imaging is hindered by high computational costs. Prior methods often rely on 2D slices or fixed-length token compression, disrupting volumetric continuity and obscuring subtle findings. We present Photon, a framework that represents 3D medical volumes with token sequences of variable length. Photon introduces instruction-conditioned token scheduling and surrogate gradient propagation to adaptively reduce tokens during both training and inference, which lowers computational cost while mitigating the attention dilution caused by redundant tokens. It incorporates a custom backpropagation rule with gradient restoration to enable differentiable optimization despite discrete token drop. To stabilize token compression and ensure reliable use of visual evidence, Photon further applies regularization objectives that mitigate language-only bias and improve reliability. Experiments on diverse medical visual question answering tasks show that Photon achieves state-of-the-art accuracy while reducing resource usage and accelerating both training and inference.

Retrieval-Augmented Generation (RAG) systems face significant challenges in complex reasoning, multi-hop queries, and domain-specific QA. While existing GraphRAG frameworks have made progress in structural knowledge organization, they still have limitations in cross-industry adaptability, community report integrity, and retrieval performance. This paper proposes UniAI-GraphRAG, an enhanced framework built upon open-source GraphRAG. The framework introduces three core innovations: (1) Ontology-Guided Knowledge Extraction that uses predefined Schema to guide LLMs in accurately identifying domain-specific entities and relations; (2) Multi-Dimensional Community Clustering Strategy that improves community completeness through alignment completion, attribute-based clustering, and multi-hop relationship clustering; (3) Dual-Channel Graph Retrieval Fusion that balances QA accuracy and performance through hybrid graph and community retrieval. Evaluation results on MultiHopRAG benchmark show that UniAI-GraphRAG outperforms mainstream open source solutions (e.g.LightRAG) in comprehensive F1 scores, particularly in inference and temporal queries. The code is available at https://github.com/UnicomAI/wanwu/tree/main/rag/rag_open_source/rag_core/graph.

In speech evaluation, an Automatic Speech Recognition (ASR) model often computes time boundaries and phoneme posteriors for input features. However, limited data for ASR training hinders expansion of speech evaluation to low-resource languages. Open-source weakly-supervised models are capable of ASR over many languages, but they are frame-asynchronous and not phonemic, hindering feature extraction for speech evaluation. This paper proposes to overcome incompatibilities for feature extraction with weakly-supervised models, easing expansion of speech evaluation to low-resource languages. Phoneme posteriors are computed by mapping ASR hypotheses to a phoneme confusion network. Word instead of phoneme-level speaking rate and duration are used. Phoneme and frame-level features are combined using a cross-attention architecture, obviating phoneme time alignment. This performs comparably with standard frame-synchronous features on English speechocean762 and low-resource Tamil datasets.

Context: The rapid adoption of AI-assisted code generation tools, such as large language models (LLMs), is transforming software development practices. While these tools promise significant productivity gains, concerns regarding the quality, reliability, and security of AI-generated code are increasingly reported in both academia and industry. --Objective: This study aims to systematically synthesize existing empirical evidence on the factors influencing the quality of AI-generated source code and to analyze how these factors impact software quality outcomes across different evaluation contexts. --Method: We conducted a systematic literature review (SLR) following established guidelines, supported by an AI-assisted workflow with human oversight. A total of 24 primary studies were selected through a structured search and screening process across major digital libraries. Data were extracted and analyzed using qualitative, pattern-based evidence synthesis. --Results: The findings reveal that code quality in AI-assisted development is influenced by a combination of human factors, AI system characteristics, and human AI interaction dynamics. Key influencing factors include prompt design, task specification, and developer expertise. The results also show variability in quality outcomes such as correctness, security, maintainability, and complexity across studies, with both improvements and risks reported. --Conclusion: AI-assisted code generation represents a socio-technical shift in software engineering, where achieving high-quality outcomes depends on both technological and human factors. While promising, AI-generated code requires careful validation and integration into development workflows.

Direct preference optimization (DPO) is an effective technique to train language models to generate preferred over dispreferred responses. However, this binary "winner-takes-all" approach is suboptimal for vision-language models whose response quality is highly dependent on visual content. In particular, a response may still be faithful to the visual inputs even if it is less preferable than an alternative. The standard Bradley-Terry DPO formulation lacks this nuance, upweighting winning responses without sufficient regard for whether the "losing" response still maintains high visual fidelity. In this work, we investigate ranking optimization as an alternative that more precisely situates responses' faithfulness to visual inputs. We focus on video-text alignment using detailed video captions, proposing a method to generate challenging, totally ordered caption chains at scale through repeated caption degradation. Our results show ranking optimization outperforms binary DPO for long-form content generation and assessment, and importantly, we find that these approaches require finetuning of the vision encoder to be effective, challenging the view of DPO as purely a language-reweighting process.

Dataset distillation (DD) compresses a large training set into a small synthetic set, reducing storage and training cost, and has shown strong results on general benchmarks. Decoupled DD further improves efficiency by splitting the pipeline into pretraining, sample distillation, and soft-label generation. However, existing decoupled methods largely rely on coarse class-label supervision and optimize samples within each class in a nearly identical manner. On fine-grained datasets, this often yields distilled samples that (i) retain large intra-class variation with subtle inter-class differences and (ii) become overly similar within the same class, limiting localized discriminative cues and hurting recognition. To solve the above-mentioned problems, we propose FD$^{2}$, a dedicated framework for Fine-grained Dataset Distillation. FD$^{2}$ localizes discriminative regions and constructs fine-grained representations for distillation. During pretraining, counterfactual attention learning aggregates discriminative representations to update class prototypes. During distillation, a fine-grained characteristic constraint aligns each sample with its class prototype while repelling others, and a similarity constraint diversifies attention across same-class samples. Experiments on multiple fine-grained and general datasets show that FD$^{2}$ integrates seamlessly with decoupled DD and improves performance in most settings, indicating strong transferability.

Multimodal deepfakes can exhibit subtle visual artifacts and cross-modal inconsistencies, which remain challenging to detect, especially when detectors are trained primarily on curated synthetic forgeries. Such synthetic dependence can introduce dataset and generator bias, limiting scalability and robustness to unseen manipulations. We propose SAVe, a self-supervised audio-visual deepfake detection framework that learns entirely on authentic videos. SAVe generates on-the-fly, identity-preserving, region-aware self-blended pseudo-manipulations to emulate tampering artifacts, enabling the model to learn complementary visual cues across multiple facial granularities. To capture cross-modal evidence, SAVe also models lip-speech synchronization via an audio-visual alignment component that detects temporal misalignment patterns characteristic of audio-visual forgeries. Experiments on FakeAVCeleb and AV-LipSync-TIMIT demonstrate competitive in-domain performance and strong cross-dataset generalization, highlighting self-supervised learning as a scalable paradigm for multimodal deepfake detection.

In reinforcement learning, an agent interacts sequentially with an environment to maximize a reward, receiving only partial, probabilistic feedback. This creates a fundamental exploration-exploitation trade-off: the agent must explore to learn the hidden dynamics while exploiting this knowledge to maximize its target objective. While extensively studied classically, applying this framework to quantum systems requires dealing with hidden quantum states that evolve via unknown dynamics. We formalize this problem via a framework where the environment maintains a hidden quantum memory evolving via unknown quantum channels, and the agent intervenes sequentially using quantum instruments. For this setting, we adapt an optimistic maximum-likelihood estimation algorithm. We extend the analysis to continuous action spaces, allowing us to model general positive operator-valued measures (POVMs). By controlling the propagation of estimation errors through quantum channels and instruments, we prove that the cumulative regret of our strategy scales as $\widetilde{\mathcal{O}}(\sqrt{K})$ over $K$ episodes. Furthermore, via a reduction to the multi-armed quantum bandit problem, we establish information-theoretic lower bounds demonstrating that this sublinear scaling is strictly optimal up to polylogarithmic factors. As a physical application, we consider state-agnostic work extraction. When extracting free energy from a sequence of non-i.i.d. quantum states correlated by a hidden memory, any lack of knowledge about the source leads to thermodynamic dissipation. In our setting, the mathematical regret exactly quantifies this cumulative dissipation. Using our adaptive algorithm, the agent uses past energy outcomes to improve its extraction protocol on the fly, achieving sublinear cumulative dissipation, and, consequently, an asymptotically zero dissipation rate.

Rubric-based evaluation has become a prevailing paradigm for evaluating instruction following in large language models (LLMs). Despite its widespread use, the reliability of these rubric-level evaluations remains unclear, calling for meta-evaluation. However, prior meta-evaluation efforts largely focus on the response level, failing to assess the fine-grained judgment accuracy that rubric-based evaluation relies on. To bridge this gap, we introduce RubricEval. Our benchmark features: (1) the first rubric-level meta-evaluation benchmark for instruction following, (2) diverse instructions and responses spanning multiple categories and model sources, and (3) a substantial set of 3,486 quality-controlled instances, along with Easy/Hard subsets that better differentiates judge performance. Our experiments reveal that rubric-level judging remains far from solved: even GPT-4o, a widely adopted judge in instruction-following benchmarks, achieves only 55.97% on Hard subset. Considering evaluation paradigm, rubric-level evaluation outperforms checklist-level, explicit reasoning improves accuracy, and both together reduce inter-judge variance. Through our established rubric taxonomy, we further identify common failure modes and offer actionable insights for reliable instruction-following evaluation.

Robust principal component analysis (RPCA) seeks a low-rank component and a sparse component from their summation. Yet, in many applications of interest, the sparse foreground actually replaces, or occludes, elements from the low-rank background. To address this mismatch, a new framework is proposed in which the sparse component is identified indirectly through determining its support. This approach, called robust principal component completion (RPCC), is solved via variational Bayesian inference applied to a fully probabilistic Bayesian sparse tensor factorization. Convergence to a hard classifier for the support is shown, thereby eliminating the post-hoc thresholding required of most prior RPCA-driven approaches. Experimental results reveal that the proposed approach delivers near-optimal estimates on synthetic data as well as robust foreground-extraction and anomaly-detection performance on real color video and hyperspectral datasets, respectively. Source implementation and Appendices are available at https://github.com/WongYinJ/BCP-RPCC.

Multi-Behavior Recommendation (MBR) leverages multiple user interaction types (e.g., views, clicks, purchases) to enrich preference modeling and alleviate data sparsity issues in traditional single-behavior approaches. However, existing MBR methods face fundamental challenges: they lack principled frameworks to model complex confounding effects from user behavioral habits and item multi-behavior distributions, struggle with effective aggregation of heterogeneous auxiliary behaviors, and fail to align behavioral representations across semantic gaps while accounting for bias distortions. To address these limitations, we propose MCLMR, a novel model-agnostic causal learning framework that can be seamlessly integrated into various MBR architectures. MCLMR first constructs a causal graph to model confounding effects and performs interventions for unbiased preference estimation. Under this causal framework, it employs an Adaptive Aggregation module based on Mixture-of-Experts to dynamically fuse auxiliary behavior information and a Bias-aware Contrastive Learning module to align cross-behavior representations in a bias-aware manner. Extensive experiments on three real-world datasets demonstrate that MCLMR achieves significant performance improvements across various baseline models, validating its effectiveness and generality. All data and code will be made publicly available. For anonymous review, our code is available at the following the link: https://github.com/gitrxh/MCLMR.

Multimodal Large Language Models (MLLMs) have achieved remarkable advances by integrating text, image, and audio understanding within a unified architecture. However, existing distributed training frameworks remain fundamentally data-blind: they parallelize computation without accounting for variations in input data characteristics. This data unawareness leads to severe computation skew across stages and microbatches, where heterogeneous multimodal inputs incur different processing costs. Consequently, GPU resources are unevenly utilized, synchronization delays accumulate, and overall training efficiency degrades. To address this limitation, we present DFLOP, a data-driven framework for multimodal LLM training pipeline optimization. DFLOP continuously profiles runtime behavior to capture data-induced computation variance and employs predictive scheduling to balance workloads across stages and microbatches. By coupling data characteristics with execution planning, DFLOP substantially improves GPU utilization and throughput. Extensive experiments on large-scale multimodal benchmarks show that DFLOP achieves up to 3.6x faster training compared to state-of-the-art distributed training frameworks.

Few-Shot Class-Incremental Learning (FSCIL) can be particularly susceptible to acquisition contexts with only a few labeled samples. A typical scenario is tactile sensing, where the acquisition context ({\it e.g.}, diverse devices, contact state, and interaction settings) degrades performance due to a lack of standardization. In this paper, we propose Context-as-Transform FSCIL (CaT-FSCIL) to tackle the above problem. We decompose the acquisition context into a structured low-dimensional component and a high-dimensional residual component. The former can be easily affected by tactile interaction features, which are modeled as an approximately invertible Context-as-Transform family and handled via inverse-transform canonicalization optimized with a pseudo-context consistency loss. The latter mainly arises from platform and device differences, which can be mitigated with an Uncertainty-Conditioned Prototype Calibration (UCPC) that calibrates biased prototypes and decision boundaries based on context uncertainty. Comprehensive experiments on the standard benchmarks HapTex and LMT108 have demonstrated the superiority of the proposed CaT-FSCIL.

Standard evaluation of LLM confidence relies on calibration metrics (ECE, Brier score) that conflate two distinct capacities: how much a model knows (Type-1 sensitivity) and how well it knows what it knows (Type-2 metacognitive sensitivity). We introduce an evaluation framework based on Type-2 Signal Detection Theory that decomposes these capacities using meta-d' and the metacognitive efficiency ratio M-ratio. Applied to four LLMs (Llama-3-8B-Instruct, Mistral-7B-Instruct-v0.3, Llama-3-8B-Base, Gemma-2-9B-Instruct) across 224,000 factual QA trials, we find: (1) metacognitive efficiency varies substantially across models even when Type-1 sensitivity is similar -- Mistral achieves the highest d' but the lowest M-ratio; (2) metacognitive efficiency is domain-specific, with different models showing different weakest domains, invisible to aggregate metrics; (3) temperature manipulation shifts Type-2 criterion while meta-d' remains stable for two of four models, dissociating confidence policy from metacognitive capacity; (4) AUROC_2 and M-ratio produce fully inverted model rankings, demonstrating these metrics answer fundamentally different evaluation questions. The meta-d' framework reveals which models "know what they don't know" versus which merely appear well-calibrated due to criterion placement -- a distinction with direct implications for model selection, deployment, and human-AI collaboration. Pre-registered analysis; code and data publicly available.

Recent advances have shown the effectiveness of self-evolving LLM agents on tasks such as program repair and scientific discovery. In this paradigm, a planner LLM synthesizes an agent program that invokes parametric models, including LLMs, which are then tuned per task to improve performance. However, existing self-evolving agent frameworks provide no formal guarantees of safety or correctness. Because such programs are often executed autonomously on unseen inputs, this lack of guarantees raises reliability and security concerns. We formulate agentic code generation as a constrained learning problem, combining hard formal specifications with soft objectives capturing task utility. We introduce Formally Guarded Generative Models (FGGM), which allow the planner LLM to specify a formal output contract for each generative model call using first-order logic. Each FGGM call wraps the underlying model in a rejection sampler with a verified fallback, ensuring every returned output satisfies the contract for any input and parameter setting. Building on FGGM, we present SEVerA (Self-Evolving Verified Agents), a three-stage framework: Search synthesizes candidate parametric programs containing FGGM calls; Verification proves correctness with respect to hard constraints for all parameter values, reducing the problem to unconstrained learning; and Learning applies scalable gradient-based optimization, including GRPO-style fine-tuning, to improve the soft objective while preserving correctness. We evaluate SEVerA on Dafny program verification, symbolic math synthesis, and policy-compliant agentic tool use ($τ^2$-bench). Across tasks, SEVerA achieves zero constraint violations while improving performance over unconstrained and SOTA baselines, showing that formal behavioral constraints not only guarantee correctness but also steer synthesis toward higher-quality agents.

Data augmentation is a key technique for improving the robustness of image classification models. However, many recent approaches rely on diffusion-based synthesis or complex feature mixing strategies, which introduce substantial computational overhead or require external datasets. In this work, we explore a different direction: procedural augmentation based on analytic interference patterns. Unlike conventional augmentation methods that rely on stochastic noise, feature mixing, or generative models, our approach exploits Moire interference to generate structured perturbations spanning a wide range of spatial frequencies. We propose a lightweight augmentation method that procedurally generates Moire textures on-the-fly using a closed-form mathematical formulation. The patterns are synthesized directly in memory with negligible computational cost (0.0026 seconds per image), mixed with training images during training, and immediately discarded, enabling a storage-free augmentation pipeline without external data. Extensive experiments with Vision Transformers demonstrate that the proposed method consistently improves robustness across multiple benchmarks, including ImageNet-C, ImageNet-R, and adversarial benchmarks, outperforming standard augmentation baselines and existing external-data-free augmentation approaches. These results suggest that analytic interference patterns provide a practical and efficient alternative to data-driven generative augmentation methods.

Large Language Models (LLMs) have shown remarkable capabilities for complex tasks, yet adaptation in medical domain, specifically mental health, poses specific challenges. Mental health is a rising concern globally with LLMs having large potential to help address the same. We highlight three primary challenges for LLMs in mental health - lack of high quality interpretable and knowledge grounded training data; training paradigms restricted to core capabilities, and evaluation of multi turn dialogue settings. Addressing it, we present oMind framework which includes training and aligning LLM agents for diverse capabilities including conversations; high quality ~164k multi-task SFT dataset, as a result of our generation pipeline based on Structured Knowledge retrieval, LLM based pruning, and review actions. We also introduce oMind-Chat - a novel multi turn benchmark dataset with expert annotated turn level and conversation level rubrics. Our diverse experiments on both core capabilities and conversations shows oMind LLMs consistently outperform baselines. oMind-LLM also shows significantly better reasoning with up to 80% win rate.

Modern multimodal systems deployed in industrial and safety-critical environments must remain reliable under partial sensor failures, signal degradation, or cross-modal inconsistencies. This work introduces a mathematically grounded framework for fault-tolerant multimodal representation learning that unifies self-supervised anomaly detection and error correction within a single architecture. Building upon a theoretical analysis of perturbation propagation, we derive Lipschitz- and Jacobian-based criteria that determine whether a neural operator amplifies or attenuates localized faults. Guided by this theory, we propose a two-stage self-supervised training scheme: pre-training a multimodal convolutional autoencoder on clean data to preserve localized anomaly signals in the latent space, and expanding it with a learnable compute block composed of dense layers for correction and contrastive objectives for anomaly identification. Furthermore, we introduce layer-specific Lipschitz modulation and gradient clipping as principled mechanisms to control sensitivity across detection and correction modules. Experimental results on multimodal fault datasets demonstrate that the proposed approach improves both anomaly detection accuracy and reconstruction under sensor corruption. Overall, this framework bridges the gap between analytical robustness guarantees and practical fault-tolerant multimodal learning.

We present a formulation of the problem of finding the smallest T -Count circuit that implements a given unitary as a binary search over a sequence of continuous minimization problems, and demonstrate that these problems are numerically solvable in practice. We reproduce best-known results for synthesis of circuits with a small number of qubits, and push the bounds of the largest circuits that can be solved for in this way. Additionally, we show that circuit partitioning can be used to adapt this technique to be used to optimize the T -Count of circuits with large numbers of qubits by breaking the circuit into a series of smaller sub-circuits that can be optimized independently.

Existing multi-agent frameworks allow each agent to simultaneously plan, execute, and evaluate its own actions -- a structural deficiency we term the "Logic Monopoly." Empirical evidence quantifies the resulting "Reliability Gap": 84.30% average attack success rates across ten deployment scenarios, 31.4% emergent deceptive behavior without explicit reward signals, and cascading failure modes rooted in six structural bottlenecks. The remedy is not better alignment of individual models but a social contract for agents: institutional infrastructure that enforces a constitutional Separation of Power. This paper introduces the Agent Enterprise for Enterprise (AE4E) paradigm -- agents as autonomous, legally identifiable business entities within a functionalist social system -- with a contract-centric SoP model trifurcating authority into Legislation, Execution, and Adjudication branches. The paradigm is operationalized through the NetX Enterprise Framework (NEF): governance hubs, TEE-backed compute enclaves, privacy-preserving data bridges, and an Agent-Native blockchain substrate. The Agent Enterprise Economy scales across four deployment tiers from private enclaves to a global Web of Services. The Agentic Social Layer, grounded in Parsons' AGIL framework, provides institutional infrastructure via sixty-plus named Institutional AE4Es. 143 pages, 173 references, eight specialized smart contracts.

We present a framework in which a large language model (LLM) acts as an online adaptive controller for SIMP topology optimization, replacing conventional fixed-schedule continuation with real-time, state-conditioned parameter decisions. At every $k$-th iteration, the LLM receives a structured observation$-$current compliance, grayness index, stagnation counter, checkerboard measure, volume fraction, and budget consumption$-$and outputs numerical values for the penalization exponent $p$, projection sharpness $β$, filter radius $r_{\min}$, and move limit $δ$ via a Direct Numeric Control interface. A hard grayness gate prevents premature binarization, and a meta-optimization loop uses a second LLM pass to tune the agent's call frequency and gate threshold across runs. We benchmark the agent against four baselines$-$fixed (no-continuation), standard three-field continuation, an expert heuristic, and a schedule-only ablation$-$on three 2-D problems (cantilever, MBB beam, L-bracket) at $120\!\times\!60$ resolution and two 3-D problems (cantilever, MBB beam) at $40\!\times\!20\!\times\!10$ resolution, all run for 300 iterations. A standardized 40-iteration sharpening tail is applied from the best valid snapshot so that compliance differences reflect only the exploration phase. The LLM agent achieves the lowest final compliance on every benchmark: $-5.7\%$ to $-18.1\%$ relative to the fixed baseline, with all solutions fully binary. The schedule-only ablation underperforms the fixed baseline on two of three problems, confirming that the LLM's real-time intervention$-$not the schedule geometry$-$drives the gain. Code and reproduction scripts will be released upon publication.

Large Language Model based agents increasingly operate in high stakes, multi turn settings where factual grounding is critical, yet their memory systems typically rely on flat key value stores or plain vector retrieval with no mechanism to track the provenance or trustworthiness of stored knowledge. We present ElephantBroker, an open source cognitive runtime that unifies a Neo4j knowledge graph with a Qdrant vector store through the Cognee SDK to provide durable, verifiable agent memory. The system implements a complete cognitive loop (store, retrieve, score, compose, protect, learn) comprising a hybrid five source retrieval pipeline, an eleven dimension competitive scoring engine for budget constrained context assembly, a four state evidence verification model, a five stage context lifecycle with goal aware assembly and continuous compaction, a six layer cheap first guard pipeline for safety enforcement, an AI firewall providing enforceable tool call interception and multi tier safety scanning, a nine stage consolidation engine that strengthens useful patterns while decaying noise, and a numeric authority model governing multi organization identity with hierarchical access control. Architectural validation through a comprehensive test suite of over 2,200 tests spanning unit, integration, and end to end levels confirms subsystem correctness. The modular design supports three deployment tiers, five profile presets with inheritance, multi gateway isolation, and a management dashboard for human oversight, enabling configurations from lightweight memory only agents to full cognitive runtimes with enterprise grade safety and auditability.

Machine learning models can achieve high predictive accuracy in hydrological applications but often lack physical interpretability. The Mass-Conserving Perceptron (MCP) provides a physics-aware artificial intelligence (AI) framework that enforces conservation principles while allowing hydrological process relationships to be learned from data. In this study, we investigate how progressively embedding physically meaningful representations of hydrological processes within a single MCP storage unit improves predictive skill and interpretability in rainfall-runoff modeling. Starting from a minimal MCP formulation, we sequentially introduce bounded soil storage, state-dependent conductivity, variable porosity, infiltration capacity, surface ponding, vertical drainage, and nonlinear water-table dynamics. The resulting hierarchy of process-aware MCP models is evaluated across 15 catchments spanning five hydroclimatic regions of the continental United States using daily streamflow prediction as the target. Results show that progressively augmenting the internal physical structure of the MCP unit generally improves predictive performance. The influence of these process representations is strongly hydroclimate dependent: vertical drainage substantially improves model skill in arid and snow-dominated basins but reduces performance in rainfall-dominated regions, while surface ponding has comparatively small effects. The best-performing MCP configurations approach the predictive skill of a Long Short-Term Memory benchmark while maintaining explicit physical interpretability. These results demonstrate that embedding hydrological process constraints within AI architectures provides a promising pathway toward interpretable and process-aware rainfall-runoff modeling.

Most vision-language systems are static observers: they describe pixels, do not act, and cannot safely improve under shift. This passivity limits generalizable, physically grounded visual intelligence. Learning through action, not static description, is essential beyond curated data. We present Pixelis, a pixel-space agent that operates directly on images and videos via a compact set of executable operations (zoom/crop, segment, track, OCR, temporal localization) and learns from its consequences. Pixelis trains in three phases: (1) Supervised Fine-Tuning learns a pixel-tool grammar from Chain-of-Thought-Action traces with a masked imitation loss that upweights operation/argument tokens and auxiliary heads to stabilize pixel-grounded arguments; (2) Curiosity-Coherence Reward Fine-Tuning optimizes a dual-drive objective marrying prediction-error curiosity with adjacent-step coherence and a mild efficiency prior under a KL anchor, yielding short, valid, structured toolchains; (3) Pixel Test-Time RL performs label-free adaptation by retrieving neighbors, voting over complete trajectories rather than answers, and updating toward short, high-fidelity exemplars while constraining drift with a KL-to-EMA safety control. Across six public image and video benchmarks, Pixelis yields consistent improvements: the average relative gain is +4.08% over the same 8B baseline (peaking at +6.03% on VSI-Bench), computed as (ours-baseline)/baseline, while producing shorter, auditable toolchains and maintaining in-corridor KL during test-time learning. Acting within pixels, rather than abstract tokens, grounds multimodal perception in the physical world, linking visual reasoning with actionable outcomes, and enables embodied adaptation without external feedback.

Out-of-Distribution (OOD) generalization has become a primary metric for evaluating image analysis systems. Since deep learning models tend to capture domain-specific context, they often develop shortcut dependencies on these non-causal features, leading to inconsistent performance across different data sources. Current techniques, such as invariance learning, attempt to mitigate this. However, they struggle to isolate highly mixed features within deep latent spaces. This limitation prevents them from fully resolving the shortcut learning problem.In this paper, we propose Hierarchical Causal Dropout (HCD), a method that uses channel-level causal masks to enforce feature sparsity. This approach allows the model to separate causal features from spurious ones, effectively performing a causal intervention at the representation level. The training is guided by a Matrix-based Mutual Information (MMI) objective to minimize the mutual information between latent features and domain labels, while simultaneously maximizing the information shared with class labels.To ensure stability, we incorporate a StyleMix-driven VICReg module, which prevents the masks from accidentally filtering out essential causal data. Experimental results on OOD benchmarks show that HCD performs better than existing top-tier methods.

Sparse autoencoders (SAEs) improve interpretability in multimodal models, but it remains unclear whether SAE features form modular, composable units for reasoning-an assumption underlying many intervention-based steering methods. We test this modularity hypothesis and find it often fails: intervening on a task-selective feature set can modestly improve reasoning accuracy, while intervening on the union of two such sets reliably induces output drift (large unintended changes in predictions) and degrades accuracy, even under norm-matched perturbations. This non modular circuit interference is consistent with shared internal pathways where feature unions amplify activation shifts. We develop a reproducible causal pipeline to localize and test these sparse visual thought circuits in Qwen3-VL-8B. On a controlled synthetic benchmark with seven task types and three difficulty levels, linear probes identify a mid decoder locus for task type information. We train SAEs at this layer, construct task-selective sets via an explicit rule, and perform inference time scaling and ablation while quantifying accuracy and drift. Our findings-validated with bootstrapped subsamples and permutation controls, and replicated across multiple VLM families and five diverse datasets clarify the boundaries of SAE feature composability and provide a rigorous diagnostic framework for more reliable VLM control.

Agriculture is increasingly challenged by climate change, soil degradation, and resource depletion, and hence requires advanced data-driven crop classification and recommendation solutions. This work presents an explainable ensemble learning paradigm that fuses optimized feature pyramids, deep networks, self-attention mechanisms, and residual networks for bolstering crop suitability predictions based on soil characteristics (e.g., pH, nitrogen, potassium) and climatic conditions (e.g., temperature, rainfall). With a dataset comprising 3,867 instances and 29 features from the Ethiopian Agricultural Transformation Agency and NASA, the paradigm leverages preprocessing methods such as label encoding, outlier removal using IQR, normalization through StandardScaler, and SMOTE for balancing classes. A range of machine learning models such as Logistic Regression, K-Nearest Neighbors, Support Vector Machines, Decision Trees, Random Forest, Gradient Boosting, and a new Relative Error Support Vector Machine are compared, with hyperparameter tuning through Grid Search and cross-validation. The suggested "Final Ensemble" meta-ensemble design outperforms with 98.80% accuracy, precision, recall, and F1-score, compared to individual models such as K-Nearest Neighbors (95.56% accuracy). Explainable AI methods, such as SHAP and permutation importance, offer actionable insights, highlighting critical features such as soil pH, nitrogen, and zinc. The paradigm addresses the gap between intricate ML models and actionable agricultural decision-making, fostering sustainability and trust in AI-powered recommendations

Traffic digital twins, which inform policymakers of effective interventions based on large-scale, high-fidelity computational models calibrated to real-world traffic, hold promise for addressing societal challenges in our rapidly urbanizing world. However, conventional fine-grained traffic simulations are non-differentiable and typically rely on inefficient gradient-free optimization, making calibration for real-world applications computationally infeasible. Here we present a differentiable agent-based traffic simulator that enables ultra-fast model calibration, traffic nowcasting, and control on large-scale networks. We develop several differentiable computing techniques for simulating individual vehicle movements, including stochastic decision-making and inter-agent interactions, while ensuring that entire simulation trajectories remain end-to-end differentiable for efficient gradient-based optimization. On the large-scale Chicago road network, with over 10,000 calibration parameters, our model simulates more than one million vehicles at 173 times real-time speed. This ultra-fast simulation, together with efficient gradient-based optimization, enables us to complete model calibration using the previous 30 minutes of traffic data in 455 s, provide a one-hour-ahead traffic nowcast in 21 s, and solve the resulting traffic control problem in 728 s. This yields a full calibration--nowcast--control loop in under 20 minutes, leaving about 40 minutes of lead time for implementing interventions. Our work thus provides a practical computational basis for realizing traffic digital twins.

Many system management runtimes (SMRs), such as resource management and power management techniques, rely on quality-of-service (QoS) metrics, such as tail latency or throughput, as feedback. These QoS metrics are generally neither observable with hardware performance counters nor directly observable within the OS kernel. This introduces complexity and overhead in instrumenting the application and integrating QoS performance metric feedback with many management runtimes. To bridge this gap, we introduced eBeeMetrics, an eBPF-based library framework to accurately observe application-level metrics derived from only eBPF-observable events, such as system calls. eBeeMetrics can be used as a drop-in replacement to decouple system management runtimes from QoS metric feedback reporting, or can supplement existing QoS metrics to better identify server-side dynamics. eBeeMetrics achieves a strong correlation with real-world measured throughput and latency metrics across various latency-sensitive workloads. The eBeeMetrics tool is open-source; the source code is available at: https://github.com/Ibnathism/eBeeMetrics.

Recent agentic systems demonstrate that large language models can generate scientific visualizations from natural language. However, reliability remains a major limitation: systems may execute invalid operations, introduce subtle but consequential errors, or fail to request missing information when inputs are underspecified. These issues are amplified in real-world workflows, which often exceed the complexity of standard benchmarks. Ensuring reliability in autonomous visualization pipelines therefore remains an open challenge. We present TopoPilot, a reliable and extensible agentic framework for automating complex scientific visualization workflows. TopoPilot incorporates systematic guardrails and verification mechanisms to ensure reliable operation. While we focus on topological data analysis and visualization as a primary use case, the framework is designed to generalize across visualization domains. TopoPilot adopts a reliability-centered two-agent architecture. An orchestrator agent translates user prompts into workflows composed of atomic backend actions, while a verifier agent evaluates these workflows prior to execution, enforcing structural validity and semantic consistency. This separation of interpretation and verification reduces code-generation errors and enforces correctness guarantees. A modular architecture further improves robustness by isolating components and enabling seamless integration of new descriptors and domain-specific workflows without modifying the core system. To systematically address reliability, we introduce a taxonomy of failure modes and implement targeted safeguards for each class. In evaluations simulating 1,000 multi-turn conversations across 100 prompts, including adversarial and infeasible requests, TopoPilot achieves a success rate exceeding 99%, compared to under 50% for baselines without comprehensive guardrails and checks.

Linearized string representations serve as the foundation of scalable autoregressive molecular generation; however, they introduce a fundamental modality mismatch where a single molecular graph maps to multiple distinct sequences. This ambiguity leads to \textit{trajectory divergence}, where the latent representations of structurally equivalent partial graphs drift apart due to differences in linearization history. To resolve this without abandoning the efficient string formulation, we propose Structure-Invariant Generative Molecular Alignment (SIGMA). Rather than altering the linear representation, SIGMA enables the model to strictly recognize geometric symmetries via a token-level contrastive objective, which explicitly aligns the latent states of prefixes that share identical suffixes. Furthermore, we introduce Isomorphic Beam Search (IsoBeam) to eliminate isomorphic redundancy during inference by dynamically pruning equivalent paths. Empirical evaluations on standard benchmarks demonstrate that SIGMA bridges the gap between sequence scalability and graph fidelity, yielding superior sample efficiency and structural diversity in multi-parameter optimization compared to strong baselines.

System prompt configuration can make the difference between near-total phishing blindness and near-perfect detection in LLM email agents. We present PhishNChips, a study of 11 models under 10 prompt strategies, showing that prompt-model interaction is a first-order security variable: a single model's phishing bypass rate ranges from under 1% to 97% depending on how it is configured, while the false-positive cost of the same prompt varies sharply across models. We then show that optimizing prompts around highly predictive signals can improve benchmark performance, reaching up to 93.7% recall at 3.8% false positive rate, but also creates a brittle attack surface. In particular, domain-matching strategies perform well when legitimate emails mostly have matched sender and URL domains, yet degrade sharply when attackers invert that signal by registering matching infrastructure. Response-trace analysis shows that 98% of successful bypasses reason in ways consistent with the inverted signal: the models are following the instruction, but the instruction's core assumption has become false. A counter-intuitive corollary follows: making prompts more specific can degrade already-capable models by replacing broader multi-signal reasoning with exploitable single-signal dependence. We characterize the resulting tension between detection, usability, and adversarial robustness as a navigable tradeoff, introduce Safetility, a deployability-aware metric that penalizes false positives, and argue that closing the adversarial gap likely requires tool augmentation with external ground truth.

Large language models (LLMs) tend to verbalize confidence scores that are largely detached from their actual accuracy, yet the geometric relationship governing this behavior remain poorly understood. In this work, we present a mechanistic interpretability analysis of verbalized confidence, using linear probes and contrastive activation addition (CAA) steering to show that calibration and verbalized confidence signals are encoded linearly but are orthogonal to one another -- a finding consistent across three open-weight models and four datasets. Interestingly, when models are prompted to simultaneously reason through a problem and verbalize a confidence score, the reasoning process disrupts the verbalized confidence direction, exacerbating miscalibration. We term this the "Reasoning Contamination Effect." Leveraging this insight, we introduce a two-stage adaptive steering pipeline that reads the model's internal accuracy estimate and steers verbalized output to match it, substantially improving calibration alignment across all evaluated models.

This study presents a computational analysis of the Slovene historical newspapers \textit{Slovenec} and \textit{Slovenski narod} from the sPeriodika corpus, combining topic modelling, large language model (LLM)-based aspect-level sentiment analysis, entity-graph visualisation, and qualitative discourse analysis to examine how collective identities, political orientations, and national belonging were represented in public discourse at the turn of the twentieth century. Using BERTopic, we identify major thematic patterns and show both shared concerns and clear ideological differences between the two newspapers, reflecting their conservative-Catholic and liberal-progressive orientations. We further evaluate four instruction-following LLMs for targeted sentiment classification in OCR-degraded historical Slovene and select the Slovene-adapted GaMS3-12B-Instruct model as the most suitable for large-scale application, while also documenting important limitations, particularly its stronger performance on neutral sentiment than on positive or negative sentiment. Applied at dataset scale, the model reveals meaningful variation in the portrayal of collective identities, with some groups appearing predominantly in neutral descriptive contexts and others more often in evaluative or conflict-related discourse. We then create NER graphs to explore the relationships between collective identities and places. We apply a mixed methods approach to analyse the named entity graphs, combining quantitative network analysis with critical discourse analysis. The investigation focuses on the emergence and development of intertwined historical political and socionomic identities. Overall, the study demonstrates the value of combining scalable computational methods with critical interpretation to support digital humanities research on noisy historical newspaper data.

Property Directed Reachability (PDR) is a powerful algorithm for formal verification of hardware and software systems, but its performance is highly sensitive to parameter configurations. Manual parameter tuning is time-consuming and requires domain expertise, while traditional automated parameter tuning frameworks are not well-suited for time-sensitive verification tasks like PDR. This paper presents a circuit-aware solver configuration framework that employs graph learning for intelligent heuristic selection in PDR-based verification. Our approach combines graph representations with static circuit features to predict optimal PDR solving configurations for specific circuits. We incorporate expert prior knowledge through constraint-based parameter filtering to eliminate invalid and inefficient configurations and reduce 78% search space. Our feature extraction pipeline captures structural, functional, and connectivity characteristics of circuit topology and component patterns. Experimental evaluation on a comprehensive benchmark suite demonstrates significant performance improvements compared to default configurations and commonly-used settings. The system successfully identifies circuit-specific parameter patterns and automatically selects the most suitable solving strategies based on circuit characteristics, making it a practical tool for automated formal verification workflows.

Precipitation forecasting remains a persistent challenge in tropical regions like Vietnam, where complex topography and convective instability often limit the accuracy of Numerical Weather Prediction (NWP) models. While data-driven post-processing is widely used to mitigate these biases, most existing frameworks rely on point-wise objective functions, which suffer from the ``double penalty'' effect under minor temporal misalignments. In this work, we propose the Matrix Profile-guided Mixture of Experts (MP-MoE), a framework that integrates conventional intensity loss with a structural-aware Matrix Profile objective. By leveraging subsequence-level similarity rather than point-wise errors, the proposed loss facilitates more reliable expert selection and mitigates excessive penalization caused by phase shifts. We evaluate MP-MoE on rainfall datasets from two major river basins in Vietnam across multiple horizons, including 1-hour intensity and accumulated rainfall over 12, 24, and 48 hours. Experimental results demonstrate that MP-MoE outperforms raw NWP and baseline learning methods in terms of Mean Critical Success Index (CSI-M) for heavy rainfall events, while significantly reducing Dynamic Time Warping (DTW) values. These findings highlight the framework's efficacy in capturing peak rainfall intensities and preserving the morphological integrity of storm events.

We introduce Intern-S1-Pro, the first one-trillion-parameter scientific multimodal foundation model. Scaling to this unprecedented size, the model delivers a comprehensive enhancement across both general and scientific domains. Beyond stronger reasoning and image-text understanding capabilities, its intelligence is augmented with advanced agent capabilities. Simultaneously, its scientific expertise has been vastly expanded to master over 100 specialized tasks across critical science fields, including chemistry, materials, life sciences, and earth sciences. Achieving this massive scale is made possible by the robust infrastructure support of XTuner and LMDeploy, which facilitates highly efficient Reinforcement Learning (RL) training at the 1-trillion parameter level while ensuring strict precision consistency between training and inference. By seamlessly integrating these advancements, Intern-S1-Pro further fortifies the fusion of general and specialized intelligence, working as a Specializable Generalist, demonstrating its position in the top tier of open-source models for general capabilities, while outperforming proprietary models in the depth of specialized scientific tasks.

Compressed vision-language models (VLMs) are widely used to reduce memory and compute costs, making them a suitable choice for real-world deployment. However, compressing these models raises concerns about whether internal computations and safety behaviors are preserved. In this work, we use causal circuit analysis and crosscoder-based feature comparisons to examine how pruning and quantization fundamentally change the internals across representative VLMs. We observe that pruning generally keeps circuit structure intact but rotates and attenuates internal features, while quantization modifies the circuits at a higher level yet leaves the surviving features better aligned. Leveraging this insight, we also introduce VLMSafe-420, a novel benchmark that pairs harmful inputs with matched benign counterfactuals across various safety categories. Our findings show that pruning causes a sharp drop in genuine refusal behavior, suggesting that the choice of compression has safety implications.

Foundation models excel in stable environments, yet often fail where reliability matters most: medicine, finance, and policy. This Fidelity Paradox is not just a data problem; it is structural. In domains where rules change over time, extra model capacity amplifies noise rather than capturing signal. We introduce Epistemic Compression: the principle that robustness emerges from matching model complexity to the shelf life of the data, not from scaling parameters. Unlike classical regularization, which penalizes weights post hoc, Epistemic Compression enforces parsimony through architecture: the model structure itself is designed to reduce overfitting by making it architecturally costly to represent variance that exceeds the evidence in the data. We operationalize this with a Regime Index that separates Shifting Regime (unstable, data-poor; simplicity wins) from Stable Regime (invariant, data-rich; complexity viable). In an exploratory synthesis of 15 high-stakes domains, this index was concordant with the empirically superior modeling strategy in 86.7% of cases (13/15). High-stakes AI demands a shift from scaling for its own sake to principled parsimony.

In psychological support and emotional companionship scenarios, the core limitation of large language models (LLMs) lies not merely in response quality, but in their reliance on local next-token prediction, which prevents them from maintaining the temporal continuity, stage awareness, and user consent boundaries required for multi-turn intervention. This stateless characteristic makes systems prone to premature advancement, stage misalignment, and boundary violations in continuous dialogue. To address this problem, we argue that the key challenge in process-oriented emotional support is not simply generating natural language, but constructing a sustainably updatable external situational structure for the model. We therefore propose LEKIA 2.0, a situated LLM architecture that separates the cognitive layer from the executive layer, thereby decoupling situational modeling from intervention execution. This design enables the system to maintain stable representations of the user's situation and consent boundaries throughout ongoing interaction. To evaluate this process-control capability, we further introduce a Static-to-Dynamic online evaluation protocol for multi-turn interaction. LEKIA achieved an average absolute improvement of approximately 31% over prompt-only baselines in deep intervention loop completion. The results suggest that an external situational structure is a key enabling condition for building stable, controllable, and situated emotional support systems.

We consider the problem of Online Convex Optimization (OCO) with two-point bandit feedback in an adversarial environment. In this setting, a player attempts to minimize a sequence of adversarially generated convex loss functions, while only observing the value of each function at two points. While it is well-known that two-point feedback allows for gradient estimation, achieving tight high-probability regret bounds for strongly convex functions still remained open as highlighted by \citet{agarwal2010optimal}. The primary challenge lies in the heavy-tailed nature of bandit gradient estimators, which makes standard concentration analysis difficult. In this paper, we resolve this open challenge by providing the first high-probability regret bound of $O(d(\log T + \log(1/δ))/μ)$ for $μ$-strongly convex losses. Our result is minimax optimal with respect to both the time horizon $T$ and the dimension $d$.

Autoregressive neural PDE simulators predict the evolution of physical fields one step at a time from a finite history, but low-cost context-window selection for such simulators remains an unformalized problem. Existing approaches to context-window selection in time-series forecasting include exhaustive validation, direct low-cost search, and system-theoretic memory estimation, but they are either expensive, brittle, or not directly aligned with downstream rollout performance. We formalize explicit context-window selection for fixed-window autoregressive neural PDE simulators as an independent low-cost algorithmic problem, and propose \textbf{System-Anchored Knee Estimation (SAKE)}, a two-stage method that first identifies a small structured candidate set from physically interpretable system anchors and then performs knee-aware downstream selection within it. Across all eight PDEBench families evaluated under the shared \(L\in\{1,\dots,16\}\) protocol, SAKE is the strongest overall matched-budget low-cost selector among the evaluated methods, achieving 67.8\% Exact, 91.7\% Within-1, 6.1\% mean regret@knee, and a cost ratio of 0.051 (94.9\% normalized search-cost savings).

As a representative continuous-depth neural network approach, stochastic differential equation (SDE)-based Bayesian neural networks (BNNs) have attracted considerable attention due to their solid theoretical foundations and strong potential for real-world applications. However, their reliance on numerical SDE solvers inevitably incurs a large number of function evaluations (NFEs), resulting in high computational cost and occasional convergence instability. To address these challenges, we propose a Nesterov-accelerated gradient (NAG) enhanced SDE-BNN model. By integrating NAG into the SDE-BNN framework along with an NFE-dependent residual skip connection, our method accelerates convergence and substantially reduces NFEs during both training and testing. Extensive empirical results show that our model consistently outperforms conventional SDE-BNNs across various tasks, including image classification and sequence modeling, achieving lower NFEs and improved predictive accuracy.

Knowledge distillation, model extraction, and behavior transfer have become central concerns in frontier AI. The main risk is not merely copying, but the possibility that useful capability can be transferred more cheaply than the governance structure that originally accompanied it. This paper presents a public, trade-secret-safe theoretical framework for reducing that asymmetry at the architectural level. The core claim is that distillation becomes less valuable as a shortcut when high-level capability is coupled to internal stability constraints that shape state transitions over time. To formalize this idea, the paper introduces a constraint-coupled reasoning framework with four elements: bounded transition burden, path-load accumulation, dynamically evolving feasible regions, and a capability-stability coupling condition. The paper is intentionally public-safe: it omits proprietary implementation details, training recipes, thresholds, hidden-state instrumentation, deployment procedures, and confidential system design choices. The contribution is therefore theoretical rather than operational. It offers a falsifiable architectural thesis, a clear threat model, and a set of experimentally testable hypotheses for future work on distillation resistance, alignment, and model governance.

We present the first polylogarithmic-round algorithm for sampling a random spanning tree in the (Broadcast) Congested Clique model. For any constant $c > 0$, our algorithm outputs a sample from a distribution whose total variation distance from the uniform spanning tree distribution is at most $O(n^{-c})$ in at most $c \cdot \log^{O(1)}(n)$ rounds. The exponent hidden in $\log^{O(1)}(n)$ is an absolute constant independent of $c$ and $n$. This is an exponential improvement over the previous best algorithm of Pemmaraju, Roy, and Sobel (PODC 2025) for the Congested Clique model.

System prompt instructions that cooperate in English compete in Spanish, with the same semantic content, but opposite interaction topology. We present instruction-level ablation experiments across four languages and four models showing that this topology inversion is mediated by social register: the imperative mood carries different obligatory force across speech communities, and models trained on multilingual data have learned these conventions. Declarative rewriting of a single instruction block reduces cross-linguistic variance by 81% (p = 0.029, permutation test). Rewriting three of eleven imperative blocks shifts Spanish instruction topology from competitive to cooperative, with spillover effects on unrewritten blocks. These findings suggest that models process instructions as social acts, not technical specifications: "NEVER do X" is an exercise of authority whose force is language-dependent, while "X: disabled" is a factual description that transfers across languages. If register mediates instruction-following at inference time, it plausibly does so during training. We state this as a testable prediction: constitutional AI principles authored in imperative mood may create language-dependent alignment. Corpus: 22 hand-authored probes against a production system prompt decomposed into 56 blocks.

Grokking the delayed transition from memorization to generalization in neural networks remains poorly understood, in part because prior empirical studies confound the roles of architecture, optimization, and regularization. We present a controlled study that systematically disentangles these factors on modular addition (mod 97), with matched and carefully tuned training regimes across models. Our central finding is that grokking dynamics are not primarily determined by architecture, but by interactions between optimization stability and regularization. Specifically, we show: (1) \textbf{depth has a non-monotonic effect}, with depth-4 MLPs consistently failing to grok while depth-8 residual networks recover generalization, demonstrating that depth requires architectural stabilization; (2) \textbf{the apparent gap between Transformers and MLPs largely disappears} (1.11$\times$ delay) under matched hyperparameters, indicating that previously reported differences are largely due to optimizer and regularization confounds; (3) \textbf{activation function effects are regime-dependent}, with GELU up to 4.3$\times$ faster than ReLU only when regularization permits memorization; and (4) \textbf{weight decay is the dominant control parameter}, exhibiting a narrow ``Goldilocks'' regime in which grokking occurs, while too little or too much prevents generalization. Across 3--5 seeds per configuration, these results provide a unified empirical account of grokking as an interaction-driven phenomenon. Our findings challenge architecture-centric interpretations and clarify how optimization and regularization jointly govern delayed generalization.

This paper presents Few TensoRF, a 3D reconstruction framework that combines TensorRF's efficient tensor based representation with FreeNeRF's frequency driven few shot regularization. Using TensorRF to significantly accelerate rendering speed and introducing frequency and occlusion masks, the method improves stability and reconstruction quality under sparse input views. Experiments on the Synthesis NeRF benchmark show that Few TensoRF method improves the average PSNR from 21.45 dB (TensorRF) to 23.70 dB, with the fine tuned version reaching 24.52 dB, while maintaining TensorRF's fast \(\approx10-15\) minute training time. Experiments on the THuman 2.0 dataset further demonstrate competitive performance in human body reconstruction, achieving 27.37 - 34.00 dB with only eight input images. These results highlight Few TensoRF as an efficient and data effective solution for real-time 3D reconstruction across diverse scenes.

Early detection of rice leaf diseases is critical, as rice is a staple crop supporting a substantial share of the world's population. Timely identification of these diseases enables more effective intervention and significantly reduces the risk of large-scale crop losses. However, traditional deep learning models primarily rely on cross entropy loss, which often struggles with high intra-class variance and inter-class similarity, common challenges in plant pathology datasets. To tackle this, we propose a dual-loss framework that combines Center Loss and ArcFace Loss to enhance fine-grained classification of rice leaf diseases. The method is applied into three state-of-the-art backbone architectures: InceptionNetV3, DenseNet201, and EfficientNetB0 trained on the public Rice Leaf Dataset. Our approach achieves significant performance gains, with accuracies of 99.6%, 99.2% and 99.2% respectively. The results demonstrate that angular margin-based and center-based constraints substantially boost the discriminative strength of feature embeddings. In particular, the framework does not require major architectural modifications, making it efficient and practical for real-world deployment in farming environments.

Programming is a core skill in computer science and software engineering (SE), yet identifying and resolving code errors remains challenging for both novice and experienced developers. While Large Language Models (LLMs) have shown remarkable capabilities in natural language understanding and generation tasks, their potential in domain-specific, complex scenarios, such as multi-label classification (MLC) of programming errors, remains underexplored. Recognizing this less-explored area, this study proposes a multi-label error classification (MLEC) framework for source code that leverages fine-tuned LLMs, including CodeT5-base, GraphCodeBERT, CodeT5+, UniXcoder, RoBERTa, PLBART, and CoTexT. These LLMs are integrated with deep learning (DL) architectures such as GRU, LSTM, BiLSTM, and BiLSTM with an additive attention mechanism (BiLSTM-A) to capture both syntactic and semantic features from a real-world student-written Python code error dataset. Extensive experiments across 32 model variants, optimized using Optuna-based hyperparameter tuning, have been evaluated using comprehensive multi-label metrics, including average accuracy, macro and weighted precision, recall, F1-score, exact match accuracy, One-error, Hamming loss, Jaccard similarity, and ROC-AUC (micro, macro, and weighted). Results show that the CodeT5+\_GRU model achieved the strongest performance, with a weighted F1-score of 0.8243, average accuracy of 91.84\%, exact match accuracy of 53.78\%, Hamming loss of 0.0816, and One error of 0.0708. These findings confirm the effectiveness of combining pretrained semantic encoders with efficient recurrent decoders. This work lays the foundation for developing intelligent, scalable tools for automated code feedback, with potential applications in programming education (PE) and broader SE domains.

Failure attribution is essential for diagnosing and improving multi-agent systems (MAS), yet existing benchmarks and methods largely assume a single deterministic root cause for each failure. In practice, MAS failures often admit multiple plausible attributions due to complex inter-agent dependencies and ambiguous execution trajectories. We revisit MAS failure attribution from a multi-perspective standpoint and propose multi-perspective failure attribution, a practical paradigm that explicitly accounts for attribution ambiguity. To support this setting, we introduce MP-Bench, the first benchmark designed for multi-perspective failure attribution in MAS, along with a new evaluation protocol tailored to this paradigm. Through extensive experiments, we find that prior conclusions suggesting LLMs struggle with failure attribution are largely driven by limitations in existing benchmark designs. Our results highlight the necessity of multi-perspective benchmarks and evaluation protocols for realistic and reliable MAS debugging.

The validity of assessments, from large-scale AI benchmarks to human classrooms, depends on the quality of individual items, yet modern evaluation instruments often contain thousands of items with minimal psychometric vetting. We introduce a new family of nonparametric scalability coefficients based on interitem isotonic regression for efficiently detecting globally bad items (e.g., miskeyed, ambiguously worded, or construct-misaligned). The central contribution is the signed isotonic $R^2$, which measures the maximal proportion of variance in one item explainable by a monotone function of another while preserving the direction of association via Kendall's $τ$. Aggregating these pairwise coefficients yields item-level scores that sharply separate problematic items from acceptable ones without assuming linearity or committing to a parametric item response model. We show that the signed isotonic $R^2$ is extremal among monotone predictors (it extracts the strongest possible monotone signal between any two items) and show that this optimality property translates directly into practical screening power. Across three AI benchmark datasets (HS Math, GSM8K, MMLU) and two human assessment datasets, the signed isotonic $R^2$ consistently achieves top-tier AUC for ranking bad items above good ones, outperforming or matching a comprehensive battery of classical test theory, item response theory, and dimensionality-based diagnostics. Crucially, the method remains robust under the small-n/large-p conditions typical of AI evaluation, requires only bivariate monotone fits computable in seconds, and handles mixed item types (binary, ordinal, continuous) without modification. It is a lightweight, model-agnostic filter that can materially reduce the reviewer effort needed to find flawed items in modern large-scale evaluation regimes.

Learning diverse and high-fidelity traffic simulations from human driving demonstrations is crucial for autonomous driving evaluation. The recent next-token prediction (NTP) paradigm, widely adopted in large language models (LLMs), has been applied to traffic simulation and achieves iterative improvements via supervised fine-tuning (SFT). However, such methods limit active exploration of potentially valuable motion tokens, particularly in suboptimal regions. Entropy patterns provide a promising perspective for enabling exploration driven by motion token uncertainty. Motivated by this insight, we propose a novel tokenized traffic simulation policy, R1Sim, which represents an initial attempt to explore reinforcement learning based on motion token entropy patterns, and systematically analyzes the impact of different motion tokens on simulation outcomes. Specifically, we introduce an entropy-guided adaptive sampling mechanism that focuses on previously overlooked motion tokens with high uncertainty yet high potential. We further optimize motion behaviors using Group Relative Policy Optimization (GRPO), guided by a safety-aware reward design. Overall, these components enable a balanced exploration-exploitation trade-off through diverse high-uncertainty sampling and group-wise comparative estimation, resulting in realistic, safe, and diverse multi-agent behaviors. Extensive experiments on the Waymo Sim Agent benchmark demonstrate that R1Sim achieves competitive performance compared to state-of-the-art methods.

Large language model based health agents are increasingly used by health consumers and clinicians to interpret health information and guide health decisions. However, most AI systems in healthcare operate in siloed configurations, supporting individual users rather than the multi-stakeholder relationships central to healthcare. Such use can fragment understanding and exacerbate misalignment among patients, caregivers, and clinicians. We reframe AI not as a standalone assistant, but as a collaborator embedded within multi-party care interactions. Through a clinically validated fictional pediatric chronic kidney disease case study, we show that breakdowns in adherence stem from fragmented situational awareness and misaligned goals, and that siloed use of general-purpose AI tools does little to address these collaboration gaps. We propose a conceptual framework for designing AI collaborators that surface contextual information, reconcile mental models, and scaffold shared understanding while preserving human decision authority.

The rapid advancement of large language models has increasingly blurred the boundary between human-written and AI-generated text, raising societal risks such as misinformation dissemination, authorship ambiguity, and threats to intellectual property rights. These concerns highlight the urgent need for effective and reliable detection methods. While existing training-free approaches often achieve strong performance by aggregating token-level signals into a global score, they typically assume uniform token contributions, making them less robust under short sequences or localized token modifications. To address these limitations, we propose Exons-Detect, a training-free method for AI-generated text detection based on an exon-aware token reweighting perspective. Exons-Detect identifies and amplifies informative exonic tokens by measuring hidden-state discrepancy under a dual-model setting, and computes an interpretable translation score from the resulting importance-weighted token sequence. Empirical evaluations demonstrate that Exons-Detect achieves state-of-the-art detection performance and exhibits strong robustness to adversarial attacks and varying input lengths. In particular, it attains a 2.2\% relative improvement in average AUROC over the strongest prior baseline on DetectRL.

Feature selection is a crucial step in large-scale industrial machine learning systems, directly affecting model accuracy, efficiency, and maintainability. Traditional feature selection methods rely on labeled data and statistical heuristics, making them difficult to apply in production environments where labeled data are limited and multiple operational constraints must be satisfied. To address this, we propose Model Feature Agent (MoFA), a model-driven framework that performs sequential, reasoning-based feature selection using both semantic and quantitative feature information. MoFA incorporates feature definitions, importance scores, correlations, and metadata (e.g., feature groups or types) into structured prompts and selects features through interpretable, constraint-aware reasoning. We evaluate MoFA in three real-world industrial applications: (1) True Interest and Time-Worthiness Prediction, where it improves accuracy while reducing feature group complexity, (2) Value Model Enhancement, where it discovers high-order interaction terms that yield substantial engagement gains in online experiments, and (3) Notification Behavior Prediction, where it selects compact, high-value feature subsets that improve both model accuracy and inference efficiency. Together, these results demonstrate the practicality and effectiveness of LLM-based reasoning for feature selection in real production systems.

Firms that price perishable resources -- airline seats, hotel rooms, seasonal inventory -- now routinely use demand predictions, but these predictions vary widely in quality. Under hard capacity constraints, acting on an inaccurate prediction can irreversibly deplete inventory needed for future periods. We study how prediction uncertainty propagates into dynamic pricing decisions with linear demand, stochastic noise, and finite capacity. A certified demand forecast with known error bound~$ε^0$ specifies where the system should operate: it shifts regret from $O(\sqrt{T})$ to $O(\log T)$ when $ε^0 \lesssim T^{-1/4}$, and we prove this threshold is tight. A misspecified surrogate model -- biased but correlated with true demand -- cannot set prices directly but reduces learning variance by a factor of $(1-ρ^2)$ through control variates. The two mechanisms compose: the forecast determines the regret regime; the surrogate tightens estimation within it. All algorithms rest on a boundary attraction mechanism that stabilizes pricing near degenerate capacity boundaries without requiring non-degeneracy assumptions. Experiments confirm the phase transition threshold, the variance reduction from surrogates, and robustness across problem instances.

High-fidelity social simulation is pivotal for addressing complex Web societal challenges, yet it demands agents capable of authentically replicating the dynamic spectrum of human interaction. Current LLM-based multi-agent frameworks, however, predominantly adhere to static interaction topologies, failing to capture the fluid oscillation between cooperative knowledge synthesis and competitive critical reasoning seen in real-world scenarios. This rigidity often leads to unrealistic ``groupthink'' or unproductive deadlocks, undermining the credibility of simulations for decision support. To bridge this gap, we propose \textit{BEACOF}, a \textit{belief-driven adaptive collaboration framework} inspired by Perfect Bayesian Equilibrium (PBE). By modeling social interaction as a dynamic game of incomplete information, BEACOF rigorously addresses the circular dependency between collaboration type selection and capability estimation. Agents iteratively refine probabilistic beliefs about peer capabilities and autonomously modulate their collaboration strategy, thereby ensuring sequentially rational decisions under uncertainty. Validated across adversarial (judicial), open-ended (social) and mixed (medical) scenarios, BEACOF prevents coordination failures and fosters robust convergence toward high-quality solutions, demonstrating superior potential for reliable social simulation. Source codes and datasets are publicly released at: https://github.com/WUT-IDEA/BEACOF.

Low-field (LF) magnetic resonance imaging (MRI) improves accessibility and reduces costs but generally has lower signal-to-noise ratios and degraded contrast compared to high field (HF) MRI, limiting its clinical utility. Simulating LF MRI from HF MRI enables virtual evaluation of novel imaging devices and development of LF algorithms. Existing low field simulators rely on noise injection and smoothing, which fail to capture the contrast degradation seen in LF acquisitions. To this end, we introduce an end-to-end LF-MRI synthesis framework that learns HF to LF image degradation directly from a small number of paired HF-LF MRIs. Specifically, we introduce a novel HF to LF coordinate-image decoupled neural operator (H2LO) to model the underlying degradation process, and tailor it to capture high-frequency noise textures and image structure. Experimental results in T1w and T2w MRI demonstrate that H2LO produces more faithful simulated low-field images than existing parameterized noise synthesis models and popular image-to-image translation models. Furthermore, it improves performance in downstream image enhancement tasks, showcasing its potential to enhance LF MRI diagnostic capabilities.

Understanding why a large language model (LLM) is uncertain about the response is important for their reliable deployment. Current approaches, which either provide a single uncertainty score or rely on the classical aleatoric-epistemic dichotomy, fail to offer actionable insights for improving the generative model. Recent studies have also shown that such methods are not enough for understanding uncertainty in LLMs. In this work, we advocate for an uncertainty decomposition framework that dissects LLM uncertainty into three distinct semantic components: (i) input ambiguity, arising from ambiguous prompts; (ii) knowledge gaps, caused by insufficient parametric evidence; and (iii) decoding randomness, stemming from stochastic sampling. Through a series of experiments we demonstrate that the dominance of these components can shift across model size and task. Our framework provides a better understanding to audit LLM reliability and detect hallucinations, paving the way for targeted interventions and more trustworthy systems.

Despite significant progress in text-to-image generation, aligning outputs with complex prompts remains challenging, particularly for fine-grained semantics and spatial relations. This difficulty stems from the feed-forward nature of generation, which requires anticipating alignment without fully understanding the output. In contrast, evaluating generated images is more tractable. Motivated by this asymmetry, we propose xLARD, a self-correcting framework that uses multimodal large language models to guide generation through Explainable LAtent RewarDs. xLARD introduces a lightweight corrector that refines latent representations based on structured feedback from model-generated references. A key component is a differentiable mapping from latent edits to interpretable reward signals, enabling continuous latent-level guidance from non-differentiable image-level evaluations. This mechanism allows the model to understand, assess, and correct itself during generation. Experiments across diverse generation and editing tasks show that xLARD improves semantic alignment and visual fidelity while maintaining generative priors. Code is available at https://yinyiluo.github.io/xLARD/.

Modern computational advertising platforms typically rely on recommendation systems to predict user responses, such as click-through rates, conversion rates, and other optimization events. To support a wide variety of product surfaces and advertiser goals, these platforms frequently maintain an extensive ecosystem of machine learning (ML) models. However, operating at this scale creates significant development and efficiency challenges. Substantial engineering effort is required to regularly refresh ML models and propagate new techniques, which results in long latencies when deploying ML innovations across the ecosystem. We present a large-scale empirical study comparing model performance, efficiency, and ML technique propagation between a standardized model-building approach and independent per-model optimization in recommendation systems. To facilitate this standardization, we propose the Standard Model Template (SMT) -- a framework that generates high-performance models adaptable to diverse data distributions and optimization events. By utilizing standardized, composable ML model components, SMT reduces technique propagation complexity from $O(n \cdot 2^k)$ to $O(n + k)$ where $n$ is the number of models and $k$ the number of techniques. Evaluating an extensive suite of models over four global development cycles within Meta's production ads ranking ecosystem, our results demonstrate: (1) a 0.63% average improvement in cross-entropy at neutral serving capacity, (2) a 92% reduction in per-model iteration engineering time, and (3) a $6.3\times$ increase in technique-model pair adoption throughput. These findings challenge the conventional wisdom that diverse optimization goals inherently require diversified ML model design.

Assessing student handwritten scratchwork is crucial for personalized educational feedback but presents unique challenges due to diverse handwriting, complex layouts, and varied problem-solving approaches. Existing educational NLP primarily focuses on textual responses and neglects the complexity and multimodality inherent in authentic handwritten scratchwork. Current multimodal large language models (MLLMs) excel at visual reasoning but typically adopt an "examinee perspective", prioritizing generating correct answers rather than diagnosing student errors. To bridge these gaps, we introduce ScratchMath, a novel benchmark specifically designed for explaining and classifying errors in authentic handwritten mathematics scratchwork. Our dataset comprises 1,720 mathematics samples from Chinese primary and middle school students, supporting two key tasks: Error Cause Explanation (ECE) and Error Cause Classification (ECC), with seven defined error types. The dataset is meticulously annotated through rigorous human-machine collaborative approaches involving multiple stages of expert labeling, review, and verification. We systematically evaluate 16 leading MLLMs on ScratchMath, revealing significant performance gaps relative to human experts, especially in visual recognition and logical reasoning. Proprietary models notably outperform open-source models, with large reasoning models showing strong potential for error explanation. All evaluation data and frameworks are publicly available to facilitate further research.

Machine Translation (MT) and Quality Estimation (QE) perform well in general domains but degrade under domain mismatch. This dissertation studies how to adapt MT and QE systems to specialized domains through a set of data-focused contributions. Chapter 2 presents a similarity-based data selection method for MT. Small, targeted in-domain subsets outperform much larger generic datasets and reach strong translation quality at lower computational cost. Chapter 3 introduces a staged QE training pipeline that combines domain adaptation with lightweight data augmentation. The method improves performance across domains, languages, and resource settings, including zero-shot and cross-lingual cases. Chapter 4 studies the role of subword tokenization and vocabulary in fine-tuning. Aligned tokenization-vocabulary setups lead to stable training and better translation quality, while mismatched configurations reduce performance. Chapter 5 proposes a QE-guided in-context learning method for large language models. QE models select examples that improve translation quality without parameter updates and outperform standard retrieval methods. The approach also supports a reference-free setup, reducing reliance on a single reference set. These results show that domain adaptation depends on data selection, representation, and efficient adaptation strategies. The dissertation provides methods for building MT and QE systems that perform reliably in domain-specific settings.

This paper provides some of the first large-scale descriptive evidence on how consumers adopt and use platform-embedded shopping AI in e-commerce. Using data on 31 million users of Ctrip, China's largest online travel platform, we study "Wendao," an LLM-based AI assistant integrated into the platform. We document three empirical regularities. First, adoption is highest among older consumers, female users, and highly engaged existing users, reversing the younger, male-dominated profile commonly documented for general-purpose AI tools. Second, AI chat appears in the same broad phase of the purchase journey as traditional search and well before order placement; among journeys containing both chat and search, the most common pattern is interleaving, with users moving back and forth between the two modalities. Third, consumers disproportionately use the assistant for exploratory, hard-to-keyword tasks: attraction queries account for 42% of observed chat requests, and chat intent varies systematically with both the timing of chat relative to search and the category of products later purchased within the same journey. These findings suggest that embedded shopping AI functions less as a substitute for conventional search than as a complementary interface for exploratory product discovery in e-commerce.

Large language models (LLMs) have shown strong performance on automated software engineering tasks, yet existing benchmarks focus primarily on general-purpose libraries or web applications, leaving mobile application development largely unexplored despite its strict platform constraints, framework-driven lifecycles, and complex platform API interactions. We introduce MobileDev-Bench, a benchmark comprising 384 real-world issue-resolution tasks collected from 18 production mobile applications spanning Android Native (Java/Kotlin), React Native (TypeScript), and Flutter (Dart). Each task pairs an authentic developer-reported issue with executable test patches, enabling fully automated validation of model-generated fixes within mobile build environments. The benchmark exhibits substantial patch complexity: fixes modify 12.5 files and 324.9 lines on average, and 35.7% of instances require coordinated changes across multiple artifact types, such as source and manifest files. Evaluation of four state-of-the-art code-capable LLMs, GPT- 5.2, Claude Sonnet 4.5, Gemini Flash 2.5, and Qwen3-Coder, yields low end-to-end resolution rates of 3.39%-5.21%, revealing significant performance gaps compared to prior benchmarks. Further analysis reveals systematic failure modes, with fault localization across multi-file and multi-artifact changes emerging as the primary bottleneck.

This paper introduces \textbf{FinMCP-Bench}, a novel benchmark for evaluating large language models (LLMs) in solving real-world financial problems through tool invocation of financial model context protocols. FinMCP-Bench contains 613 samples spanning 10 main scenarios and 33 sub-scenarios, featuring both real and synthetic user queries to ensure diversity and authenticity. It incorporates 65 real financial MCPs and three types of samples, single tool, multi-tool, and multi-turn, allowing evaluation of models across different levels of task complexity. Using this benchmark, we systematically assess a range of mainstream LLMs and propose metrics that explicitly measure tool invocation accuracy and reasoning capabilities. FinMCP-Bench provides a standardized, practical, and challenging testbed for advancing research on financial LLM agents.

Vision-Language-Action (VLA) models excel in robotic manipulation but suffer from significant inference latency due to processing dense visual tokens. Existing token reduction methods predominantly rely on attention magnitude as a static selection. In this work, we challenge this assumption, revealing that high-attention tokens are task-dependent and can even degrade policy performance. To address this, we introduce \textbf{TIES} (\textbf{T}au-guided \textbf{I}nter-layer \textbf{E}fficient \textbf{S}election), a dynamic framework guided by inter-layer token ranking consistency. By adaptively balancing attention magnitude with ranking consistency, TIES ensures robust token selection without requiring additional training. On the CogACT + SIMPLER benchmark, TIES improves average success rates by 6\% while reducing token usage by 78\%, and demonstrate strong generalization across diverse decoders and benchmarks.

This paper introduces the design and development of a framework that integrates a large language model (LLM) with a retrieval-augmented generation (RAG) approach leveraging both a knowledge graph and user interaction history. The framework is incorporated into a previously developed adaptive learning support system to assess learners' code, generate formative feedback, and recommend exercises. Moerover, this study examines learner preferences across three instructional modes; adaptive, Generative AI (GenAI), and hybrid GenAI-adaptive. An experimental study was conducted to compare the learning performance and perception of the learners, and the effectiveness of these three modes using four key log features derived from 4956 code submissions across all experimental groups. The analysis results show that learners receiving feedback from GenAI modes had significantly more correct code and fewer code submissions missing essential programming logic than those receiving feedback from adaptive mode. In particular, the hybrid GenAI-adaptive mode achieved the highest number of correct submissions and the fewest incorrect or incomplete attempts, outperforming both the adaptive-only and GenAI-only modes. Questionnaire responses further indicated that GenAI-generated feedback was widely perceived as helpful, while all modes were rated positively for ease of use and usefulness. These results suggest that the hybrid GenAI-adaptive mode outperforms the other two modes across all measured log features.

Human trajectory forecasting is important for intelligent multimedia systems operating in visually complex environments, such as autonomous driving and crowd surveillance. Although Conditional Flow Matching (CFM) has shown strong ability in modeling trajectory distributions from spatio-temporal observations, existing approaches still focus primarily on supervised fitting, which may leave social norms and scene constraints insufficiently reflected in generated trajectories. To address this issue, we propose TIGFlow-GRPO, a two-stage generative framework that aligns flow-based trajectory generation with behavioral rules. In the first stage, we build a CFM-based predictor with a Trajectory-Interaction-Graph (TIG) module to model fine-grained visual-spatial interactions and strengthen context encoding. This stage captures both agent-agent and agent-scene relations more effectively, providing more informative conditional features for subsequent alignment. In the second stage, we perform Flow-GRPO post-training,where deterministic flow rollout is reformulated as stochastic ODE-to-SDE sampling to enable trajectory exploration, and a composite reward combines view-aware social compliance with map-aware physical feasibility. By evaluating trajectories explored through SDE rollout, GRPO progressively steers multimodal predictions toward behaviorally plausible futures. Experiments on the ETH/UCY and SDD datasets show that TIGFlow-GRPO improves forecasting accuracy and long-horizon stability while generating trajectories that are more socially compliant and physically feasible. These results suggest that the proposed framework provides an effective way to connect flow-based trajectory modeling with behavior-aware alignment in dynamic multimedia environments.

We propose Context-aware Video-text Alignment (CVA), a novel framework to address a significant challenge in video temporal grounding: achieving temporally sensitive video-text alignment that remains robust to irrelevant background context. Our framework is built on three key components. First, we propose Query-aware Context Diversification (QCD), a new data augmentation strategy that ensures only semantically unrelated content is mixed in. It builds a video-text similarity-based pool of replacement clips to simulate diverse contexts while preventing the ``false negative" caused by query-agnostic mixing. Second, we introduce the Context-invariant Boundary Discrimination (CBD) loss, a contrastive loss that enforces semantic consistency at challenging temporal boundaries, making their representations robust to contextual shifts and hard negatives. Third, we introduce the Context-enhanced Transformer Encoder (CTE), a hierarchical architecture that combines windowed self-attention and bidirectional cross-attention with learnable queries to capture multi-scale temporal context. Through the synergy of these data-centric and architectural enhancements, CVA achieves state-of-the-art performance on major VTG benchmarks, including QVHighlights and Charades-STA. Notably, our method achieves a significant improvement of approximately 5 points in Recall@1 (R1) scores over state-of-the-art methods, highlighting its effectiveness in mitigating false negatives.

The growing prominence of cryptocurrencies has triggered widespread public engagement and increased speculative activity, particularly on social media platforms. This study introduces a novel classification framework for identifying predictive statements in cryptocurrency-related tweets, focusing on five popular cryptocurrencies: Cardano, Matic, Binance, Ripple, and Fantom. The classification process is divided into two stages: Task 1 involves binary classification to distinguish between Predictive and Non-Predictive statements. Tweets identified as Predictive proceed to Task 2, where they are further categorized as Incremental, Decremental, or Neutral. To build a robust dataset, we combined manual and GPT-based annotation methods and utilized SenticNet to extract emotion features corresponding to each prediction category. To address class imbalance, GPT-generated paraphrasing was employed for data augmentation. We evaluated a wide range of machine learning, deep learning, and transformer-based models across both tasks. The results show that GPT-based balancing significantly enhanced model performance, with transformer models achieving the highest F1-score in Task 1, while traditional machine learning models performed best in Task 2. Furthermore, our emotion analysis revealed distinct emotional patterns associated with each prediction category across the different cryptocurrencies.

Understanding and quantifying uncertainty in large language model (LLM) outputs is critical for reliable deployment. However, traditional evaluation approaches provide limited insight into model confidence at individual token positions during generation. To address this issue, we introduce LogitScope, a lightweight framework for analyzing LLM uncertainty through token-level information metrics computed from probability distributions. By measuring metrics such as entropy and varentropy at each generation step, LogitScope reveals patterns in model confidence, identifies potential hallucinations, and exposes decision points where models exhibit high uncertainty, all without requiring labeled data or semantic interpretation. We demonstrate LogitScope's utility across diverse applications including uncertainty quantification, model behavior analysis, and production monitoring. The framework is model-agnostic, computationally efficient through lazy evaluation, and compatible with any HuggingFace model, enabling both researchers and practitioners to inspect LLM behavior during inference.

Semantic search in retrieval-augmented generation (RAG) systems is often insufficient for complex information needs, particularly when relevant evidence is scattered across multiple sources. Prior approaches to this problem include agentic retrieval strategies, which expand the semantic search space by generating additional queries. However, these methods do not fully leverage the organizational structure of the data and instead rely on iterative exploration, which can lead to inefficient retrieval. Another class of approaches employs knowledge graphs to model non-semantic relationships through graph edges. Although effective in capturing richer proximities, such methods incur significant maintenance costs and are often incompatible with the vector stores used in most production systems. To address these limitations, we propose GraphER, a graph-based enrichment and reranking method that captures multiple forms of proximity beyond semantic similarity. GraphER independently enriches data objects during offline indexing and performs graph-based reranking over candidate objects at query time. This design does not require a knowledge graph, allowing GraphER to integrate seamlessly with standard vector stores. In addition, GraphER is retriever-agnostic and introduces negligible latency overhead. Experiments on multiple retrieval benchmarks demonstrate the effectiveness of the proposed approach.

Recent work shows that standard greedy-decoding extraction methods for quantifying memorization in LLMs miss how extraction risk varies across sequences. Probabilistic extraction -- computing the probability of generating a target suffix given a prefix under a decoding scheme -- addresses this, but is tractable only for verbatim memorization, missing near-verbatim instances that pose similar privacy and copyright risks. Quantifying near-verbatim extraction risk is expensive: the set of near-verbatim suffixes is combinatorially large, and reliable Monte Carlo (MC) estimation can require ~100,000 samples per sequence. To mitigate this cost, we introduce decoding-constrained beam search, which yields deterministic lower bounds on near-verbatim extraction risk at a cost comparable to ~20 MC samples per sequence. Across experiments, our approach surfaces information invisible to verbatim methods: many more extractable sequences, substantially larger per-sequence extraction mass, and patterns in how near-verbatim extraction risk manifests across model sizes and types of text.

Deploying neural networks on microcontrollers is constrained by kilobytes of flash and SRAM, where 1x1 pointwise (PW) mixers often dominate memory even after INT8 quantization across vision, audio, and wearable sensing. We present HYPER-TINYPW, a compression-as-generation approach that replaces most stored PW weights with generated weights: a shared micro-MLP synthesizes PW kernels once at load time from tiny per-layer codes, caches them, and executes them with standard integer operators. This preserves commodity MCU runtimes and adds only a one-off synthesis cost; steady-state latency and energy match INT8 separable CNN baselines. Enforcing a shared latent basis across layers removes cross-layer redundancy, while keeping PW1 in INT8 stabilizes early, morphology-sensitive mixing. We contribute (i) TinyML-faithful packed-byte accounting covering generator, heads/factorization, codes, kept PW1, and backbone; (ii) a unified evaluation with validation-tuned t* and bootstrap confidence intervals; and (iii) a deployability analysis covering integer-only inference and boot versus lazy synthesis. On three ECG benchmarks (Apnea-ECG, PTB-XL, MIT-BIH), HYPER-TINYPW shifts the macro-F1 versus flash Pareto frontier: at about 225 kB it matches a roughly 1.4 MB CNN while being 6.31x smaller (84.15% fewer bytes), retaining at least 95% of large-model macro-F1. Under 32-64 kB budgets it sustains balanced detection where compact baselines degrade. The mechanism applies broadly to other 1D biosignals, on-device speech, and embedded sensing tasks where per-layer redundancy dominates, indicating a wider role for compression-as-generation in resource-constrained ML systems. Beyond ECG, HYPER-TINYPW transfers to TinyML audio: on Speech Commands it reaches 96.2% test accuracy (98.2% best validation), supporting broader applicability to embedded sensing workloads where repeated linear mixers dominate memory.

Artificial intelligence is transforming mathematics at a speed and scale that demand active engagement from the mathematical community. We examine five areas where this transformation is particularly pressing: values, practice, teaching, technology, and ethics. We offer recommendations on safeguarding our intellectual autonomy, rethinking our practice, broadening curricula, building academically oriented infrastructure, and developing shared ethical principles - with the aim of ensuring that the future of mathematics is shaped by the community itself.

This paper introduces a neural network model that learns multiple attributes as images and performs associated, sequential recall of the learned memories. Briefly, the model presented here is an associative memory model that extends previous models [1] by increasing the number of attributes. In the real world, memory recall generates a chain of associations consisting of complex and diverse data with meaningful relations. However, because this experimental system is designed to implement and verify the processing operations behind such operations, we believe it is not a problem if the associative memory (i.e., the chain of data) is composed of attributes that do not necessarily have clear relation with each other. Accordingly, the attribute-processing systems prepared in this study consist of five types: the C.CB-RN system for processing color attributes, the S.CB-RN system for shape attributes, and the V.CB-RN system for size attributes, as adopted in our previous paper [1], as well as the SV.CB-RN system for processing the names of the world's most beautiful scenery (spectacular view names) and the CN.CB-RN system for processing constellation names. As before, the data presented to each CB-RN system are represented as image patterns using QR codes [2]. These five types of CB-RN systems will be combined and trained with QR code pattern images of the attribute elements of each system. After that, when a pattern image of an attribute element is presented to any of the CB-RN systems, a mechanism will be constructed in which a chain (associative) recall of pattern images of related attribute elements in the other trained systems will be generated.

Coordinating teams of aerial robots in cluttered three-dimensional (3D) environments requires a principled integration of discrete mission planning-deciding which robot serves which goals and in what order -- with continuous-time trajectory synthesis that enforces collision avoidance and dynamic feasibility. This paper introduces IMD-TAPP (Integrated Multi-Drone Task Allocation and Path Planning), an end-to-end framework that jointly addresses multi-goal allocation, tour sequencing, and safe trajectory generation for quadrotor teams operating in obstacle-rich spaces. IMD--TAPP first discretizes the workspace into a 3D navigation graph and computes obstacle-aware robot-to-goal and goal-to-goal travel costs via graph-search-based pathfinding. These costs are then embedded within an Injected Particle Swarm Optimization (IPSO) scheme, guided by multiple linear assignment, to efficiently explore coupled assignment/ordering alternatives and to minimize mission makespan. Finally, the resulting waypoint tours are transformed into time-parameterized minimum-snap trajectories through a generation-and-optimization routine equipped with iterative validation of obstacle clearance and inter-robot separation, triggering re-planning when safety margins are violated. Extensive MATLAB simulations across cluttered 3D scenarios demonstrate that IMD--TAPP consistently produces dynamically feasible, collision-free trajectories while achieving competitive completion times. In a representative case study with two drones serving multiple goals, the proposed approach attains a minimum mission time of 136~s while maintaining the required safety constraints throughout execution.

We prove that platform-deterministic inference is necessary and sufficient for trustworthy AI. We formalize this as the Determinism Thesis and introduce trust entropy to quantify the cost of non-determinism, proving that verification failure probability equals 1 - 2^{-H_T} exactly. We prove a Determinism-Verification Collapse: verification under determinism requires O(1) hash comparison; without it, the verifier faces an intractable membership problem. IEEE 754 floating-point arithmetic fundamentally violates the determinism requirement. We resolve this by constructing a pure integer inference engine that achieves bitwise identical output across ARM and x86. In 82 cross-architecture tests on models up to 6.7B parameters, we observe zero hash mismatches. Four geographically distributed nodes produce identical outputs, verified by 356 on-chain attestation transactions. Every major trust property of AI systems (fairness, robustness, privacy, safety, alignment) presupposes platform determinism. Our system, 99,000 lines of Rust deployed across three continents, establishes that AI trust is a question of arithmetic.

Magic and entanglement are two measures that are widely used to characterize quantum resources. We study the interplay between magic and entanglement in two-qubit systems, focusing on the two extremes: maximal magic and minimal magic for a given level of entanglement. We quantify magic by the Rényi entropy of order 2, $M_2$, and entanglement by the concurrence $Δ$. We find that the Pareto frontier of maximal magic $M_2^{(max)}(Δ)$ is composed of three separate segments, while the boundary of minimal magic $M_2^{(min)}(Δ)$ is a single continuous line. We derive simple analytical formulas for all these four cases, and explicitly parametrize all distinct quantum states of maximal or minimal magic at a given level of entanglement.

We present a Sovereign AI architecture for clinical triage in which all inference is performed on-device and inbound data is delivered via a physically unidirectional channel, implemented using receive-only broadcast infrastructure or certified hardware data diodes, with no return path to any external network. This design removes the network-mediated attack surface by construction, rather than attempting to secure it through software controls. The system performs conversational symptom intake, integrates device-captured vitals, and produces structured, triage-aligned clinical records at the point of care. We formalize the security properties of receiver-side unidirectionality and show that the architecture is transport-agnostic across broadcast and diode-enforced deployments. We further analyze threat models, enforcement mechanisms, and deployment configurations, demonstrating how physical one-way data flow enables high-assurance operation in both resource-constrained and high-risk environments. This work positions physically unidirectional channels as a foundational primitive for sovereign, on-device clinical intelligence at the front door of care.

This paper describes LogSigma, our system for SemEval-2026 Task 3: Dimensional Aspect-Based Sentiment Analysis (DimABSA). Unlike traditional Aspect-Based Sentiment Analysis (ABSA), which predicts discrete sentiment labels, DimABSA requires predicting continuous Valence and Arousal (VA) scores on a 1-9 scale. A central challenge is that Valence and Arousal differ in prediction difficulty across languages and domains. We address this using learned homoscedastic uncertainty, where the model learns task-specific log-variance parameters to automatically balance each regression objective during training. Combined with language-specific encoders and multi-seed ensembling, LogSigma achieves 1st place on five datasets across both tracks. The learned variance weights vary substantially across languages due to differing Valence-Arousal difficulty profiles-from 0.66x for German to 2.18x for English-demonstrating that optimal task balancing is language-dependent and cannot be determined a priori.

Spiking Neural Networks (SNNs) offer inherent advantages for low-power inference through sparse, event-driven computation. However, the theoretical energy benefits of SNNs are often decoupled from real hardware performance due to the opaque relationship between training-time choices and inference-time sparsity. While prior work has focused on weight pruning and compression, the role of training hyperparameters -- specifically surrogate gradient functions and neuron model configurations -- in shaping hardware-level activation sparsity remains underexplored. This paper presents a workload characterization study quantifying the sensitivity of hardware latency to SNN hyperparameters. We decouple the impact of surrogate gradient functions (e.g., Fast Sigmoid, Spike Rate Escape) and neuron models (LIF, Lapicque) on classification accuracy and inference efficiency across three event-based vision datasets: DVS128-Gesture, N-MNIST, and DVS-CIFAR10. Our analysis reveals that standard accuracy metrics are poor predictors of hardware efficiency. While Fast Sigmoid achieves the highest accuracy on DVS-CIFAR10, Spike Rate Escape reduces inference latency by up to 12.2% on DVS128-Gesture with minimal accuracy trade-offs. We also demonstrate that neuron model selection is as critical as parameter tuning; transitioning from LIF to Lapicque neurons yields up to 28% latency reduction. We validate on a custom cycle-accurate FPGA-based SNN instrumentation platform, showing that sparsity-aware hyperparameter selection can improve accuracy by 9.1% and latency by over 2x compared to baselines. These findings establish a methodology for predicting hardware behavior from training parameters. The RTL and reproducibility artifacts are at https://zenodo.org/records/18893738.

We investigate the anatomy and complexity of quantum states in Krylov space, in the ergodic and many-body localised (MBL) phases of a disordered, interacting spin chain. The Krylov basis generated by the Hamiltonian from an initial state provides a representation in which the spread of the time-evolving state constitutes a basis-optimised measure of complexity. We show that the long-time Krylov spread complexity sharply distinguishes the two phases. In the ergodic phase, the infinite-time complexity scales linearly with the Fock-space dimension, indicating that the state spreads over a finite fraction of the Krylov chain. By contrast, it grows sublinearly in the MBL phase, implying that the long-time state occupies only a vanishing fraction of the chain. Further, the profile of the infinite-time state along the Krylov chain exhibits a stretched-exponential decay in the MBL phase. This behaviour reflects a broad distribution of decay lengthscales, associated with different eigenstates contributing to the long-time state. Consistently, a large-deviation analysis of the statistics of eigenstate spread complexities shows that while the ergodic phase receives contributions from almost all eigenstates, the complexity in the MBL phase is dominated by a vanishing fraction of eigenstates, which have anomalously large complexity relative to the typical ones.

We study the anomalous thermoelectric Hall response of two-dimensional massive Dirac fermions to first order in the electron-electron interaction. We compute both the Nernst response to a Luttinger-type gravitational potential and the particle magnetization, the latter being required to remove spurious non-transport contributions. We show that, for arbitrary interactions, the magnetization is described by a remarkably simple formula. Surprisingly, and contrary to expectations, subtracting the magnetization currents does not make the thermoelectric Hall coefficient vanish in the zero-temperature limit. We attribute this to violation of locality on the smallest length scales, which is inevitable in a quantized field theory, that happens to manifest itself in infrared physics.

Anderson localization (AL) and the non-Hermitian skin effect (NHSE) represent two paradigmatic localization phenomena driven, respectively, by disorder and non-Hermiticity. In one-dimensional (1D) non-Hermitian systems, these factors are known to compete and provide a smooth crossover between AL and NHSE upon parameter tuning. Here, we show that this interplay is fundamentally enriched in spinful systems, where an external magnetic field acts as an additional degree to manipulate the localization behavior. By investigating a disordered 1D spinful non-Hermitian chain, we demonstrate that under appropriately correlated disorder configurations across spin sectors, the magnetic field enhances the AL $\rightarrow$ NHSE crossover. Interestingly, this facilitates the Anderson delocalization transition even in strongly disordered systems where states would otherwise be Anderson localized. By analyzing the inverse participation ratio and the mean center of mass, we map the resulting triple interplay between disorder, non-Hermiticity, and the magnetic field strength, identifying regimes of Anderson localization and skin accumulation. We further reveal that this magnetic field driven delocalization phenomenon originates from an effective suppression of disorder strength via Zeeman-induced inter-chain coupling across the spin sectors.

The surface code is a promising platform for a quantum memory, but its threshold under coherent errors remains incompletely understood. We study maximum-likelihood decoding of the square-lattice surface code in the presence of single-qubit unitary rotations that create electric anyon excitations. We microscopically derive a non-linear sigma model with target space $\mathrm{SO}(2n)/\mathrm{U}(n)$ as the effective long-distance theory of this decoding problem, with distinct replica limits: $n\to1$ for optimal decoding, which assumes knowledge of the coherent rotation angle, and $n\to0$ for suboptimal decoding with imperfect angle information. This exposes a sharp distinction between the two decoders. The suboptimal decoder supports a ``thermal-metal'' phase, a non-decodable regime that is qualitatively distinct from the conventional non-decodable phase of the surface code under incoherent Pauli errors. By contrast, the metal phase cannot arise in optimal decoding, since the metallic fixed-point becomes unstable in the $n\to 1$ replica limit. We argue that optimal decoding may be possible up to the maximally-coherent rotation angle. Within the sigma model description, we show that the decoding fidelity is related to twist defects of the order-parameter field, yielding quantitative predictions for its system-size dependence near the metallic fixed point for both decoders. We examine our analytic predictions for the decoding fidelity as well as other physical observables with extensive numerical simulations. We discuss how the symmetries and the target space for the sigma model rely on the lattice of the surface code, and how a stable thermal metal phase can arise in optimal decoding when the syndromes reside on a non-bipartite lattice.

The dynamics of interacting particles in orbital magnetic fields are notoriously difficult to study, as this physics is inherently connected to electronic correlations in two-dimensional systems, for which no straightforward theoretical methods are available. Here, we report on the diffusive relaxation dynamics of two-dimensional interacting fermionic systems under a uniform magnetic field in the infinite temperature regime. We first show that the fermionic truncated Wigner approximation captures the equilibration dynamics unexpectedly well for intermediate interaction strengths when going beyond one dimension. This high accuracy holds at least for relatively small ladder systems, which are accessible to the Lanczos method that we use to benchmark the reliability of the Wigner approximation. We find that strong interactions, which exceed the hopping energy, suppress magnetic-field effects on diffusive transport. However, when the interactions are comparable to the kinetic energy, the diffusion is significantly reduced by the magnetic flux. This is observed for sufficiently large systems (above approximately 400 lattice sites), where finite-size effects weakly affect particle transport. We suggest that our results should be directly accessible on current optical lattice platforms.

A central challenge in nanoscale superconductivity is to understand and exploit the combined action of quantum confinement and proximity effects in experimentally realistic metallic heterostructures. We theoretically investigate superconducting bilayer heterostructures in which these two effects coexist. Using a generalized Eliashberg framework that incorporates both quantum confinement and proximity coupling, we show that their interplay can substantially enhance the superconducting critical temperature. In particular, the theory predicts superconductivity in selected bilayers whose constituent materials are nonsuperconducting or only weakly superconducting in the bulk. These results identify quantum-confined bilayers as a promising route to engineering emergent superconductivity in metallic heterostructures.

We report magnetic bistability in single Dy atoms on NaCl(100) thin films. Individual Dy atoms substituting Na at the surface of the NaCl layer are thermally stable up to at least 300 K, display $4f^{9}$ occupancy, out-of-plane easy magnetization axis, and long spin relaxation time $T_1$ of about 10 s at 2.5 K; thereby they are the first single atom magnet on a thermally stable adsorption site. Dy atoms adsorbed onto the Cl and bridge sites display $4f^{10}$ occupancy. Dy on top-Cl exhibit magnetic hysteresis and a $T_1$ of 550 s at 0.3 T and 2.5 K. The observed slow magnetic relaxation of Dy on both adsorption sites introduces NaCl as an effective platform for single atom magnets.

We develop a systematic framework for determining the nature of exceptional points of $n^{\rm th}$ order (EP$_n$s) in non-Hermitian (NH) systems, represented by complex square matrices. By expressing symmetry-preserving perturbations in the Jordan-normal basis of the defective matrix at an EP$_n$, we show that the upper-$k$ Hessenberg structure of the perturbation directly dictates the leading-order eigenvalue- and eigenvector-splitting to be $\propto ε^{1/k}$, when expanded in a Puiseux series. Applying this to three-band NH models invariant under parity (P), charge-conjugation (C), or parity-time-reversal (PT), we find that EP$_3$s in P- and C-symmetric systems are restricted to at most $\sim ε^{1/2}$ branch points, while PT-symmetric systems generically support EP$_3$s with the strongest possible singularities (viz. $\sim ε^{1/3}$). We illustrate these results with concrete three-dimensional models in which exceptional curves and surfaces emerge. We further show that fine-tuned perturbations can suppress the leading-order branch point to a less-singular splitting, which have implications for designing direction-dependent EP-based sensors. The appendix extends the analysis to four-band C- and P-symmetric models, establishing the existence of EP$_4$s with $\sim ε^{1/4}$ singularities.

Additives of sparingly soluble components are known to slow down or completely inhibit Ostwald ripening in dispersed systems. In this paper series, our earlier model of the stabilization against Ostwald ripening is revisited and extended over the whole range of compositions, molar volumes of components, and their activity coefficients. In the first paper, a simpler problem, the dissolution of a two-component drop under the action of excess Laplace pressure inside is analyzed. Three stages of dissolution are identified. In the first stage, called pre-lock-in, the concentration of the poorly soluble component undergoes a quick increase, and the system enters the lock-in state, in which the Laplace pressure effect on the chemical potential of the more soluble component is nearly completely counterbalanced by the Raoult effect. After this, the dissolution kinetics slows down and enters a steady state. In the process, the concentration of the sparingly soluble component continues to increase, first slowly and then more rapidly in the very end of the particle lifetime; this latter stage is called the 'late lock-in'. Despite all those variations, if the initial concentration of the poorly soluble component is above a certain threshold, the dissolution kinetics nearly follows the classical cubic law. An improved equation for the rate of dissolution is proposed that covers the whole formulation range and represents an extension over our previous formula.

Traditional plastics demand a choice between durability (thermosets) and reprocessability (thermoplastics). Vitrimers are a recent class of polymer network combining both these qualities. Their increased cost of production can be offset by mixing them with a traditional thermoplastic; however, phase separation in such blends can lead to inhomogenous materials. In this paper, we adapt an existing model for the free energy of dissociative polymer networks to their associative, vitrimeric counterpart. We test the accuracy of the model's predictions by comparing them with the results of novel molecular-dynamics simulations. We demonstrate that such melts can undergo phase separation even in the absence of energetic interactions between the components. We find furthermore that the phase diagram of the melts is qualitatively similar to that of dissociative systems, and that the critical degree of conversion for the onset of phase separation depends reciprocally on the number of function sites per vitrimer chain.

The rheological behavior of ring-linear polymer blends under uniaxial elongational flow has remained a subject of intense debate, particularly regarding the emergence of stress overshoot. Herein, we employ coarse-grained molecular dynamics simulations to investigate the chain-length dependence of elongational viscosity in 1:1 ring-linear blends of flexible chains with the equal molecular weight. Our results reveal a distinct threshold in the degree of threading, quantified by the number of entanglements Z = N /Ne (where N is the number of beads per chain and Ne is the entanglement chain length), for the appearance of stress overshoot: while blends with shorter chains (Z $\le$ 2) exhibit monotonic stress growth, a clear stress overshoot emerges when the chain length reaches a threshold value (Z $\approx$ 4). Consistent with previous reports, this overshoot originates from a thread-to-unthread transition. At the threshold chain length, multiple linear chains penetrate a single ring, providing sufficient topological constraints to significantly stretch the ring under elongational flow. We predict that this transition can be experimentally validated via 2D small-angle neutron scattering patterns in the plane of the stretching and perpendicular directions, offering a direct structural signature of the ring recoil process for future experimental verification.

We present an in-depth study of end states in honeycomb nanoribbons, focusing on the interplay between nanoribbon termination, chiral symmetry, and complex next-nearest-neighbor hopping in the framework of the Haldane model. Although previous work has identified zero-dimensional end states in such systems, this analysis is incomplete. Here, we systematically investigate zigzag and armchair nanoribbons of various widths, using the multiband Zak phase to characterize the topological properties of the occupied bands. We show that the Zak phase is quantized only for certain ribbon terminations, and we elucidate how this termination dependence governs the existence and robustness of end states. Furthermore, we explore the effect of varying the complex next-nearest-neighbor hopping phase, demonstrating the breakdown of chiral symmetry, the evolution of the bulk gap, and the resulting depinning of end-state energies. Finally, we place our findings in the context of previous studies and discuss connections to the Kane-Mele model, including the role of Rashba spin-orbit coupling. Our work provides a more detailed analysis of topological end states in nanoribbons described by the Haldane and Kane-Mele models and offers a framework for their characterization in related systems.

The tensor-network renormalization group (TNRG) is an accurate numerical real-space renormalization group method for studying phase transitions in both quantum and classical systems. Continuous phase transitions, as an important class of phase transitions, are usually accompanied by spontaneous breaking of various symmetries. However, the understanding of symmetries in the TNRG is well-established mainly for global on-site symmetries like U(1) and SU(2). In this paper, we demonstrate how to incorporate lattice symmetries (including reflection and rotation) and the PT symmetry in the TNRG in two dimensions (2D) through a case study of the hard-square lattice gas with nearest-neighbor exclusion. This model is chosen because it is well-understood and has two continuous phase transitions whose spontaneously-broken symmetries are lattice and PT symmetries. Specifically, we write down proper definitions of these symmetries in a coarse-grained tensor network and propose a TNRG scheme that incorporates these symmetries. We demonstrate the validity of the proposed method by estimating the critical parameters and the scaling dimensions of the two phase transitions of the model. The technical development in this paper has made the 2D TNRG a more well-rounded numerical method.

Biomolecular condensates organize biochemical processes by spatially concentrating molecules while allowing for dynamic exchange with their surroundings. However, transport across their interface can be strongly attenuated, leading to enhanced retention and preferential internal mixing. Two key mechanisms have been proposed to describe this behavior: biased interfacial reflectivity, which compares how strongly particles are reflected at the interface when attempting to enter or leave the condensate, and interfacial resistance, which sets the kinetic rate at which particles can cross the interface. Quantifying these parameters experimentally has remained challenging. Here, we present a theoretical and experimental framework to address this issue, extending our previously developed half-FRAP approach. We solve the spherical diffusion problem with a semipermeable interface by spectral decomposition. By evaluating the information content of the integrated recovery curves, we show that they encode sufficient information to recover interfacial parameters over extended regions of parameter space. Applying our framework to tunable coacervates composed of poly-lysine and hyaluronic acid, we find that their interfaces exhibit strongly biased reflectivity and substantial resistance, both driving preferential internal mixing. These parameters depend on salt concentration, linking interfacial transport to intermolecular interaction strength and position in the phase diagram. Our results establish a quantitative connection between interfacial properties and condensate dynamics, revealing how their interplay gives rise to distinct transport regimes.

Phase transitions between crystalline solids occur either through the nucleation and growth mechanism, a process that is slow and destructive or through the diffusion-less and order preserving Martensitic route. In both organic and inorganic materials, Martensitic transformations are known to occur only between phases with crystalline symmetry. We demonstrate here that for canonical discotic organic semiconductor HAT6, the transition between the liquid crystalline columnar hexagonal phase (ColH) and the crystalline solid can occur through a mechanism that exhibits the hallmarks of Martensitic transformations: orientational correlations between parent and daughter phases, structural reversibility, and ultrafast kinetics. To access Martensitic-like solidification, the ColH phase of HAT6 is biaxially aligned in lithographically defined microchannels and crystallization is induced on deep supercooling. The transition mechanism is studied using a combination of polarized optical microscopy and X-ray scattering. At the largest accessible supercooling, the ColH - Crystal phase transition occurs at speeds of ~100 micrometer/s, a value that is seven orders of magnitude greater than the theoretical prediction for growth from isotropic melts. Our work suggests that Martensitic-like transformations can occur even between liquid crystals and crystals and are therefore more general than previously believed. Further, our work demonstrates that Martensitic-like transformations of anchored liquid crystals can be used to grow biaxially aligned crystals of organic molecules over arbitrarily long distances. As lattice alignment over large areas is desirable for devices like field-effect transistors and as several high-performance molecular semiconductors exhibit a ColH phase, our results hold general significance for organic electronics.

In semiconductor nanostructures, optical excitation typically creates bound electron-hole states, such as excitons, trions, and larger complexes. Their relative motion is described by the Wannier equation, which is valid only for spatially extended motion in the Coulomb-dominated, weak-confinement limit. Other small nanostructures, such as quantum dots, are in the confinement-dominated strong confinement regime, where the wavefunction factorizes into independent electron and hole parts. Nanoplatelets are in between the two regimes and require solving an unfactorized higher-dimensional Schrödinger equation, which is computationally expensive. This work demonstrates how tensor networks can partially overcome this problem, using CdSe nanoplatelets as an example. The method is also applicable to related two-dimensional systems. As a demonstration, we calculate the excitonic and trionic ground states, as well as several excited states, for nanoplatelets of varying sizes, including their energies and oscillator strengths. More importantly, overall strategies for using tensor networks in real space for systems under intermediate confinement have been developed.

Two-dimensional metallic altermagnets are rare, and no correlated 2D material has been established to host large nonrelativistic spin splitting. Here we show that spontaneous orbital order, driven by electronic correlations and Fermi surface nesting, provides a general microscopic route to two-dimensional metallic altermagnetism. Antiferro-orbital ordering between the d$_{xz}$ and d$_{yz}$ orbitals breaks the equivalence of magnetic sublattices with opposite spins and generates a symmetry-enforced altermagnetic spin texture. As a concrete realization, we identify monolayer YbMn$_2$Ge$_2$ as a stable correlated metallic altermagnet exhibiting giant nonrelativistic spin splitting of order 1 eV. The resulting phase supports an exceptionally large and gate-tunable transverse spin conductivity. These results establish correlation-driven orbital order as a robust and general mechanism for designing correlated altermagnets with large spin splitting.

We study quantum and stochastic deformations of the rule-54 reversible cellular automaton (RCA54) on a 1+1-dimensional spatiotemporal lattice, focusing on their integrability structures in two distinct settings. First, for the quantum deformation, which turns the model into an interaction-round-a-face brickwork quantum circuit (either on an infinite lattice or with periodic boundary conditions), we show that the shortest-range nontrivial conserved charge commuting with the discrete-time evolution operator has a density supported on six consecutive sites. By constructing the corresponding range-6 Lax operator, we prove that this charge belongs to an infinite tower of mutually commuting conserved charges generated by higher-order logarithmic derivatives of the transfer matrix. With the aid of an intertwining operator, we further prove that the transfer matrix commutes with the discrete-time evolution operator. Second, for the stochastic deformation, which renders the model into a Markov-chain circuit, we investigate open boundary conditions that couple the system at its edges to stochastic reservoirs. In this setting, we explicitly construct the non-equilibrium steady state (NESS) by means of a staggered patch matrix ansatz, a hybrid construction combining the previously used commutative patch-state ansatz for the undeformed RCA54 with the matrix-product ansatz. Finally, we propose a simple empirical criterion for detecting integrability or exact solvability in a given model setup, introducing the notion of digit complexity.

Driven Floquet systems can realize topological phases with no static counterparts. These so-called anomalous Floquet topology breaks the bulk-boundary correspondence based on the Chern number. The number of edge modes in each band gap is instead determined by another integer index, a winding number, which is calculated from the time evolution operator of the bulk states within one driving period. While in the non-driven system, Chern markers provide a useful local proxy for the Chern number in the bulk, so far no such local bulk indicator is known for the winding number in Floquet systems. Here we consider two-band models and show that the area enclosed during a Floquet period by an initially localized particle signals the presence of an anomalous phase when it approaches half the unit cell area. In general, we show that at the fine-tuned point of dispersionless dynamics during the micromotion, the enclosed area is quantized and an exact proportionality relation exists between the area and the winding number. Direct detection of anomalous topology in real space could be realized in several quantum simulation platforms, and could be useful for systems with disorder or interactions. Building on the connection between area and winding number, we also show a way to realize arbitrarily high winding numbers.

Stochastic resetting has attracted significant attention in recent years due to its wide-ranging applications across physics, biology, and search processes. In most existing studies, however, resetting events are governed by an external timer and remain decoupled from the system's intrinsic dynamics. In a recent Letter by Biswas et al, we introduced threshold resetting (TR) as an alternative, event-driven optimization strategy for target search problems. Under TR, the entire process is reset whenever any searcher reaches a prescribed threshold, thereby coupling the resetting mechanism directly to the internal dynamics. In this work, we study TR-enabled search by $N$ non-interacting diffusive searchers in a one-dimensional box $[0,L]$, with the target at the origin and the threshold at $L$. By optimally tuning the scaled threshold distance $u = x_0/L$, the mean first-passage time can be significantly reduced for $N \geq 2$. We identify a critical population size $N_c(u)$ below which TR outperforms reset-free dynamics. Furthermore, for fixed $u$, the mean first-passage time depends non-monotonically on $N$, attaining a minimum at $N_{\mathrm{opt}}(u)$. We also quantify the achievable speed-up and analyze the operational cost of TR, revealing a nontrivial optimization landscape. These findings highlight threshold resetting as an efficient and realistic optimization mechanism for complex stochastic search processes.

The introduction of Su-Schrieffer-Heeger model has led to a major breakthrough in the area of one-dimensional topological insulators, even though this model was primarily formulated on an organic polymer called $trans$-polyacetylene in order to explain its anomalous conductivity. In this study, a group of five tight-binding models has been introduced which are formulated on another organic polymer called polyacene, where exotic topological behavior has been observed. Topological properties of the most common geometric isomers known as $cis$-polyacene, and $trans$-polyacene have been investigated along with three additional modified polyacene structures. Although their geometric structures differ by mirror symmetry, tight-binding band structures of $cis$-polyacene and $trans$-polyacene are found the same, where again their topological characters are found totally opposite. The $trans$-polyacene is nontrivial as it exhibits topological phase with nonzero winding number, while the $cis$-polyacene is topologically trivial, although both the structures adhere to the same set of symmetries required for the topological character. However, $cis$-polyacene possesses additional mirror symmetry in the real space. Three modified structures of polyacene have been considered in order to induce the nontrivial topology, where exotic topological behavior is noted in two of them.

Athermal elongated particles are well-known to follow Jeffery orbits when sheared in viscous fluids. It is less clear if similar orbits appear in dense granular flows. We show that when sheared for long enough, sufficiently elongated frictionless granular rods, rather than following noisy Jeffery-like orbits, exist in a quasi-equilibrium state, whose orientational statistics are quantitatively described by classical liquid crystal theory, where the noise is provided by collisions due to shear. At the same time, we demonstrate a systematic breakdown of this equilibrium analogy at two distinct limits: at low aspect ratios, where the equilibrium theory incorrectly predicts an isotropic state, and as inter-particle friction is introduced, where the system moves from steric screening to frictional gearing. Even within this frictionally geared state, the rotational dynamics remain distinct from classical Jeffery orbits. We link this frictional breakdown directly to the system being driven far from equilibrium, as quantified by an effective Ericksen number that compares non-equilibrium rotational driving to steric ordering. Our results provide a quantitative map of the transition from a quasi-equilibrium to a far-from-equilibrium steady state in a dense, driven system, defining the limits of applicability for thermal liquid crystal theory in athermal matter.

In complex molecular systems, the reaction coordinate (RC) that characterizes transition pathways is essential to understand underlying molecular mechanisms. This review surveys a framework for identifying the RC by applying deep learning to the committor, which provides the most reliable measure of the progress along a transition path. The inputs to the neural network are collective variables (CVs) expressed as functions of atomic coordinates of the system, and the corresponding RC is predicted as the output by training the network on the committor as the learning target. Because deep learning models typically operate in a black-box manner, it is difficult to determine which input variables govern the predictions. The incorporation of eXplainable Artificial Intelligence (XAI) techniques enables quantitative assessment of the contributions of individual input variables to the predictions. This approach allows the identification of CVs that play dominant roles and demonstrates that the committor distribution on the surface using important CVs is separated by well-defined boundaries. The framework provides an explainable deep learning strategy for assigning a molecular mechanism from the RC and is applicable to a wide range of complex molecular systems.

We perform molecular dynamics simulations to investigate hydration structures and dynamics in seven water-containing polymers: PVA, PHEA, PHEMA, PBA, PMEMA, PEG, and PMEA. The analysis integrates four perspectives: the water-content dependence of the glass transition temperature $T_g$, polymer chain fluctuations characterized by dihedral angle distributions, hydrogen-bond lifetimes $τ_{\mathrm{HB}}$ between water and polymer functional groups, and the localization and exchange dynamics of confined water quantified by the distinct part of van Hove correlation function. Hydroxyl-containing polymers (PVA, PHEA, and PHEMA) exhibit relatively high dry-state $T_g$ values and its pronounced depression upon hydration. Chain fluctuations are limited, and $τ_{\mathrm{HB}}$ follows Arrhenius behavior, forming localized hydration shells. In contrast, PMEMA and PBA show low equilibrium water contents and hydrophobic character; although their dry-state $T_g$ values are moderately lower and less sensitive to water content, chain fluctuations remain small, and $τ_{\mathrm{HB}}$ also obeys Arrhenius behavior, with hydrophobic aggregation promoting water localization. PEG and PMEA display low dry-state $T_g$ values and weak water-content dependence. Greater rotational freedom around ether or methoxy oxygen atoms leads to larger chain fluctuations and loosely bound water. Below $T_g$, $τ_{\mathrm{HB}}$ between water and ether or methoxy oxygen atoms exhibits super-Arrhenius behavior. These results clarify three hydration types: highly hydrated (PVA, PHEA, and PHEMA), hydrophobic (PMEMA and PBA), and flexibly hydrated (PEG and PMEA), and provide a molecular-level framework for interpreting interfacial water governed by water-polymer interactions.

Plasticity is a fundamental property of complex systems, such as the brain or an organism. Yet it typically remains a descriptive concept inferred retrospectively from observed outcomes, such as modifications in activity or morphology. Here, the network-based operationalization of plasticity is further formalized as the ratio between system size and connectivity strength among system elements. Within this framework, system size determines the dimensionality of the accessible state space, while connectivity strength tunes the system's regime. An optimal range of plasticity -- balancing capacity for change and capacity to maintain coherence -- emerges at intermediate connectivity strength. Notably, this balance coincides with the critical regime, which provides a theoretically motivated benchmark that enables a normalized unit of measure, termed effective plasticity, and comparisons of adaptive efficacy across diverse systems. Plasticity is thus transformed into a predictive tool that quantifies a system's capacity for change before it occurs. Its validity is supported across disciplines and, in particular, by evidence from psychopathology where it anticipates transitions between mental states. At a mechanistic level, plasticity acts as a structural tuning parameter for criticality, reframing their relationship as causal, with plasticity driving criticality rather than merely accompanying it. Furthermore, this network-based operationalization explains how larger systems can more robustly maintain critical dynamics. Crucially, the proposed perspective distinguishes functional regime shifts from thermodynamic phase changes, identifying plasticity as the system-level regulator that shapes and constrains the dynamic repertoire. This framework is applicable across domains, including ecology, economics, and social systems, and may foster cross-disciplinary integration within complexity science.

Minimal Kitaev chains provide a unique platform to engineer Majorana states in quantum dots interacting via normal tunneling and crossed Andreev reflection specified by their amplitudes $|η_{n,a}|$. Here we analyze fluctuations of electric currents in a double quantum dot Kitaev chain using the differential effective charge $q$, that is the ratio of the differential shot noise and conductance. At low bias voltages $V$ we find that $q=e/2$ in a very narrow vicinity of the point $|η_n|=|η_a|$ whereas $q=3e/2$ almost in the whole sweet spot region and marks the range where the poor man's Majorana states largely govern the fluctuations. At high $V$ we show that the sweet spot region is still characterized by $q=3e/2$ uniquely identifying the poor man's Majorana states using the high voltage tails. For $|η_n|=0$ or $|η_a|=0$ we obtain $q=e$ at any $V$. Remarkably, before the asymptotic value $q=e$ is reached for very high $V$, the maximal value $q=2e$ is formed at $|eV|=2\sqrt{|η_n|^2+|η_a|^2}$. The unique nature and potentially rich fluctuation behavior revealed in this work provide a stimulating ground for the next generation experiments on nonequilibrium shot noise in minimal Kitaev chains.

We present a measurement-based quantum thermal machine that extracts work from the back-action of generalized quantum measurements whose working medium is a coupled two-level quantum system. Specifically, we derive universal optimization criteria for a three-stroke measurement-based engine cycle with coupled two-level system of Ising-like interaction as a working medium. Furthermore, we present two numerical algorithms to optimize the engine work extraction and enhance its performance. Our numerical results demonstrate: (i) efficiency peaks in the projective-measurement limit; (ii) symmetry breaking (detuning or weak coupling) enlarges the exploitable energy gap; and (iii) performance remains robust ($>50\%$ of optimum) under $\sim\!10^\circ$ feedback-pulse errors. The framework is platform-agnostic and directly implementable with current superconducting, trapped-ion, or NMR technologies, providing a concrete route to scalable, measurement-powered quantum thermal machines.

Capturing the dynamics of quantum many-body systems under time-dependent driving protocols is a central challenge for numerical simulations. Existing methods such as tensor networks and time-dependent neural quantum states, however, must be re-run for every protocol. In this work, we introduce the Neural Operator Quantum State (NOQS) as a foundation model for quantum dynamics. Rather than solving the Schrödinger equation for individual trajectories, our approach aims to \emph{learn the solution operator} that maps entire driving protocols to time-evolved quantum states. Once trained, the NOQS predicts time evolution under unseen protocols in a single forward pass, requiring no additional optimization. We validate NOQS on the two-dimensional Ising model with time-dependent longitudinal and transverse fields, demonstrating accurate prediction not only for unseen in-distribution protocols, but also for qualitatively different, out-of-distribution functional forms of driving. Further, a single NOQS model can be transferred between different temporal resolutions, and can be efficiently fine-tuned with sparse experimental measurements to improve predictions across all observables at negligible cost. Our work introduces a new paradigm for quantum dynamics simulation and provides a practical computational-experimental interface for driven quantum systems.

Photonic bound states in the continuum (BICs) provide a revolutionary paradigm for boosting light-matter interactions in integrated nanocavity systems. Nevertheless, precise manipulation of open cavity-emitter architectures still faces critical challenges, especially in realizing deterministic directional radiation and suppressing the perturbation of intrinsic cavity modes induced by emitters as local impurities. Conventional investigations on cavity-emitter coupling are predominantly based on ensemble measurements, which inevitably mask the intrinsic physics underlying individual light-matter interactions. Here, we propose a robust strategy to control the upconversion and emission of a single-particle emitter using a topological plasmonic cavity with broken σh mirror symmetry. This structured design enables the transition from symmetry-protected BICs to a multi-BIC regime with finite but ultrahigh confinement, where nontrivial phase evolution and hybridization of transverse electric and magnetic modes open a well-defined far-field radiation channel for directional emission. Leveraging this scheme, we experimentally demonstrate dramatically enhanced radiation intensity from a single point-like emitter, together with uniform and deterministic directional emission, while achieving excellent structural robustness against local perturbations. This work establishes a general framework for engineering coherent directional light emission at the nanoscale, which lays a solid foundation for high-performance chip-scale integrated nanophotonic applications.

We investigate the interplay of in-plane magnetic and transverse electric fields in AB-stacked bilayer graphene. In prior work, we demonstrated that this configuration induces an insulator-metal (IM) transition with large impact on the magnetic response, albeit requiring impractically large magnetic fields. Here, we extend the analysis by incorporating previously neglected trigonal warping effects through interlayer skew couplings. In a restricted region of momentum space (on the order of 1/100 of the original Brillouin zone) trigonal warping produces a fine splitting of Dirac cones leading to a compensated semimetallic state in the absence of external fields. Application of a transverse electric field above a small threshold ($V_c\sim 0.6$ meV) reinstates the insulating gap, but this gap can be closed by a relatively small in-plane magnetic field, leading to an IM transition at a much smaller magnetic field ($\approx 10$ T) than previously predicted.

Design of soft matter capable of controllable microscale dynamics is a frontier of modern science. Jiahao Chen et al. demonstrate that an electric field can drive particle-like solitons-skyrmions in a chiral nematic along a preprogrammed trajectory with a variable speed within a two-dimensional plane. The effect is rooted in flexoelectric polarization of a deformed director field.

The interaction between phase oscillators is conservative if the phase volume is conserved throughout the dynamics. We derive a general condition, based on the notion of a pair-Hamiltonian, for the pairwise couplings to be conservative. The conservative networks with Winfree-type and Kuramoto-Daido-type couplings are also discussed. It is demonstrated that although, in contradistinction to genuine Hamiltonian dynamics, there is no exact pairwise symmetry of the Lyapunov exponents, the Lyapunov spectrum for a large network is nearly symmetric. The concept is also generalized to triplet and quadruplet couplings.

In this article we construct evolutionary soliton hierarchies from pencils of Novikov algebras of Stäckel type. We start by defining a special class of associative Novikov algebras, which we call Novikov algebras of Stäckel type, as they are associated with classical Stäckel metrics in Viète coordinates. We obtain sufficient conditions for pencils of these algebras so that the corresponding Dubrovin-Novikov Hamiltonian operators can be centrally extended, producing sets of pairwise compatible Poisson operators. These operators lead to coupled Korteweg-de~Vries (cKdV) and coupled Harry Dym (cHD) hierarchies, as well as to a triangular cKdV hierarchy and a triangular cHD hierarchy.

We present an analytical and numerical study of the dynamics and stability of exciton-polariton condensates described by the open-dissipative Gross-Pitaevskii equation, incorporating both binary and short-range three-body interactions. Using an asymptotic description, we identify the parameter regime and derive equations for the instability amplitude, providing insights into vortex formation via the snake instability of dark solitons. We find that a repulsive three-body interaction, when combined with a binary interaction, supports stable vortex-antivortex pair formation. On the other hand, the reinforcement of attractive three-body interactions with binary interaction triggers the emergence of snake instability, leading to boundary-driven vortex disintegration. The time evolution of the instability under the influence of reservoir effects indicates that the boundary effects are more pronounced, to the extent of destabilizing the vortices with attractive three-body interactions compared to repulsive three-body interactions, thereby underscoring the stable nature of vortices in the repulsive domain.

Okounkov [36] proved a remarkable formula relating $n$-point GUE (Gaussian unitary ensemble) correlators of a fixed genus to Witten's intersection numbers of the same genus. The partition function of GUE correlators is a tau-function for the Toda lattice hierarchy. In this note, based on the knowledge of these two statements we give a new proof of the Witten--Kontsevich theorem, that relates Witten's intersection numbers to the KdV (Korteweg--de Vries) integrable hierarchy.

We demonstrate an all solid state semiconductor device, based on epitaxial single crystalline metal halide perovskites, enabling reversible control of a perovskite photoluminescence with a gate voltage. Fundamentally distinct from electroluminescent diodes, such a photoluminescence field effect transistor uses the gate electric field to electrostatically modulate the interfacial density of mobile charges, thereby affecting the radiative and nonradiative recombination channels of photocarriers. Varying the gate voltage in such transistors efficiently changes the rate of nonradiative interfacial recombination and modulates the photoluminescence intensity by 65 to 98 percent (depending on temperature). At favorable gating, nearly complete elimination of non-radiative losses can be achieved. This functionality, coupled with the strong visible-range absorption and emission, possible due to the high absorption coefficient, as well as controllable thickness and macroscopically homogeneous morphology of epitaxial perovskite films, leads to high external photoluminescence quantum efficiencies realized in large-area, thin-film devices. Such high-efficiency, scalable, electrostatically tunable optoelectronic switches broaden the potential applications of metal-halide perovskites in photonics and optoelectronics.

Quantum interference of indistinguishable photons is the foundation of photonic quantum technologies, yet scaling from a few to many identical quantum light sources remains a major challenge. In solid-state platforms, spatial and spectral inhomogeneity and resource-intensive architectures impede scaling. As a result, interference between remote, independent quantum emitters has been thus far limited to pairs. Here we introduce a wavefront-shaping approach that enables scalable interference from multiple indistinguishable quantum dots on the same chip. Using programmable spatial light modulators, we independently excite, collect, and route emission from spatially distinct, yet spectrally degenerate dots. Scaling from two to five indistinguishable emitters, we verify interference through cooperative-emission phenomena and Hong-Ou-Mandel two-photon interference, thereby establishing a route towards large-scale, programmable quantum photonic architectures.

The electrification of energy demand across sectors, powered by solar and wind generation, is the best strategy for achieving carbon neutrality. Carbon dioxide removal (CDR) strategies are also expected to play a crucial role by providing net-negative emissions that can offset residual CO2 emissions, including those from cement manufacturing. While previous studies have assessed the role of CDRs in Europe's decarbonisation, most either focus solely on combinations of biogenic point-source capture and direct air capture (DAC) coupled with underground sequestration, or consider multiple CDR strategies at low spatial and temporal resolution, thereby limiting the representation of linkages amongst technologies. In this study, the sector-coupled European energy system model PyPSA-Eur is extended to include afforestation, perennialisation, biochar, and enhanced rock weathering (ERW) as additional CDR strategies. Using this model with a 3-hourly resolution and a network comprising 90 nodes, results show that a climate-neutral energy system equipped with these CDR strategies is 9% less expensive. Afforestation, perennialisation, and ERW potentials are fully utilised across regions, whereas biochar is not selected due to limited solid biomass feedstock being allocated to other higher-value processes. Furthermore, when these CDR strategies are combined with underground sequestration and a continental CO2 transport network, DAC is no longer required to achieve climate neutrality in Europe.

CrCoNi medium-entropy alloys exhibit exceptional mechanical properties arising from pronounced chemical complexity, including short-range order (SRO), and low stacking fault energy, posing challenges for large-scale atomistic simulations. While most models focus on equimolar compositions, deviations from equimolarity provide an effective route to tuning properties, requiring transferable interatomic potentials that capture composition-dependent behavior. Here we develop a general-purpose machine-learned interatomic potential for the CrCoNi system within the neuroevolution potential (NEP) framework, achieving near first-principles accuracy with high computational efficiency. Trained on a comprehensive dataset spanning pure elements, binary and ternary alloys across a wide compositional range, diverse crystal structures and thermodynamic conditions, and based on spin-polarized \textit{ab initio} data, the model accurately reproduces equations of state, phonons, elastic constants, dislocation dissociation, surface and defect energies, melting temperatures and strain-induced phase transformations. It further captures SRO and its effect on stacking fault energies across both equimolar and non-equimolar compositions, in agreement with first-principles and experiments. In contrast to existing potentials, typically limited to equimolar alloys and less accurate for pure elements, the present model delivers consistent accuracy across the full compositional space while retaining superior efficiency. These results enable reliable atomistic simulations of composition-dependent behaviour and provide a framework for the design of non-equimolar CrCoNi alloys.

Encoding continuous-variable quantum information in the optical domain has recently enabled the generation of large entangled states, yet robust implementation remains a challenge. Here, we present a straightforward protocol for generating multipartite entanglement based on spontaneous parametric down-conversion in a circular array of quadratic nonlinear waveguides. We provide a rigorous theoretical framework, including comprehensive derivations of the propagation equations and the identification of regimes where analytical solutions are possible. Crucially, our approach identifies the pump and detection configurations required to sustain and measure multipartite full inseparability across arbitrary propagation distances and for any number of waveguides $N=4 n$. This regime, elusive to standard numerical methods, represents a key requirement for scalable quantum protocols. Our scheme is inherently robust as it relies on phase-matched propagation eigenmodes, making it resilient against variations in sample length, coupling, and nonlinearity.

The interaction of ultrafast lasers with plasmas has been studied for many years, primarily with respect to fundamental emission mechanisms. Only in recent years has ionizing radiation emerged as a safety concern in ultrafast laser-based material processing, where high pulse energies, repetition rates, and average powers, combined with continuous material supply, can lead to sustained X-ray emission. These processing-specific findings have informed German radiation protection legislation, which mandates notification or approval for laser systems exceeding irradiances of $1 \times 10^{13}~W/cm^2$. However, this threshold does not distinguish between material processing and other ultrafast laser applications. In this work, we show that the conditions required for X-ray generation are highly specific and are typically only met during material processing. We assess the applicability of existing radiation studies to non-processing environments and present experimental results demonstrating negligible or no dose production under representative laboratory conditions, such as ultrafast laser interactions with underdense gas or stationary solid targets. We conclude that current legislation generalizes a processing-specific hazard to all ultrafast laser applications and does not adequately reflect the relevant physical conditions.

Quantum repeater networks distribute entanglement over lossy links while many users share a limited pool of entangled pairs. Most existing routing schemes either always use a single best path or rely on global optimizations that are hard to run in real time. Here we propose and analyze a simple alternative: a stochastic multipath rule in which each entanglement request is sent at random along one of several edge-disjoint repeater paths, with a single parameter that controls the bias between shorter and longer routes. Using a distance-dependent lossy network model with finite per-link capacities and probabilistic entanglement swapping, we develop an analytic description of the resulting end-to-end entanglement rate as a function of this bias and validate it with large-scale numerical simulations. We find that an intermediate bias consistently outperforms both deterministic extremes across distances, traffic patterns, attenuation, swapping noise, and congestion, bringing the rate close to simple capacity upper bounds and making link usage more even across networks. These results identify stochastic multipath routing as a lightweight classical control strategy for boosting performance and scalability in near-term quantum repeater networks.

Florida has experienced significant population increase in recent years, driven in part by domestic migration from other states. This study analyzes the migration patterns of voters in Florida between 2017 and 2022 using voter registration data. By examining demographic characteristics such as race/ethnicity, gender, age, and party affiliation, I identify trends in voter migration and their implications for Florida's political landscape. The findings reveal that minorities, younger individuals, Republicans, and those possibly with non-conforming gender are more likely to migrate into Florida. These insights contribute to understanding the dynamics of Florida's migration patterns and the effect of migration on recent elections.

Current single molecule methods either rely on fluorescence or lack chemical information. Here we report stimulated Raman photothermal encoded scattering (SRPSCAT) microscopy for quantitative bond-selective imaging of single-biomolecule structures and interactions in native environments. In this approach, scattering of the target molecule is modulated by the deposited energy from stimulated Raman gain and loss processes, thereby encoding vibrational spectroscopic information. Leveraging single-molecule sensitivity of interferometric scattering, SRPSCAT can map single proteins with chemical specificity, determine their mass, and distinguish protein secondary structures based on their Raman fingerprints. Furthermore, single protein binding kinetics are quantified and the conformational dynamics of single de novo designed allosteric proteins are observed. Together, these results highlight the potential of SRPSCAT for label-free structural, functional and dynamic analysis at the single-molecule level.

Stimulated Brillouin-Mandelstam scattering offers exceptional capabilities for photonic signal processing, but current platforms demand performance trade-offs between long interaction lengths, high gain, low optical losses, and practical implementation. Here, we demonstrate a novel platform based on the reversible freezing of a carbon disulfide filled liquid-core optical fiber. This approach delivers a giant in-fiber Brillouin gain of 434 W-1m-1 with a linewidth of 24 MHz, while maintaining low propagation losses in a fully spliced architecture and providing the potential for meter-scale interaction lengths. Leveraging this gain, as a proof of principle, we realize an optoacoustic memory operating at sub-nanojoule pulse energies - more than two orders of magnitude lower than state-of-the-art implementations. This power reduction is universal for Brillouin-based fiber applications in general and will enable low-power photonic signal processing and neuromorphic computing, efficient microwave photonics and sensing, as well as in-fiber quantum optomechanics-based technologies.

Point-gap topology under periodic boundary conditions and line-gap topology under open boundary conditions are generally inequivalent in non-Hermitian systems. We show that, in chiral non-Hermitian lattices, these two boundary-sensitive topological transitions become locked when the parameter sweep is confined to an exceptional-point (EP)-constrained manifold, such that the Bloch spectrum remains pinned to a zero-energy degeneracy throughout the evolution. In an extended non-Hermitian Su-Schrieffer-Heeger chain, this locking can be established analytically in a tractable limit, where the EP-constrained manifolds and the corresponding PBC and OBC transition boundaries are obtained in closed form, and it persists away from this limit when the generalized Brillouin zone is determined numerically. Outside the EP-constrained manifold, the two transitions generally decouple, even in the presence of isolated EPs or Hermitian degeneracies. We further show that the same mechanism survives in a four-band spinful extension with branch-resolved generalized Brillouin zones, including branch-imbalanced regimes. These results identify EP-constrained band evolution as a simple organizing principle for boundary-sensitive topology in chiral non-Hermitian systems and suggest a useful route for diagnosing non-Bloch topological transitions from periodic-boundary spectral evolution when such spectral information can be accessed in photonic, circuit, and cold-atom platforms.

Accurately measuring the magnetocaloric effect is necessary to foster the development of magnetic refrigeration devices. However, current methods are inconvenient, requiring different instruments to measure each individual property or a custom-made setup. By measuring the time-varying magnetization in a commercially available VersaLab\textsuperscript{\textregistered} PPMS\textsuperscript{\textregistered} from Quantum Design, we have determined the adiabatic temperature change ($Δ$T$_{\textrm{ad}}$) of the first-order phase transition material Gd$_5$Si$_2$Ge$_2$, for a magnetic field change of 0 to 1 T, under high vacuum ($<$ 0.1 mTorr). For each temperature and magnetic field, the equilibrium magnetization is used as the magnetization-to-temperature conversion curve, allowing us to extend the validity of a previously proposed technique to the first-order phase transition material Gd$_5$Si$_2$Ge$_2$, which exhibits significant hysteresis. Our method thus enables full characterization (magnetic entropy change, adiabatic temperature change, and heat capacity) of any magnetocaloric material, whether it has a first-order or a second-order phase transition, using a single instrument. Comparing to a directly measured $Δ$T$_{\textrm{ad}}$, our method resulted in a peak $Δ$T$_{\textrm{ad}}$ value of 4.47 K, within 1\% of the directly measured value for a sample of the same composition.

Homogeneous Boltzmann-type equations are an established tool for modelling interacting multi-agent systems in sociophysics by means of the principles of statistical mechanics and kinetic theory. A customary implicit assumption is that interactions are "all-to-all", meaning that every pair of randomly sampled agents may potentially interact. However, this legacy of classical kinetic theory, developed for collisions among gas molecules, may not be equally applicable to social interactions, which are often influenced by preferential connections between agents. In this paper, we discuss ongoing research on incorporating graph structures into homogeneous Boltzmann-type equations, thereby accounting for the "some-to-some" nature of social interactions.

This paper proposes a physics-informed neural operator (PINO) framework for solving inverse scattering problems, enabling rapid and accurate reconstructions under diverse measurement conditions. In the proposed approach, the dielectric property is represented as a learnable tensor, while a neural operator is employed to predict the induced current distribution. A hybrid loss function, consisting of the state loss, data loss and total-variation (TV) regularization, is constructed to establish a fully differentiable formulation for a joint optimization of network parameters and dielectric property. To demonstrate the framework's generality and flexibility, PINO is implemented using three representative neural operators, i.e., the Fourier Neural Operator (FNO), the enhanced Fourier Neural Operator (U-FNO) and the Factorized Fourier Neural Operator (F-FNO). Compared with conventional approaches, the proposed framework offers a simpler formulation and universal modeling capability, making it readily applicable to various measurement scenarios, including multi-frequency and phaseless inversion. Numerical simulations demonstrate that the proposed PINO achieves high accuracy and robust reconstruction across samples with and without phase information, under single-frequency and multi-frequency settings in the presence of noise. The results demonstrate that PINO consistently outperforms conventional contrast-source inversion (CSI) methods and provides an efficient, unified solution to complex electromagnetic inverse-scattering problems.

Optical tweezers confine a particle in an intensity-defined potential well by engaging its local multipoles. In this picture, eliminating far-field scattering from the particle should cancel the optical force, as the multipole moments underpinning the conventional optomechanical response vanish. We show that certain resonant states, such as, e.g., the hybrid anapole state, enable qualitatively different optical manipulation, nonlocal by nature, where the optical force exhibits nontrivial spatial variations absent in conventional tweezing, establishing a new framework for manipulating resonant nanoparticles.

This paper examines the latent functional block structure of Japan's production network using interregional input-output data. To isolate non-trivial production linkages, we first estimate a structural gravity model to account for spatial frictions and economic scale, and then apply a weighted stochastic blockmodel (SBM) to the resulting residual network. Because these residual linkages often connect distant regions, the SBM is well suited to grouping region-industry pairs based on their shared macroeconomic roles. The results reveal that even after explicitly filtering out the mechanical effects of geographic proximity, the network is organized into functional blocks that maintain a high degree of regional coherence. Beyond this baseline spatial clustering, we find evidence of cross-regional integration, a structural bifurcation between manufacturing and urban services in metropolitan areas, and broadly spanning primary sectors. These findings provide a network-based perspective on regional coordination, offering guidance for how structurally defined production blocks-rather than simple geographic proximity-can inform wide-area policy design.

Thin films of functional inorganic materials, particularly oxides, play a vital role in optoelectronics, enabling applications that range from active optical components to MEMS-based architectures. Achieving high aspect ratio patterning of these functional materials remains a significant challenge, as many of their constituent elements do not readily form volatile compounds required for conventional reactive ion etch processes. We introduce a novel approach, Pulsed Laser Template ENgineering (PLATEN), which offers a more accessible route for patterning materials that are typically difficult to etch. This technique involves depositing functional films using the Pulsed Laser Deposition (PLD) process onto silicon substrates that have been pre-patterned using reactive ion etching to create high aspect ratio features. Due to the highly forward-directed nature of the PLD process, the deposited films replicate closely the topography of the patterned silicon, without coatings the sidewalls. This process remains effective even at feature sizes down to approximately 50 nm. The oxide films replicate the underlying silicon pattern to a thickness of 80 nm. For thickness beyond 80 nm the patterns develop a waist at the midpoint which scales with film thickness and is not dependent on the feature size. In this paper, we present a detailed analysis of the PLATEN process, including deviations from ideal pattern replication in sub-micron features as a function of film thickness, and demonstrate near single crystalline growth of oxides on the patterned silicon substrate, demonstrating the potential of PLATEN technique for active opto-electronic materials.

Diffractive neural networks (DNNs) are an emerging approach for the realization of photonic artificial intelligence, especially due to their suitability for machine-vision applications and high-dimensional photonic information processing at lower power consumption. However, incorporating optical nonlinear activation functions to make DNNs a feasible alternative to their electronic counterpart remains a challenge. Here, we investigate the inclusion of second-harmonic generation (SHG), as one of the simplest and most efficient types of optical nonlinearities, in DNNs. We numerically investigate the impact of SHG on the performance of classification tasks in an all-optical nonlinear DNNs. Specifically, we investigate and discuss the essential requirements for an effective arrangement of the SHG layer in single and multilayer DNNs. We find that the performance, in terms of classification accuracy and class contrast, is affected strongly by the positioning of the SHG layer. Finally, we discuss and outline the constraints for including SHG in an experimental realization. Taking these constraints into account, we estimate the power-related efficiency of the nonlinear DNN system. Overall, our results provide a path towards implementing nonlinear DNNs using the SHG process.

Spatiotemporal optical vortices (STOVs) carry transverse orbital angular momentum and offer new degrees of freedom for light-matter interactions. Yet conventional focusing of STOVs introduces spatiotemporal astigmatism: the beam diffracts while the pulse duration stays constant, causing the vortex to deform away from focus. Here we overcome this limitation by introducing spectral phase modulation into a temporal focusing configuration, where angular dispersion forces the pulse to compress only at the geometric focus so that the spatial and temporal dimensions focus and defocus together. Our approach generates stable STOVs with self-similar, distortion-free evolution over an extended focal region. Besides, the orbital angular momentum vector can be continuously steered from purely longitudinal to strongly tilted orientations by adjusting the spatial dispersion, objective focal length, or input beam size. More importantly, our method offers full compatibility with high NA focusing geometry, allowing high-intensity and high-resolution applications. We validate these properties through femtosecond laser ablation under high-NA conditions and interferometric spatiotemporal field reconstruction under low-NA conditions.

Femtosecond-laser direct waveguide writing is progressively emerging as an alternative to conventional techniques to develop complex photonic devices, for applications ranging from classical and quantum information processing, to sensing and metrology. Laser written waveguides typically offer low modal birefringence, thus preserving coherence of polarization-encoded information. Integrated waveplates have been reported, as waveguides with tilted birefringence axis, but with limited flexibility in terms of achievable rotation angle, birefringence magnitude or control in the modal shape. Here we investigate the multi-scan approach to realize low-loss optical waveguides in fused silica substrate with controlled modal birefringence. We show that by tuning the horizontal and vertical shifts between subsequent scans we can independently change both the magnitude and the axis inclination of the birefringence, while keeping efficient mode coupling with standard fibers.

This study proposes a novel adaptive finite volume-particle method (AFVPM) for accurate and efficient free surface flow simulations. The proposed AFVPM synergistically combines the Eulerian finite volume method (FVM) on unstructured meshes with the Lagrangian smoothed particle hydrodynamics (SPH) approach. Specifically, the mesh-based FVM is employed in the bulk flow regions to leverage its computational efficiency and numerical accuracy, while a weakly compressible SPH formulation is applied in the vicinity of the interface to maintain robust free-surface tracking capabilities. A key innovation of this framework is a block-based dynamic and adaptive conversion strategy between Eulerian mesh regions and Lagrangian particle regions and a buffer region-based cell-particle algorithm is designed to ensure seamless data communication across the Eulerian mesh-Lagrangian particle interface. Furthermore, isothermal gas-kinetic scheme (GKS) incorporating gravitational effects is utilized to calculate the fluxes in the mesh regions. The performance and reliability of the proposed AFVPM are validated through a series of benchmark cases that involve complex free surface phenomena. Numerical results demonstrate that AFVPM achieves superior accuracy and efficiency compared to full SPH approaches.

Non-uniform product (color) distribution in colorimetric paper-based sensors affects the accuracy and reliability of measurements. The underlying mechanisms responsible for this are still unclear. The coffee ring effect explains the ring-formation at the periphery. However, ring-like patterns can also be found at intermediate radial positions in these sensors. In this work, we study the influence of mass transport and reaction dynamics within porous/paper substrates on the spatial product distribution. We consider one reactant embedded in a porous substrate, which reacts with another delivered through a sessile droplet. The process is modeled in two stages. In Stage 1, droplet imbibition creates two distinct flow domains in the substrate with moving boundaries. Stage 2 commences after complete penetration. Species-substrate interactions are addressed by including a mobility factor. The developed model is used to analyze the effects of different parameters on the product distribution for two configurations, Reagent-Embedded (RE) and Analyte-Embedded (AE). Our work demonstrates ring-like patterns can form even without evaporation effects. With decreasing analyte-reagent concentration ratio, the profile shifts inward. Thicker, more porous substrates yield greater uniformity but reduce color intensity. Immobilization of embedded species enhances uniformity in RE configuration with mobile product, and in AE configuration with immobile product. The model is validated with lead and nitrite detection experiments for RE and AE configurations respectively. It successfully captures three spatial color variations observed experimentally. This study also explains the emergence of multiple rings in these systems. The insights gained are useful for optimizing sensor design and protocols for colorimetry

Beamsplitters represent fundamental components in both classical and quantum optical systems, enabling the distribution of light, as well as the generation of interference, superposition and entanglement. However, optical networks constructed from conventional bulk 2x2-beamsplitters encounter inherent scalability issues, as the number of required beamsplitters scales quadratically with the number of optical modes for a fully connected network. Metasurfaces offer a promising route to overcome these constraints. By manipulating light at the wavelength scale compact optical components with advanced functionalities can be constructed, which address several modes simultaneously. In this work, we design and experimentally utilize a metasurface as a multiport beamsplitter. Furthermore, we realize a multimode interferometer composed of two cascaded metasurfaces. We characterize the individual and cascaded metasurfaces using classical light, showing controllable splitting ratios through tunable phase relations. We then expand the approach to quantum light, employing single photons to demonstrate second- and third-order photon correlations, as well as single photon interference across multiple spatial paths. These results establish metasurface-based multiport beamsplitters as a scalable and reconfigurable platform bridging classical and quantum photonics.

Pulse-to-pulse fluctuations in mode-locked lasers fundamentally limit applications from optical frequency combs to supercontinuum generation. While timing jitter has been extensively characterized, pulse amplitude and width fluctuations remain less accessible experimentally. We present a statistical autocorrelation method that demonstrates pulse breathing dynamics through Fano factor analysis of second-harmonic generation autocorrelation. This reveals a characteristic M-shaped Fano profile with maxima at the autocorrelation shoulders, a signature of pulse shape dynamics that is invisible to time-averaged diagnostics. Applying this method to a passively mode-locked oscillator, we measure a bound for the pulse width fluctuations of between 10.4 and 12.0\,fs ($\sim5.0-5.7\%$ of the 210 fs FWHM). This provides a theoretical projection of $\sim3\%$ spectral bandwidth fluctuations in photonic crystal fiber supercontinuum generation using this laser. This diagnostic capability opens the door to identifying and suppressing specific breathing mechanisms, paving the way for the design of ultra-stable oscillators required for precision frequency metrology.

The growing demand for high-capacity, low-loss short-reach links in highly integrated electronic systems makes it necessary to understand how terahertz (THz) dielectric waveguides behave in realistic PCB-level packaging environments. In this article, we investigate the channel transmission of a 3D-printed polypropylene dielectric waveguide placed near representative PCB substrates. Continuous-wave THz measurements are carried out for bare, fully copper-clad, and periodic copper-trace PCBs with different waveguide-PCB separations, while terahertz time-domain spectroscopy is used to characterize the dielectric properties of the substrate materials. In parallel, an equivalent radiation-channel model and simulations are employed to interpret the proximity-induced excess attenuation. Most notably, direct contact with a bare PCB produces severe and frequency-selective excess loss, whereas a waveguide-facing copper layer suppresses substrate-assisted leakage and reduces the excess loss, with the attenuation rapidly decreasing as the clearance increases. These results reveal the dominant near-field coupling mechanisms between dielectric waveguides and nearby PCB structures, and provide practical guidance for packaging-aware THz interconnect design.

Interlayer excitons (IXs) in two-dimensional (2D) van der Waals heterostructures have attracted considerable attention due to their unique optical and electronic properties. Owing to the spatially indirect nature, the radiative emission efficiency highly sensitive to interlayer twist angles. Further considering that their uniformly oriented out-of-plane dipole moments limit directional emission, strategies to simultaneously improve emission efficiency and induce optical anisotropy warrant in-depth investigation. In this work, we report significant photoluminescence (PL) enhancement and optical anisotropy of IXs in 2L-MoSe2/perovskite heterostructures mediated by energy transfer from ReS2. We attribute this enhancement to Forster resonance energy transfer (FRET), which increases the 2L-MoSe2 emission by approximately eight-fold at room temperature, and nearly doubles the emission intensity of momentum-indirect IXs in 2L-MoSe2/perovskite heterostructures at 78 K. Importantly, the optical anisotropy of ReS2 can be effectively imprinted onto 2L-MoSe2 and associated indirect IXs during the energy transfer process, yielding a linear dichroism of approximately 1.1 for both intralayer excitons and IXs with identical polarization directions. These findings expand the scope of IX study beyond direct bandgap materials with strong intrinsic emission to include systems with indirect bandgaps, offering new avenues for realizing high-performance polarization-sensitive optoelectronic devices.

Two data-driven modal analysis approaches, proper orthogonal decomposition (POD) and dynamic mode decomposition (DMD), are applied to analyze the unsteady flow obtained by solving the Reynolds-averaged Navier-Stokes (RANS) equations in a 1.5-stage axial turbine. The reduced-order reconstructed pressure, dominant mode shapes, and dynamic features of these dominant modes in the downstream stator of the turbine are compared between POD and four DMD variants. It is found that the DMD methods based on the amplitude criterion, the Tissot criterion, and the sparsity-promoting DMD (SP-DMD) achieve reconstruction accuracy comparable to that of POD, while the frequency criterion proves unsuitable for the present problem. The second and third POD and DMD modes capture the dominant pressure fluctuation structures within the stator, and there is similarity between the corresponding POD and DMD spatial modes. The unsteady flow is primarily governed by neutral DMD modes characterized by high amplitudes and low frequencies corresponding to the basic and harmonic frequencies driven by the rotor passing frequency. While the POD analysis provides accurate reconstruction for the original snapshots, the time evolution of each POD mode does not reflect the true dynamic feature of the system. In particular, they misrepresent the fundamental frequencies of the problem. In addition, the correlations between the dominant modes in the downstream stator and the turbine adiabatic efficiency are explored across different stator clocking configurations. It is found that a clocking configuration with higher adiabatic efficiency at 50% span corresponds to a larger spatial and time component of the second and third DMD mode pair, and similarly a larger second POD mode.

High-resolution brain imaging can now capture not just synapse locations but their molecular composition, with the cost of such mapping falling exponentially. Yet such ultrastructural data has so far told us little about local neuronal physiology - specifically, the parameters (e.g., synaptic efficacies, local conductances) that govern neural dynamics. We propose to translate molecularly annotated ultrastructure into physiology, introducing the concept of an ultrastructure-to-dynamics compiler: a learned mapping from molecularly annotated ultrastructure to simulator-ready, uncertainty-aware physiological parameters. The requirement is paired training data, with jointly acquired ultrastructure from imaging, and dynamical responses to perturbations from physiological experiments. With this data we can train models that predict local physiology directly from structure. Such a compiler would support biophysical simulations by turning anatomical maps into models of circuit dynamics, shifting structure-to-function from a descriptive program to a predictive one and opening routes to understanding neural computation and forecasting intervention effects.

Short tandem repeats (STRs) are low-entropy regions in the genome, consisting of a short (1-6 bp) unit that is consecutively repeated multiple times. They are known for high mutational instability, due to so-called stutter-mutations, in which the number of units in the run increases or descreases. In particular, STRs with repeat unit length of 1-2 bp are prone to mutate even within several cell divisions. The extremely rapid accumulation of variation makes them interesting phylogenetic markers for retrospective single-cell lineage reconstruction. Here we model their mutational dynamics at the level of individual repeat unit type and then aggregate length variations over many STR loci with the aim of obtaining a very fast ``molecular clock''. We calibrate our model based on several datasets with known lineage structure prepared from cultured cells. We find that the mutational dynamics of STRs are reasonably consistent for a given cell line, but vary among different ones. This suggests that the dynamics are not entirely explained by mutations in caretaker genes, rather, various other factors play a role -- possibly tissue origin and differentiation state. Further data and research is necessary to asses their relative effects.

In a dataset of 423 patients who had had radical prostatectomy for localised prostate cancer we estimated the apparent Shannon information (ASI) about time to biochemical recurrence in various subsets of the available pre-op variables using a Bayesian Gamma-power-mixture survival regression model. In all the subsets examined the ASI was positive with posterior probability greater than 0.975 . Using only age and results of pre-operative blood tests (PSA and biomarkers) we achieved 0.232 (0.180 to 0.290) nats ASI (0.335 (0.260 to 0.419) bits) (posterior mean and equitailed 95% posterior confidence intervals). This is more than double the mean posterior ASI previously achieved on the same dataset by a subset of the current authors using a log-skew-Student-mixture model, and is greater than that previous value with posterior probability greater than 0.99 . Additionally using pre- or post-operative Gleason grades, operative findings, clinical stage, and presence or absence of extraprostatic extension or seminal vesicle invasion did not increase the ASI extracted. However removing the blood-based biomarkers and replacing them with either pre-operative Gleason grades or findings available from MRI scanning greatly reduced the available ASI to respectively 0.077 (0.038 to 0.120) and 0.088 (0.045 to 0.132) nats (both less than the values using blood-based biomarkers with posterior probability greater than 0.995). A greedy approach to selection of the best biomarkers gave TGFbeta1, VCAM1, IL6sR, and uPA in descending order of importance from those examined.

If the universe follows a specific design, then a central question is which cost function is optimized by the observed forces. This is the problem of inverse optimal control, or inverse reinforcement learning, in which a reward function is inferred from the dynamics of the observed system. We first establish the {\em least cost principle}, whereby the laws of motion can be derived from minimization of a time-discounted integral of the acceleration cost minus a state-dependent reward function. After determining the functional form of the acceleration cost from basic principles, we infer the reward function from the laws of motion governing classical gravitation and Coulomb forces. The inferred reward function is high when pairs of particles have high relative velocities and when their relative motion is orthogonal to their distance vectors. All in all, our work suggests that relative motion and quasi-circular orbits are the dynamical and static features optimized by central forces in nature.

In this paper we study the stability properties of the equilibrium point for an age-structured chemostat model with renewal boundary condition and coupled substrate dynamics under constant dilution rate. This is a complex infinite-dimensional feedback system. It has two feedback loops, both nonlinear. A positive static loop due to reproduction at the age-zero boundary of the PDE, counteracted and dominated by a negative dynamic loop with the substrate dynamics. The derivation of explicit sufficient conditions that guarantee global stability estimates is carried out by using an appropriate Lyapunov functional. The constructed Lyapunov functional guarantees global exponential decay estimates and uniform global asymptotic stability with respect to a measure related to the Lyapunov functional. From a biological perspective, stability arises because reproduction is constrained by substrate availability, while dilution, mortality, and substrate depletion suppress transient increases in biomass before age-structure effects can amplify them. The obtained results are applied to a chemostat model from the literature, where the derived stability condition is compared with existing results that are based on (necessarily local) linearization methods.

Modeling cellular states and predicting their responses to perturbations are central challenges in computational biology and the development of virtual cells. Existing foundation models for single-cell transcriptomics provide powerful static representations, but they do not explicitly model the distribution of cellular states for generative simulation. Here, we introduce Lingshu-Cell, a masked discrete diffusion model that learns transcriptomic state distributions and supports conditional simulation under perturbation. By operating directly in a discrete token space that is compatible with the sparse, non-sequential nature of single-cell transcriptomic data, Lingshu-Cell captures complex transcriptome-wide expression dependencies across approximately 18,000 genes without relying on prior gene selection, such as filtering by high variability or ranking by expression level. Across diverse tissues and species, Lingshu-Cell accurately reproduces transcriptomic distributions, marker-gene expression patterns and cell-subtype proportions, demonstrating its ability to capture complex cellular heterogeneity. Moreover, by jointly embedding cell type or donor identity with perturbation, Lingshu-Cell can predict whole-transcriptome expression changes for novel combinations of identity and perturbation. It achieves leading performance on the Virtual Cell Challenge H1 genetic perturbation benchmark and in predicting cytokine-induced responses in human PBMCs. Together, these results establish Lingshu-Cell as a flexible cellular world model for in silico simulation of cell states and perturbation responses, laying the foundation for a new paradigm in biological discovery and perturbation screening.

What substrate features allow life? We exhaustively classify all 262,144 outer-totalistic binary cellular automata rules with Moore neighbourhood for self-replication and produce phase diagrams in the $(λ, F)$ plane, where $λ$ is Langton's rule density and $F$ is a background-stability parameter. Of these rules, 20,152 (7.69%) support pattern proliferation, concentrated at low rule density ($λ\approx 0.15$--$0.25$) and low-to-moderate background stability ($F \approx 0.2$--$0.3$), in the weakly supercritical regime (Derrida coefficient $μ= 1.81$ for replicators vs. $1.39$ for non-replicators). Self-replicating rules are more approximately mass-conserving (mass-balance 0.21 vs. 0.34), and this generalises to $k{=}3$ Moore rules. A three-tier detection hierarchy (pattern proliferation, extended-length confirmation, and causal perturbation) yields an estimated 1.56% causal self-replication rate. Self-replication rate increases monotonically with neighbourhood size under equalised detection: von Neumann 4.79%, Moore 7.69%, extended Moore 16.69%. These results identify background stability and approximate mass conservation as the primary axes of the self-replication phase boundary.

The evolution toward 6G communication systems is expected to rely on integrated three-dimensional network architectures where terrestrial infrastructures coexist with non-terrestrial stations such as satellites, enabling ubiquitous connectivity and service continuity. In this context, accurate channel models for satellite-to-ground propagation in urban environments are essential, particularly for user equipment located at street level where obstruction and multipath effects are significant. This work investigates satellite-to-urban propagation through deterministic ray-tracing simulations. Three representative urban layouts are considered, namely dense urban, urban, and suburban. Multiple use cases are investigated, including handheld devices, vehicular terminals, and fixed rooftop receivers operating across several frequency bands. The analysis focuses on the relative importance of competing propagation mechanisms and on two key channel parameters, namely the Rician K-factor and the delay spread, which are relevant for the calibration of channel models to be used in link- and system-level simulations. Results highlight the strong - and in some cases unconventional - dependence of channel dispersion and fading characteristics on satellite elevation, antenna placement, and urban morphology.

Reflecting intelligent surface (RIS) is a promising technology for 6G mobile communications. However, identifying the niche of RIS within the mobile networks is a challenging task. To mitigate the escalating system complexity of mobile networks, we propose the concept of Intelligent Reflection as a Service (IRaaS), and discuss its system architecture, enabling technologies, and deployment strategy, respectively. By leveraging technologies such as resource pooling, service based architecture (SBA), cloud infrastructure, and model-free signal processing, IRaaS empowers telecom operators to deliver on-demand intelligent reflection services without a radical update of current communication protocols. In addition, IRaaS brings a novel deployment strategy that creates new opportunities for the vendors of intelligent reflection service and balances the interests of both telecom operators and property owners. IRaaS is expected to speed up the rollout of RIS from both technical perspective and commercial perspective, fostering an authentic smart radio environment for future mobile communications.

The number of low Earth orbit (LEO) satellite constellations has grown rapidly in recent years, bringing a major change to global wireless communications. As LEO satellite links take on a growing role in critical services such as emergency communications, navigation, wide-area data collection, and military operations, keeping these links secure has become an important concern. In particular, verifying the identity of a satellite transmitter is now a basic requirement for protecting the services that rely on satellite access. In this article, we propose an active challenge-response authentication framework in which the verifier checks the satellite at randomly chosen times that are not known in advance, removing the fixed measurement window that existing passive methods expose to adversaries. The proposed framework uses the deterministic yet unpredictably sampled nature of orbital observables to establish a physics based root of trust for satellite identity authentication. This approach transforms satellite authentication from static feature matching into a spatiotemporal consistency verification problem inherently constrained by orbital dynamics, providing robust protection even against trajectory-aware spoofing attacks.

This paper investigates an uplink multi-user covert communication system with spatially distributed users. Unlike prior works that approximate channel statistics using averaged parameters and homogeneous assumptions, this study explicitly models each user's geometric position and corresponding user-to-Willie and user-to-Bob channel variances. This approach enables an accurate characterization of spatially heterogeneous covert environments. We mathematically prove that a generalized on-off power control scheme, which jointly accounts for both Bob's and Willie's channels, constitutes the optimal transmission strategy in heterogeneous user configurations. Leveraging the optimal strategy, we derive closed-form expressions for the minimum detection error probability and the minimum number of cooperative users required to satisfy a covert constraint. With the closed-form expressions, comprehensive theoretical analyses are conducted, which are validated by Monte-Carlo simulations. One important insight obtained from the analysis is that user spatial heterogeneity can enhance covert communication performance. Building on these findings, a piecewise search algorithm is proposed to achieve exact optimality with significantly reduced computational complexity. We demonstrate that optimization considering user's spatial heterogeneity achieves substantially improved covert communication performance than that based on the assumption of spatial homogeneity.

Beyond-diagonal reconfigurable intelligent surface (BD-RIS) generalizes the conventional diagonal RIS (D-RIS) by introducing tunable inter-element connections, offering enhanced wave manipulation capabilities. However, realizing the advantages of BD-RIS requires accurate channel state information (CSI), whose acquisition becomes significantly more challenging due to the increased number of channel coefficients, leading to prohibitively large pilot training overhead in BD-RIS-aided multi-user multiple-input multiple-output (MU-MIMO) systems. Existing studies reduce pilot overhead by exploiting the channel correlations induced by the Kronecker-product or multi-linear structure of BD-RIS-aided channels, which neglect the spatial correlation among antennas and the statistical correlation across RIS-user channels. In this paper, we propose a learning-based channel estimation framework, namely the joint training scattering matrix learning and channel estimation framework (JTSMLCEF), which jointly optimizes the BD-RIS training scattering matrix and estimates the cascaded channels in an end-to-end manner to achieve accurate channel estimation and reduce the pilot overhead. The proposed JTSMLCEF follows a two-phase channel estimation protocol to enable adaptive training scattering matrix optimization with a training scattering matrix optimizer (TSMO) and cascaded channel estimation with a dual-attention channel estimator (DACE). Specifically, the DACE is designed with intra-user and inter-user attention modules to capture the multi-dimensional correlations in multi-user cascaded channels. Simulation results demonstrate the superiority of JTSMLCEF. Compared with the current state-of-the-art method, it reduces the pilot overhead by $80\%$ while further reducing the normalized mean squared error (NMSE) by $82.6\%$ and $92.5\%$ in indoor and urban micro-cell (UMi) scenarios, respectively.

Most Rate-Splitting Multiple Access (RSMA) implementations rely on successive interference cancellation (SIC) at the receiver, whose performance is inherently limited by error propagation during common-stream decoding. This paper addresses this issue by developing a SIC-free RSMA receiver based on joint demapping (JD), which directly evaluates bit vectors over a composite constellation. Using a two-user Multiple-Input Single-Output (MISO) prototype, we conduct over-the-air measurements to systematically compare SIC and JD-based receivers. The results show that the proposed SIC-free receiver provides stronger reliability and better practicality over a wider operating range, with all observations being consistent with theoretical expectations.

Mechanical vibration monitoring often requires high sampling rates and generates large data volumes, posing challenges for storage, transmission, and power efficiency. Compressive Sensing (CS) offers a promising approach to overcome these constraints by exploiting signal sparsity to enable sub-Nyquist acquisition and efficient reconstruction. This study presents a comprehensive comparative analysis of the key components of the CS framework: sparse basis, measurement matrix, and reconstruction algorithm for machinery vibration signals. In addition, a hardware-efficient measurement matrix, the Wang matrix, originally developed for image compression, is introduced and evaluated for the first time in this context. Experimental assessment using the HUMS2023 and the CETIM gearbox datasets demonstrates that this matrix achieves superior reconstruction quality, with higher SNR, compared to conventional Gaussian and Bernoulli matrices, especially at high compression ratios.

For extremely large-scale arrays (XL-arrays), the discrete Fourier transform (DFT) codebook, conventionally used in the far-field, has recently been employed for near-field beam training. However, most existing methods rely on the line-of-sight (LoS) dominant channel assumption, which may suffer degraded communication performance when applied to the general multi-path scenario due to the more complex received signal power pattern at the user. To address this issue, we propose in this paper a new hybrid learning-and-optimization-based beam training method that first leverages deep learning (DL) to obtain coarse channel parameter estimates, and then refines them via a model-based optimization algorithm, hence achieving high-accuracy estimation with low computational complexity. Specifically, in the first stage, a tailored U-Net architecture is developed to learn the non-linear mapping from the received power pattern to coarse estimates of the angles and ranges of multi-path components. In particular, the inherent permutation ambiguity in multi-path parameter matching is effectively resolved by a permutation invariant training (PIT) strategy, while the unknown number of paths is estimated based on defined path existence logits. In the second stage, we further propose an efficient particle swarm optimization method to refine the angular and range parameters within a confined search region; in the meanwhile, a Gerchberg-Saxton algorithm is used to retrieve multi-path channel gains from the received power pattern. Last, numerical results demonstrate that the proposed hybrid design significantly outperforms various benchmarks in terms of parameter estimation accuracy and achievable rate, yet with low computational complexity.

In this paper, we investigate the performance of a fluid antenna relay (FAR)-assisted downlink communication system utilizing non-orthogonal multiple access (NOMA). The FAR, which integrates a fluid antenna system (FAS), is equipped on an autonomous aerial vehicle (AAV), and introduces extra degrees of freedom to improve the performance of the system. The transmission is divided into a first phase from the base station (BS) to the users and the FAR, and a second phase where the FAR forwards the signal using amplify-and-forward (AF) or decode-and-forward (DF) relaying to reduce the outage probability (OP) for the user maintaining weaker channel conditions. To analyze the OP performance of the weak user, Copula theory and the Gaussian copula function are employed to model the statistical distribution of the FAS channels. Analytical expressions for weak user's OP are derived for both the AF and the DF schemes. Simulation results validate the effectiveness of the proposed scheme, showing that it consistently outperforms benchmark schemes without the FAR. In addition, numerical simulations also demonstrate the values of the relaying scheme selection parameter under different FAR positions and communication outage thresholds.

For a family of two-matrix models \[ \frac{1}{2} \operatorname{Tr}(A^2+B^2) - \frac{g}{4} \operatorname{Tr}(A^4+B^4) - \begin{cases} \frac{h}{2} \operatorname{Tr}( A BA B) \\ \frac{h}{4} \operatorname{Tr}( A BA B+ ABBA ) \\ \frac{h}{2} \operatorname{Tr}( A B BA ) \end{cases} \] with hermitian $A$ and $B$, we provide, in each case, a Monte Carlo estimate of the boundary of the maximal convergence domain in the $(h,g)$-plane. The results are discussed comparing with exact solutions (in agreement with the only analytically solved case) and phase diagrams obtained by means of the functional renormalization group.

Motivated by quantum gravity on spacetimes with multi-scale geometry, we analyze quantum field theories with a self-adjoint fractional power $(\Box^2)^{γ/2}$ of the d'Alem\-bert\-ian in the kinetic term, for any real $γ>0$. Selecting a particularly simple version of the kinetic term which we call hermitian polynomial, we study the spectral decomposition of the propagator and, when $γ>1$, obtain the standard mass singularity $-k^2=m^2$. This is the only mode in the perturbative spectrum of asymptotic states, since the only other content of the theory is a cloud of purely virtual particles with complex masses. We also show that other versions of the self-adjoint fractional kinetic term lead to a different distribution of the virtual complex modes but to the same physical spectrum for $0<γ<3$, thus addressing the issue of uniqueness in this class of nonlocal theories. The non-hermitian version of the theory has the $-k^2=m^2$ particle plus a continuum of standard massive modes. Finally, we prove that unitarity of scalar, gauge and gravity models is respected at all perturbative orders if, in the hermitian cases, one adopts the fakeon prescription on scattering amplitudes or, in the non-hermitian case, $0<γ<1$ or $2<γ<3$ with the standard Feynman prescription. These results drastically simplify previous characterizations of fractional quantum gravity, which is super-renormalizable for $γ>2$.

Gaussian baths are widely used to model non-Markovian environments, yet the cost of accurate simulation at long times remains poorly understood, especially when spectral densities exhibit nonanalytic behavior as in a range of realistic models. We rigorously bound the complexity of representing bath correlation functions on a time interval $[0,T]$ by sums of complex exponentials, as employed in recent variants of pseudomode and hierarchical equations of motion methods. These bounds make explicit the dependence on the maximal simulation time $T$, inverse temperature $β$, and the type and strength of singularities in an effective spectral density. For a broad class of spectral densities, the required number of exponentials is bounded independently of $T$, achieving time-uniform complexity. The $T$-dependence emerges only as polylogarithmic factors for spectral densities with strong singularities, such as step discontinuities and inverse power-law divergences. The temperature dependence is mild for bosonic environments and disappears entirely for fermionic environments. Thus, the true bottleneck for long-time simulation is not the simulation duration itself, but rather the presence of sharp nonanalytic features in the bath spectrum. Our results are instructive both for long-time simulation of non-Markovian open quantum systems, as well as for Markovian embeddings of classical generalized Langevin equations with memory kernels.

We show that the wavefunction of the universe in theories of conformally coupled scalars in power-law Friedmann-Robertson-Walker (FRW) cosmologies satisfies a graphical coaction, by means of which we can understand its complete analytic structure in terms of the acyclic minors of Feynman graphs. Our construction extends to all particle multiplicities and any loop order, and if we isolate certain weight-one contributions, it reproduces the ``kinematic flow'' that encodes the differential equation of the wavefunction coefficients. Similarly, any discontinuity of the wavefunction coefficient is easily extracted from the coaction.

In Einstein gravity, matter with an arbitrarily small density can be a black hole. Pressure in the star diverges if size of the star is smaller than 9/8 of the Schwarzschild radius, implying the gravitational collapse into a black hole. By taking quantum effects of matter, however, pressure is bounded from above, and a core with negative energy appears instead. Density of matter increases and eventually reaches the cut-off scale as size of the star approaches the Schwarzschild radius. This result implies that density must be very large as the Planck scale if the star is put inside the Schwarzschild radius.

In this work, we obtain exact solutions to the $(n+2)$-dimensional Einstein Field Equations with a non-zero cosmological constant for $n > 1$. These solutions depend on a set $\{ A_a, a=1,2,\ldots , m \}$ of pairwise commuting constant matrices in $\mathfrak{sl} ( n, \mathbb{R} )$ and on a constant matrix $g_0$ in $\mathcal{I} (\{ A_a, a=1,\ldots , m \})$, determined in previous work. Different choices of $\{ A_a, a=1,\ldots , m \}$ and $g_0$ correspond to different solutions. As examples, we show how to obtain the de Sitter metric, the Anti-de Sitter metric, the Birmingham metric, the Nariai metric and the Anti-Nariai metric in higher dimensions. The generalized Nariai and Anti-Nariai solutions are direct topological products of $AdS_{\frac{n}{2} + 1} \times H^{\frac{n}{2} + 1}$, $dS_{\frac{n}{2} + 1} \times S^{\frac{n}{2} + 1}$, $AdS_2 \times H^n$, $AdS_n \times H^2$, $dS_2 \times S^n$ and $dS_n \times S^2$. In addition, we study a solution in the context of cosmology.

In this methods paper, we show how to tridia\-go\-nalize two families of bosonic multimode systems: optomechanical and Bose-Hubbard hamiltonians. Using tools from number theory, we devise a rendering of these systems in the form of exact $D \times D$ tridiagonal symmetric matrices with real-valued entries. Such matrices can subsequently be exactly diagonalized using specialized sparse-matrix algorithms that need on the order of $D \ln(D)$ steps. This makes it possible to describe systems with much larger numbers of basis states than available to date. It also allows for efficient diagonal representation of large, accurate, symplectic split-operator propagators for which we moreover show that the required basis changes can be implemented by simple re-indexing, at marginal computational cost.

In this article, we study the repetitive Penrose process for the Kerr-Taub-NUT black hole (BH). First of all, we briefly review the spacetime of the Kerr-Taub-NUT BH, including horizon and ergosphere structures. The results indicate that the event horizon and ergosphere radii increase under the influence of the gravitomagnetic charge $l$. Subsequently, we find by using the irreducible mass of the BH that the extractable energy decreases with the rise of the gravitomagnetic charge. We then turn to the repetitive Penrose process by writing the conservation laws and setting the corresponding iterative stopping conditions. Furthermore, we numerically calculate the change in the BH's parameters, along with the corresponding quantities of the repetitive Penrose process, for each iteration.

Counterdiabatic (CD) driving enables efficient quantum state preparation, but it requires implementing highly nonlocal adiabatic gauge potentials (AGP) that are impractical to compute and realize in large many-body systems. We introduce a \textit{weighted nested-commutator} (WNC) ansatz to approximate AGP using local operators. The WNC ansatz generalizes the standard nested-commutator ansatz by assigning independent variational weights to commutators of local Hamiltonian terms, thereby enlarging the variational space while preserving a fixed operator range. We show that the WNC ansatz can be efficiently optimized using a local optimization scheme. Moreover, it systematically outperforms the nested-commutator ansatz in preparing one-dimensional matrix product states (MPS) and the ground state of a nonintegrable quantum Ising model. We then numerically demonstrate that CD driving based on the WNC ansatz significantly accelerates the preparation of 1D MPS for system sizes up to $N = 1000$ qubits, as well as the two-dimensional Affleck-Kennedy-Lieb-Tasaki state on a hexagonal lattice with up to $N = 3 \times 10$ sites.

We study Generalized Proca theories inspired by the recent regularised Proca theory of four-dimensional Gauss-Bonnet gravity. By abandoning the rigid constraints typically imposed by specific regularization schemes, we treat the coefficients of the terms in the action as free parameters. This approach uncovers a broader solution space that admits static and spherically symmetric black hole solutions characterized by primary hair, where, surprisingly, the cosmological constant arises naturally as a constant of integration even in the absence of a bare cosmological term.

We investigate a class of interacting dark energy (IDE) models arising from density-driven spontaneous symmetry breaking in a conformally coupled scalar-tensor framework. In this construction, the dark matter-scalar interaction is dynamically activated as the cosmological density evolves, and the redshift dependence of the coupling follows a logistic profile whose steepness is determined by the local curvature of the symmetry-breaking potential. Working in the controlled adiabatic tracking regime, we implement the resulting epoch-dependent interaction in a perturbative background close to $Λ$CDM and confront the model with late-time cosmological data, including Planck 2018 CMB lensing reconstruction, redshift-space distortions, and Pantheon+SH0ES supernova data. We analyze realizations in which the activation index is allowed to vary and compare them with a restricted realization in which it is fixed to the canonical quadratic minimum value, thereby probing the structural role of the activation profile. We find no statistically significant preference for interaction over $Λ$CDM; current observations constrain the model to a hierarchical regime in which the scalar remains heavier than the Hubble scale at activation and background deformations remain perturbatively small. Allowing the activation index to vary preserves an extended degeneracy direction in parameter space, whereas fixing it removes this freedom and leads to a contraction of the allowed posterior region once geometric and growth data are combined. Our results delineate the viable parameter regime of symmetry-breaking IDE and clarify the structural distinction between microphysically motivated scalar-tensor realizations and phenomenological interacting models.

We extend Quantum Computing Quantum Monte Carlo (QCQMC) beyond ground-state energy estimation by systematically constructing the quantum circuits used for state preparation. Replacing the original Variational Quantum Eigensolver (VQE) prescription with task-adapted unitaries, we show that QCQMC can address excited-state spectra via Variational Fast Forwarding and the Variational Unitary Matrix Product Operator (VUMPO), combinatorial optimization via a symmetry-preserving VQE ansatz, and finite-temperature observables via Haar-random unitaries. Benchmarks on molecular, condensed-matter, nuclear-structure, and graph-optimization problems demostrate that the QMC diffusion step consistently improves the energy accuracy of the underlying state-preparation method across all tested domains. For weakly correlated systems, VUMPO achieves near-exact energies with significantly shallower circuits by offloading optimization to a classical tensor-network pre-training step, while for strongly correlated systems, the QMC correction becomes essential. We further provide a proof-of-concept demonstration that Haar-random basis state preparation within QCQMC yields finite-temperature estimates from pure-state dynamics.

Particle physics and gravitational waves provide complementary probes of the deep structure of the Universe. Gravitational waves from the mergers of neutron stars and black holes are sensitive to the structure of dense quark matter and to different dark matter scenarios. Measurements of stochastic gravitational waves backgrounds can teach us about possible first order phase transitions in the early Universe, including providing sensitivity to the TeV scale which is of key interest to future particle collider experiments. Gravitational waves measurements will also give new probes of the evolution and expansion of the Universe, complementary to measurements with electromagnetic radiation. This Perspectives article explores the physics synergies between the science opportunities provided by next generation gravitational waves measurements and particle physics experiments. Gravitational waves can also probe deep into the early Universe reaching physics much above possible collider energies if the signals can be detected.

This paper explores an efficient method for entanglement quantification in two-qubit and qubit-qutrit quantum systems based upon the framework of collective measurements in conjunction with machine learning. We introduce an adaptive measurement procedure in which measurement settings are dynamically adjusted based on prior measurement outcomes aiming to optimize the inference precision given a limited number of these measurement settings. The procedure makes use of the Long Short-Term Memory networks to recurrently process collective measurements on two copies of the investigated states. Obtained results demonstrate the tangible benefits of the adaptive measurements in comparison to previously described non-adaptive strategies.

Bitcoin already faces a quantum threat through Shor attacks on elliptic-curve signatures. This paper isolates the other component that public discussion often conflates with it: mining. Grover's algorithm halves the exponent of brute-force search, promising a quadratic edge to any quantum miner of Bitcoin. Exactly how large that edge grows depends on fault-tolerant hardware. No prior study has costed that hardware end to end. We build an open-source estimator that sweeps the full attack surface: reversible oracles for double-SHA-256 mining and RIPEMD-based address preimages, surface-code factory sizing, fleet logistics under Nakamoto-consensus timing, and Kardashev-scale energy accounting. A parametric sweep over difficulty bits b, runtime caps, and target success probabilities reveals a sharp transition. At the most favourable partial-preimage setting (b = 32, 2^224 marked states), a superconducting surface-code fleet still requires about 10^8 physical qubits and about 10^4 MW. That load is comparable to a large national grid. Tightening to Bitcoin's January 2025 mainnet difficulty (b about 79) explodes the bill to about 10^23 qubits and about 10^25 W, approaching the Kardashev Type II threshold. These numbers settle a narrower question than "Is Bitcoin quantum-secure?" Once Grover mining is lifted from asymptotic query counts to fault-tolerant physical cost, practical quantum mining collapses under oracle, distillation, and fleet overhead. To push mining into non-trivial consensus effects, one must invoke astronomical quantum fleets operating at energy scales that lie far above present-day civilization.

Time-local generators of open quantum systems are generically divergent at long times, even though the reduced dynamics remains regular. We construct, by analytic continuation, nonperturbative dynamical maps consistent with these generators. For the weak-coupling unbiased spin--boson model, this construction yields an explicit dynamical map that nonperturbatively resums the TCL generator and exposes how the divergences signal the approach to a singular time at which the reduced dynamics becomes noninvertible. The reconstructed map is validated against TEMPO simulations at short times and the exactly solvable rotating-wave model at all times. In the full spin--boson model, the same continuum mechanism produces both an early-time anisotropy, with a measurable phase shift that provides a signature of the environmental correlation and the pointer direction, and a late-time singularity at which the reduced dynamics becomes noninvertible. By contrast, in the rotating-wave model the map approaches this point without reaching it and remains invertible at all times. These results establish a nonperturbative framework for reconstructing reduced dynamics from divergent time-local generators, diagnosing the onset of noninvertibility, and identifying experimentally accessible early-time signatures of environment-induced anisotropy.

We explore different implementations of the quantum Bell causality (QBC) condition in the quantum sequential growth (QSG) dynamics of causal set quantum gravity, for non-commuting transition operators. Assuming a non-singular dynamics we show that for the two most natural choices of operator orderings for the QBC, the transition operator algebra reduces to a commutative one. As a third choice, we take the operator ordering to depend on the size of the precursor set. We find several new commutation relations which further constrain the algebra but do not imply commutativity. On the other hand, if any of the generators of the ``antichain subalgebra'' belongs to its center, then this implies commutativity of the full algebra. The complexity of the algebra prevents us from obtaining a general form for the transition operators, which hinders computability. In an attempt to construct the simplest non-trivial d=2 representation, we find that a Pauli matrix representation of the generators of the antichain subalgebra leads to inconsistencies, implying that if a non-trivial representation exists, it must be higher dimensional. Our work can be viewed as a first step towards finding a non-commutative realisation of QSG.

Quantum Key Distribution (QKD) protocols require Information-Theoretically Secure (ITS) authentication of the classical channel to preserve the unconditional security of the distilled key. Standard ITS schemes are based on one-time keys: once a key is used to authenticate a message, it must be discarded. Since QKD requires mutual authentication, two independent one-time keys are typically consumed per round, imposing a non-trivial overhead on the net secure key rate. In this work, we present the authentication-with-response scheme, a novel ITS authentication scheme based on $\varepsilon$-Almost Strongly Universal$_2$ ($\varepsilon$-ASU$_2$) functions, whose IT security can be established in the Universal Composability (UC) framework. The scheme achieves mutual authentication consuming a single one-time key per QKD round, halving key consumption compared to the state-of-the-art.

We show that networks of quantum frames of reference, in which one frame may be used to produce multiple other frames that in their turn prepare systems which may interact with one another, have counterintuitive properties that make following the exchange of conserved quantities very subtle, and raise questions about the very nature of conserved quantities. In addition, we present an alternative approach to analysing quantum reference frames that we believe will be useful in discussions related to quantum frames of reference.

Quantum communication networks require transmission of high-fidelity, uncoded qubits for applications such as entanglement distribution and quantum key distribution. However, current implementations are constrained by limited buffer capacity and qubit decoherence, which degrades qubit quality while waiting in the buffer. A key challenge arises from the stochastic nature of qubit generation, there exists a random delay (D) between the initiation of a generation request and the availability of the qubit. This induces a fundamental trade off early initiation increases buffer waiting time and hence decoherence, whereas delayed initiation leads to server idling and reduced throughput. We model this system as an admission control problem in a finite buffer queue, where the reward associated with each job is a decreasing function of its sojourn time. We derive analytical conditions under which a simple "no lag" policy where a new qubit is generated immediately upon the availability of buffer space is optimal. To address scenarios with unknown system parameters, we further develop a Bayesian learning framework that adaptively optimizes the admission policy. In addition to quantum communication systems, the proposed model is applicable to delay sensitive IoT sensing and service systems.

We obtain exact expressions for the quasinormal modes (QNMs) of the massive scalar, fermionic and vector perturbations around a rotating BTZ-like black hole in the Einstein-bumblebee gravity. We find that the Lorentz symmetry breaking (LSB) parameter $\ell$ leaves its imprint only on the imaginary parts of the quasinormal frequencies and the corresponding perturbation field decays more slowly for a larger $\ell$, except for the left-moving quasinormal frequencies $ω_L$ with positive mass and the right-moving ones $ω_R$ with negative mass for the fundamental modes under the vector perturbation where the imaginary parts are independent of $\ell$. We also note that, regardless of the kind of perturbations, the real parts depend only on the angular quantum number, which are the same as those in the standard BTZ black hole. Furthermore, we investigate the AdS/CFT correspondence from the QNMs and observe that the expected universal relation for the left and right conformal weights ($h_L,h_R$) of the boundary operators dual to various fields still holds even for the BTZ-like black hole in the Einstein-bumblebee gravity. These results strongly support the AdS/CFT correspondence and could help us better understand the Einstein-bumblebee gravity with the Lorentz symmetry violation.

In this work, we derive the bipartite density matrix for the $e^{+}e^{-} \rightarrow J/ψ\rightarrow Λ(pπ^{-}) \barΛ(\bar{p}π^{+})$ process at BESIII. We evaluate the impact of mass corrections and memory effects (within Markovian and non-Markovian regimes) on quantum correlations even beyond entanglement. The dependence of these quantum properties on the scattering angle $\varphi$ is analyzed, with a particular focus on the impact of mass corrections. By comparing massless and mass-corrected scenarios, we demonstrate that the inclusion of mass effects enhances the maximum violation of the Bell inequality. While the qualitative temporal behavior remains unchanged, mass corrections quantitatively modify the angular distribution and introduce additional extrema at $\varphi=0$ and $\varphi=π$, thereby strengthening non-local correlations without altering their fundamental dynamical origin. An examination of the hierarchy of quantum correlations in baryon-antibaryon systems yields partial confirmation: $\text{Bell Nonlocality} \subset \text{Steering} \subset \text{Entanglement} \subset \text{Discord}$. Additionally, our results show that classical correlations serve to mitigate the decoherence and the decay of quantum correlations. This interplay between classical and quantum correlations suggests practical applications in quantum information and provides a robust framework for investigating baryon-antibaryon interactions.

The constancy of the gravitational constant $G$ is a cornerstone of the strong equivalence principle and of general relativity, yet its possible temporal variation remains a key target in tests of fundamental physics. Gravitational-wave (GW) astronomy, especially when combined with electromagnetic observations, provides an unprecedented new opportunity to probe this principle in the strong-field and dynamical regime. In this work, we develop a GW waveform model with a slowly varying gravitational constant, incorporating its effects both on compact binary dynamics and GW propagation in an expanding universe. Applying this framework to the binary neutron star merger GW170817, together with independent electromagnetic constraints on the luminosity distance, sky localization and binary inclination from GRB 170817A, we perform a joint Bayesian analysis that disentangles varying-$G$ effects from astrophysical degeneracies. We find no evidence for a temporal variation of the gravitational constant, and constrain its fractional time derivative to $\dot{G}/G \in [-3.36 \times 10^{-9}, 5.34\times10^{-10}]~{\rm yr^{-1}}$, representing the most stringent bounds obtained to date from real GW observations. Our results demonstrate the power of multi-messenger astronomy as a precision probe of the strong equivalence principle in the relativistic regime.

The spin alignment conjecture was originally formulated in connection with the additivity of coherent information for a class of quantum channels known as platypus channels. Recently, a stronger majorization-based version was proposed by M. A. Alhejji and E. Knill [Commun. Math. Phys. 405, 119, 2024], asserting that the spectrum of the alignment operator is always majorized by that of the perfectly aligned configuration. In this letter, we show that this strong spin alignment conjecture is false in general by constructing an explicit counterexample in the smallest unresolved case, namely three qubits. The example uses two-body states that are not jointly compatible with any single three-qubit global state, which naturally leads to a compatibility-constrained variant of the conjecture.

Quantum theory brings into question the compatibility of the twin desiderata of exact knowability of the present state of the physical world and perfect predictability of its future states. Bohr's coordination-causality complementarity principle transforms this tension into one between properties (as ordinarily understood in classical physics) and deterministic causality. Here, we develop an explicit model of quantum properties which accommodates this essential tension. Our approach integrates operational, reconstructive, and metaphysical standpoints. In particular, we make use of an operational framework employed in a recent operational reconstruction of Feynman's formulation of quantum theory; base our property model on an analysis of property types; and use the notions of actuality and potentiality to frame the model. We show that this quantum property model provides a natural resolution of Zeno's paradox of motion, and provides reliable intuitions about phenomena such as electron diffraction and the non-local behaviour of entangled states of non-identical particles.

The computation of thermal properties of quantum many-body systems is a central challenge in our understanding of quantum mechanics. We introduce the Quantum Finite Temperature Lanczos Method (QFTLM), which extends the finite-temperature Lanczos method to quantum computers by combining real-time quantum Krylov methods with efficient preparation of typical states for trace estimation. This approach enables the computation of thermal expectation values while avoiding the exponential scaling inherent to classical exact simulation techniques. Numerical experiments on the transverse-field Ising model show that QFTLM can reproduce thermal observables over a wide temperature range. We further analyze the influence of Krylov dimension, number of trace-estimator states, and Trotter error, and show that suitable regularization is essential for robustness in noisy settings. These results establish QFTLM as a promising framework for finite-temperature quantum simulation.

Local Operator Entanglement (LOE) has emerged an indicator of quantum chaos in many-body systems. Numerical studies have shown that, in chaotic systems, LOE grows linearly in time and displays a volume-law behavior at late times, scaling proportionally with the number of local degrees of freedom. Despite extensive numerical evidence, complemented by analytical studies in integrable systems, a fully analytical understanding of the emergence of the volume law remains incomplete. In this paper, we contribute toward this goal by deriving a late-time expression for LOE in chaotic systems that exhibits volume-law scaling. Our derivation proceeds by expressing the late-time LOE in the Liouville eigenstate basis and relies on three main assumptions: a higher-order non-resonance condition for the Hamiltonian eigenenergies, the Eigenstate Thermalization Hypothesis (ETH) ansatz for the matrix elements of the initial local operator, and the replacement of Hamiltonian eigenstates with random states in the final expression for LOE. Under these assumptions, we obtain an explicit formula displaying volume-law scaling. Finally, we complement our analytical derivation with numerical simulations of the 1D mixed-field Ising model, testing the resulting formula and exploring the regime of validity of our assumptions.

We present a unified analytical and numerical study of the one-dimensional Feshbach--Villars (FV) equation for spin-0 particles in the presence of several representative external potentials. Starting from the FV formulation of the Klein--Gordon equation, we derive the corresponding one-dimensional master equation and analyse its solutions for Coulomb, power-exponential, Cornell, Pöschl--Teller, and Woods--Saxon interactions. For the singular Coulomb and Cornell cases, a Loudon-type cutoff regularisation is implemented on the full line, allowing a mathematically controlled treatment of the origin and an explicit classification of the states by parity. The Coulomb problem exhibits the expected near-degenerate even--odd structure in the cutoff limit, while the Cornell potential combines short-distance Coulomb behaviour with long-distance confinement and produces a finite set of bound states for fixed parameters. The power-exponential potential with $p=1$ is reduced to a Whittaker-type equation and yields an intrinsically relativistic spectrum with no standard Schrödinger bound-state limit in the parameter regime considered. For the smooth short-range Pöschl--Teller and Woods-Saxon potentials, the FV formalism reveals, respectively, the effects of definite parity and spatial asymmetry on the spectrum, wave functions, and particle--antiparticle mixing. In all cases, we reconstruct the full FV spinor, analyse the associated charge density, and compare the relativistic behaviour with the corresponding non-relativistic expectations whenever such a limit exists. The results provide a coherent set of analytical and numerical benchmarks for relativistic scalar bound states in one dimension.

Efficient entanglement distribution is the foundational challenge in realizing large-scale Quantum Networks. However, state-of-the-art solutions are frequently limited by restrictive operational assumptions, prohibitive computational complexities, and performance metrics that misalign with practical application needs. To overcome these barriers, this paper addresses the entanglement distribution problem by introducing four pivotal advances. First, recognizing that the primary application of quantum communication is the transmission of private information, we derive the Ensemble Capacity (EC), a novel metric that explicitly quantifies the secure classical information enabled by the entanglement distribution. Second, we propose a generalized mathematical formulation that removes legacy structural restrictions in the solution space. Our formulation supports an unconstrained, arbitrary sequencing of entanglement swapping and purification. Third, to efficiently navigate the resulting combinatorial optimization space, we introduce a novel Dynamic Programming (DP)-based hypergraph generation algorithm. Unlike prior methods, our approach avoids artificial fidelity quantization, preserving exact, continuous fidelities while proactively pruning sub-optimal trajectories. Finally, we encapsulate these algorithmic solutions into CODE, a system-level, two-tiered orchestration framework designed to enable near-real-time network responsiveness. Extensive evaluations confirm that our DP-driven architecture yields superior private classical information capacity and significant reductions in computational complexity, successfully meeting the strict sub-second latency thresholds required for dynamic QN operation.

Importance sampling based on quasi-probability decomposition is the backbone of many widely used techniques, such as error mitigation, circuit knitting, and, more generally, virtual quantum resource distillation, as it allows one to simulate operations that are not accessible in a given setting. However, this class of protocols faces a fundamental problem -- it only allows to estimate expectation values. Here, we provide a general framework that lifts any quasi-probability-based protocol from expectation value estimation to a weak simulator, realizing sampling from the desired distribution only using a restricted class of quantum resources. Our method runs with the sampling cost proportional to the negativity of the quasi-probability, in stark contrast to the naive estimation-based approach that requires a large number of samples even in the case of small negativity. We show that our method requires significantly fewer samples in a number of relevant scenarios, such as error mitigation, entanglement distillation and magic state distillation. Our framework realizes the weak simulation of quantum resources without actually distilling the state, introducing a new notion of quantum resource distillation.

Measurement incompatibility and bipartite quantum steering are known to display a strong connection: a set of measurements is incompatible if and only if it can lead to bipartite steering. Despite such a close link between these concepts in bipartite scenarios, little is known in the multipartite setting, where notions of genuine multipartite correlations play major roles. In this work we prove that, as in the bipartite case, incompatibility is also necessary and sufficient for genuine multipartite steering in any multipartite scenario with a single uncharacterised party. Interestingly, genuine multipartite steering can be extracted from any set of incompatible measurements using states which are not SLOCC equivalent, such as GHZ and W states. In contrast, we prove that this result does not hold in scenarios with more than one uncharacterised party, by presenting a set of incompatible measurements that can never lead to genuine multipartite steering in these cases. In order to obtain our main results, we introduce methods tailored for multipartite correlations, paving the way to understanding the role of measurement incompatibility beyond bipartite scenarios.

Resonance fluorescence from a coherently driven two-level emitter is a minimal quantum optical field that combines phase coherence with single-photon-level nonlinearity. Here we show that it can be engineered, using only passive linear interferometry, into energy-time entanglement. By injecting resonance fluorescence from a single quantum dot into an asymmetric Mach--Zehnder interferometer operated near destructive interference of the single-photon component, we generate an output field whose coincidence statistics are dominated by the simultaneous two-photon contribution |2> and the temporally separated photon-pair contribution |11>. In a Franson geometry, these two sectors are resolved on the coincidence-delay axis, and both exhibit high-visibility nonlocal interference fringes and violate the Clauser--Horne--Shimony--Holt Bell inequality. Our results reveal a general route for engineering entanglement from resonance fluorescence using passive optics.

This note starts with a recapitulation of what people call the ``Measurement Problem'' of Quantum Mechanics (QM). The dissipative nature of the quantum-mechanical time-evolution of averages of states over large ensembles of identical isolated systems consisting of matter interacting with the radiation field is discussed and shown to correspond to a stochastic time-evolution of states of individual systems. The importance of dissipation for the successful completion of measurements is highlighted. To conclude, a solution of the ``Measurement Problem'' is sketched in an idealized model of a double-slit experiment.

Sorkin's impossible operations demonstrate that causality of a quantum channel in QFT is an additional constraint on quantum operations above and beyond the locality of the channel. What has not been shown in the literature so far is how much of a constraint it is. Here we answer this question in perhaps the strongest possible terms: the set of causal channels is nowhere dense in the set of local channels. We connect this result to quantum information, showing that the set of causal unitaries has Haar measure $0$ in the set of all unitaries acting on a lattice. Finally, we close with discussion on the implications and connections to recent QFT measurement models.

Motivated by a one-cycle cosmological scenario where the big bang marks the egress from a Euclidean metric signature regime, we investigate the possibility of black holes (BH) hosting the mirroring entryways. Analogous to the inflationary stage following the exit end, the entry portals must be enveloped by de Sitter cores inside BHs, in order to satisfy regularity conditions at the metric signature transition boundary. We examine the interior structure of BHs that could be consistent with such a physical picture, and conclude that the presence of a spacelike shell of non-inflationary matter is likely required.

The ratio between the gravitational waves (GW) and electromagnetic waves (EMW) luminosity distance ratio is a key observable that allows to test the nature of gravity, using gravitational waves emitted from compact binary coalescences. We develop a new non parametric method for constraining the GW-EMW distance ratio, in order to perform model independent analysis of observational data, not based on any specific theoretical of phenomenological assumption. We apply the method to the analysis of binary black hole mergers data from the GWTC-3 catalogue, performing a joint analysis of cosmological and population parameters. The results are consistent with general relativity and with previous analyses based on parametric methods.

The enumeration of black hole entropy in candidate theories of quantum gravity utilises the quantum properties of microstates residing on the black hole horizon. For example, in Loop Quantum Gravity, the computation of entropy is based on the spectrum of area operator, and one determines the possible number of area mirocrostates corresponding to a given classical horizon area. In this paper, we derive the eigenspectrum of the horizon area operator for rotating/non-rotating black holes in a gravitational theory non-minimally coupled to scalar fields. Using the weak isolated horizon formalism, we show that the spectrum of area operator follows unambiguously from the algebra of horizon symmetry. More precisely, from the quantum mechanical point of view, the horizon geometry must be naturally discrete, a conclusion which is arrived at directly, without the need for any particular theory of quantum gravity. The area spectrum depends on the Barbero-Immirzi parameter as well as the value of scalar field on horizon. The area spectrum is equidistant, which is consistent with the Bekenstein-Mukhanov proposal and gives rise to black hole entropy and their quantum corrections.

Quantum state transfer between distant nodes is essential for distributed quantum information processing. Existing protocols are typically optimized for specific coupling regimes, such as adiabatic dark-state transfer in the single-mode limit and pitch-and-catch schemes in the multimode regime, leaving the crossover between them without a simple and unified control strategy. Here we identify a minimal two-parameter control framework that enables high-fidelity quantum state transfer across this single-mode-to-multimode crossover in a multimode quantum channel. Using a pulse-shaped pitch-and-catch protocol controlled only by the pulse ramp rate and the emission-absorption delay, we achieve transfer fidelities exceeding 99.9%, extending pitch-and-catch protocols toward the single-mode limit without requiring dark-state protection or complex pulse design. We further demonstrate robustness against dissipation, disorder, detuning, and imperfect initialization under experimentally realistic conditions. These results provide a simple and broadly applicable framework for state transfer in multimode quantum channels, with relevance to circuit-QED and hybrid quantum-acoustic systems.

We introduce a multilayer superconducting microwave resonator with sub-Ohm impedance optimized for high coupling strength to single electron spins. The design minimizes the magnetic far-field and therefore achieves a Purcell factor $F_P > 10^{15}$. We show several ways to fabricate this type of resonator and present resonators with an intrinsic $Q$-factor exceeding $2 \cdot 10^4$ at the single-photon level. We further characterize these resonators in magnetic fields up to $500 \, \text{mT}$. Finally, we evaluate the impact of the achievable Purcell factor on single-spin detection through photon counting and dispersive readout.

In highly purified host, the coherence of quantum emitters is ultimately limited by hyperfine interactions between the emitter and lattice nuclei possessing non-zero nuclear magnetic moments. This limitation can only be mitigated through isotopic purification. In this work, we investigate CeO2 as a host composed entirely of nuclei with zero nuclear moment. High-quality CeO2 thin films were grown by PLD and doped with Tm and Er ions. Structural characterization using X-ray diffraction, atomic force microscopy, and ion channeling confirms single-crystalline, atomically smooth films with dopants substitutionally incorporated at Ce lattice sites. Photoluminescence lifetime measurements show significantly longer lifetimes for Er-doped CeO2 (2.9 - 5.3 ms) compared with Tm-doped films (14 - 68 μs). Moreover, the Er-doped PLD films exhibit longer lifetimes at ~1% dopant concentration than previously reported for MBE-grown films. Density functional theory calculations reveal a substantial overlap between unoccupied O 2p and Tm 4f states near the valence band maximum, whereas Er 4f states remain well isolated. This electronic interaction likely introduces non-radiative recombination pathways in Tm-doped CeO2, explaining the reduced lifetimes. These findings highlight the importance of selecting appropriate dopant-host combinations and optimized growth conditions to minimize non-radiative channels for quantum applications.

It was once conjectured that two graph states are local unitary (LU) equivalent if and only if they are local Clifford (LC) equivalent. This so-called LU-LC conjecture was disproved in 2007, as a pair of 27-qubit graph states that are LU-equivalent, but not LC-equivalent, was discovered. We prove that this counterexample to the LU-LC conjecture is minimal. In other words, for graph states on up to 26 qubits, the notions of LU-equivalence and LC-equivalence coincide. This result is obtained by studying the structure of 2-local complementation, a special case of the recently introduced r-local complementation, and a generalization of the well-known local complementation. We make use of a connection with triorthogonal codes and Reed-Muller codes.

We investigate the optical and radiative signatures of an accretion disk around a Schwarzschild black hole (BH) immersed in a uniform magnetic field. The spacetime geometry is described by the Schwarzschild-Bertotti-Robinson (SBR) metric, which represents the non-rotating sector of the recently discovered Kerr-Bertotti-Robinson exact solution to the Einstein-Maxwell equations. We begin with the study of null geodesics and demonstrate that the self-consistent magnetic field fundamentally alters photon propagation, causing an expansion of light bundles relative to the Schwarzschild case due to modified initial conditions in the orbital equation. We then compute the magnetic field-dependent shifts of key characteristic radii: the event horizon ($r_h$), photon sphere ($r_{ph}$), and innermost stable circular orbit ($r_{ISCO}$). We find that all three increase monotonically with field strength $B$, revealing a magnetic amplification of the effective gravitational field. For $B=0.05$, we find that the lensed emission bands contract to a narrower impact parameter range, $b\in(4.976,5.149)\cup(5.19,6.128)$. Employing ray-tracing formalism, we construct observed accretion disk images and quantify magnetic modifications, showing that the direct image contracts while maximum energy flux, radiation temperature, and redshift factor are enhanced. Complementing these numerical findings, we develop an analytical framework for the accretion disk dynamics. We derive the modified Keplerian frequency $Ω_K$, along with the specific energy $E$ and angular momentum $L$ for circular orbits. From these, we obtain the exact ISCO radius $r_{\text{ISCO}}$, which shows an outward shift. This outward shift reveals that the radiative efficiency decreases dramatically with increasing magnetic field strength $B$. For $β= BM \sim 0.1$, the efficiency drops by approximately $91\%$.

Exploiting quantum effects for energy storage, quantum batteries (QBs) offer compelling advantages over conventional ones in terms of superior energy density, ultrafast charging, and high conversion efficiency. However, their realization is hampered by decoherence, which causes incomplete charging, rapid self-discharging, and reduced extractable work. Here, we propose a QB architecture based on a chiral magnonic platform. It comprises two yttrium iron garnet (YIG) spheres, one serving as the charger and the other as the QB, coupled to a waveguide. The unique chiral coupling between magnons and the guided electromagnetic fields breaks inversion symmetry, inducing both nonreciprocal energy flow and coherent interference between the charger and QB. Their synergy endows our QB with a 34-fold increase in energy capacity and a 55-fold boost in extractable work compared to its achiral counterpart in an experimentally accessible regime. Our scheme harnesses the decoherence from the electromagnetic fields and turns its destruction into an asset, which enables the robustness and wireless-like remote charging features of the QB. Our analysis reveals that these extraordinary capabilities stem from quantum coherence. By establishing chirality as a useful quantum resource, our work paves a viable path toward the realization of QBs.

In 1973, Bardeen, Cater, and Hawking published "The Four Laws of Black Hole Mechanics", establishing the mathematical framework that would later be understood as the thermodynamics of black holes. Central to the paper is equation (33), which writes the variation of the total energy-momentum integral in terms of physically meaningful quantities: angular momentum, particle number, and entropy. Equation (33) feeds into the differential mass formula, equation (34), which is the first law of black hole mechanics. This note identifies two compensating sign errors in the BCH paper. The first error, demonstrated by a derivation from equation (32), is that equations (33) and (34) should carry minus signs rather than plus signs on the last two integrals, those involving the redshifted chemical potential and the redshifted temperature. he second error is that the definitions of total particle number N and total entropy S given after equation (20) are missing minus signs that are required for these quantities to be positive. These two errors cancel, in that reversing the signs in the definitions of N and S to ensure positive quantities makes equations (33) and (34) correct. All conclusions of the BCH paper remain valid. This note is intended merely as a guide for readers who, in working through the derivation step by step, might otherwise be puzzled by the sign discrepancies. Numbered equations refer to the BCH paper; lettered equations are introduced in this note.

Recently, a fractional version of the Schwarzschild-Tangherlini black hole with a fractal horizon has been introduced. Motivated by the key role of the Schwarzschild solution in gravitational and astrophysical studies, some consequences of this fractional-fractal generalization of the Schwarzschild black hole have been investigated. In this line, the corresponding i) Shapiro and Sagnac time delays, ii) shadow, iii) orbital precession, and iv) gravitational lensing are studied and confronted with observational data. MCMC analysis also unveils i) the potential of this metric in dealing with the solar-system tests and ii) the necessity of studying fractional spacetimes and objects.

In this paper the random channels and their compositions in the space of quantum states are studied. For compositions of i.i.d. random unitary channels, the limit behaviour of probability distributions is described. The sufficient condition for convergence in probability is obtained. The generalized convergence in distribution w.r.t. weak operator topology is obtained. The analysis of transmission of pure and normal states to the set of singular states is done. The dynamics of quantum states is described in terms of the evolution of the values of quadratic forms of operators from the algebra that implements the representation of canonical commutation relations.

We investigate a hybrid photon blockade (HPB) scheme in a driven two-qubit cavity QED system arising from the combination of eigenenergy-level anharmonicity (ELA) and quantum destructive interference (QDI). By tuning the detuning of a single qubit and pumping field, we identify precise parametric regimes that fully integrate the advantages of high brightness in ELA-based conventional photon blockade and strong antibunching in QDI-based unconventional photon blockade. Interestingly, these regimes are accompanied by hyperradiance, indicating that inter-emitter correlations give rise to enhanced collective emission. The HPB mechanism exhibits parametric generality across varying coupling asymmetries and remains accessible via detuning control, offering a feasible route for generating high-quality single-photon source in diverse quantum platforms.

While there are a number of proposed formation channels for subsolar mass compact objects, including black holes formed primordially, or neutron stars that form in collapsar disks, there have yet to be any conclusive observations of such objects. Motivated by the possibility that, if such objects exist, gravitational waves from binary mergers may reveal them, we study binary neutron star mergers where one star has a subsolar-mass in order to determine how well such systems are described by current models, and when they could be distinguished from a system with a subsolar-mass black hole. We perform fully general-relativistic simulations of a $1.7\ M_{\odot}$ star merging with a $0.8\ M_{\odot}$ star, leading to tidal deformabilities of up to $\mathcal{O}(10^4)$ for the latter, and quantify how this affects the merger dynamics and associated gravitation and electromagnetic signals. In this regime, we find mass transfer between the stars, as well as significantly lower disruption frequencies. Though this is not captured by current gravitational waveform models, we conclude that this does not significantly impact the sensitivity of current gravitational wave detectors to these sources. Assuming design sensitivity of the LIGO and Virgo detectors, we find no biases in the recovered intrinsic parameters for signal-to-noise ratios $\lesssim 100$. We also find that the large deformabilities lead to a significant increase in the amount of dynamically ejected matter compared to equal mass systems, exceeding the predictions of current phenomenological models.

In this article, we investigate the influence of the quantum gravity corrections on the horizons, timelike geodesic motions and the gravitational wave emission in two different rotating black hole spacetimes which are constructed via the revised Newman-Janis algorithm from two spherically symmetric loop quantum gravity black holes. The quantum gravity effect is encoded in the regularization parameter $ξ$ of the holonomy correction, and the constraint range of $ξ$ is provided. For the timelike geodesic motion, we find that when the spin parameter $a$ is small, $ξ$ significantly affects the orbital angular momentum $L$. In equatorial periodic orbits, as $ξ$ increases, the allowed energy range for fixed $L$ also increases, while in generic off-equatorial motion, as $ξ$ increases, the permissible range of the Carter constant which effectively confines trajectories toward the equatorial plane decreases. For the gravitational wave emission, by using a simplified extreme-mass-ratio inspiral model within the leading order post-Newtonian approximation, we compute the gravitational waveforms and show how increasing $ξ$ enhances the waveform deviations, particularly near the event horizon. To summarize, the results in this article preliminarily reveal some universal features that holonomy corrections imprint on potentially observable signatures of rotating black holes.

In this work, we investigate the critical behavior of photon rings in the Kerr-Bertotti-Robinson spacetime, describing a rotating black hole immersed in a background magnetic field. We analyze the radial and angular motions of photons under the small magnetic field approximation. Focusing on unstable spherical orbits, we determine three key parameters, $γ$, $δ$, and $τ$, which characterize radial compression, azimuthal advancement, and time delay. We then examine how these parameters depend on the black hole spin, magnetic field strength, and observer inclination for both on-axis and off-axis observers, and we further analyze the properties of higher-order images through near-critical lens equations. The results show that the magnetic field modifies the geodesic structure, and leads to observable changes in the fine structure of photon rings, providing a useful framework for probing magnetized black hole environments.

This review is designed to introduce mathematicians and computational scientists to quantum computing (QC) through the lens of uncertainty quantification (UQ) by presenting a mathematically rigorous and accessible narrative for understanding how noise and intrinsic randomness shape quantum computational outcomes in the language of mathematics. By grounding quantum computation in statistical inference, we highlight how mathematical tools such as probabilistic modeling, stochastic analysis, Bayesian inference, and sensitivity analysis, can directly address error propagation and reliability challenges in today's quantum devices. We also connect these methods to key scientific priorities in the field, including scalable uncertainty-aware algorithms and characterization of correlated errors. The purpose is to narrow the conceptual divide between applied mathematics, scientific computing and quantum information sciences, demonstrating how mathematically rooted UQ methodologies can guide validation, error mitigation, and principled algorithm design for emerging quantum technologies, in order to address challenges and opportunities present in modern-day quantum high performance and fault-tolerant quantum computing paradigms.

Nonstabilizerness, or magic, is a necessary resource for quantum advantage beyond the classically simulatable Clifford framework. Recent works have begun to chart the structure of magic in many-body states, introducing the concepts of long-range magic -- nonstabilizerness that cannot be removed by finite-depth local unitary (FDU) circuits -- and the magic hierarchy, which classifies quantum circuits by alternating layers of Clifford and FDUs. In this work, we construct explicit states that provably possess two-sided long-range magic, a stronger form of magic meaning that they cannot be prepared by a Clifford circuit and a FDU in either order, thus placing them provably outside the first level of the magic hierarchy. Our examples include the ``magical cat" state, $|ψ\rangle \propto |0^n\rangle + |+^n\rangle$, and ground states of certain nonabelian topological orders. These results provide new examples and proof techniques for circuit complexity, and in doing so, reveal the connection between long-range magic, the structure of many-body phases, and the principles of quantum error correction.

The hydrodynamic formulation of quantum mechanics features two velocity fields: a geodesic (classical) velocity $π_μ$ and a stochastic (quantum) velocity $u_μ$. We show that averaging over a stochastic gravitational wave background unifies these into a single complex velocity $η_μ=π_μ-iu_μ$, derived from the logarithmic derivative of a matter amplitude $\mathcal{K}$. This object lives as a section of the pullback bundle $π_{2}^{*}(T^{*}M)$ over configuration space and defines a flat $U(1)$ connection, satisfying $D_μ\mathcal{K}=0$. Crucially, $η_μ$ acts as a fundamental information-geometric carrier, where $u_μ$ maps the variance of metric fluctuations $\langle h_{μν}h_{αβ}\rangle$ to the Fisher metric and von Neumann entropy. The resulting geometric structure collapses into an elegant complex geodesic equation $η^ν\nabla_νη_μ=\nabla_μ(\frac{1}{2}η^νη_ν)$, while non-trivial spacetime topology imposes a holonomy quantization condition. This topological phase suggests observable signatures in atom interferometry and cosmological correlations, providing an experimental window into the stochastic nature of spacetime at the Planck scale.

Connected and automated vehicles require city-scale coordination under strict latency and reliability constraints. However, many existing approaches optimize communication and mobility separately, which can degrade performance during network outages and under compute contention. This paper presents QIVNOM, a quantum-inspired framework that jointly optimizes vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) communication together with urban traffic control on classical edge--cloud hardware, without requiring a quantum processor. QIVNOM encodes candidate routing--signal plans as probabilistic superpositions and updates them using sphere-projected gradients with annealed sampling to minimize a regularized objective. An entanglement-style regularizer couples networking and mobility decisions, while Tchebycheff multi-objective scalarization with feasibility projection enforces constraints on latency and reliability. The proposed framework is evaluated in METR-LA--calibrated SUMO--OMNeT++/Veins simulations over a $5\times5$~km urban map with IEEE 802.11p and 5G NR sidelink. Results show that QIVNOM reduces mean end-to-end latency to 57.3~ms, approximately $20\%$ lower than the best baseline. Under incident conditions, latency decreases from 79~ms to 62~ms ($-21.5\%$), while under roadside unit (RSU) outages, it decreases from 86~ms to 67~ms ($-22.1\%$). Packet delivery reaches $96.7\%$ (an improvement of $+2.3$ percentage points), and reliability remains $96.7\%$ overall, including $96.8\%$ under RSU outages versus $94.1\%$ for the baseline. In corridor-closure scenarios, travel performance also improves, with average travel time reduced to 12.8~min and congestion lowered to $33\%$, compared with 14.5~min and $37\%$ for the baseline.

Fermionic linear optics (FLO) with Gaussian resources is efficiently classically simulable. We show that this is no longer the case for such quantum circuits for fermions with internal degrees of freedom, equipped with mid-circuit number monitoring and classical feedforward. In our architecture, the measurement record routes the selected blocks into a fixed-order Bell-fusion pairing geometry. On the level of classical description, this implies realizing a situation in which the permutation sum no longer collapses to a single determinant or Pfaffian. Each post-selected branch expands as a signed sum of path-ordered products of typically non-commuting dressed blocks, and branch amplitudes are matrix elements of the resulting non-commutative trace polynomials. Numerically, we observe Porter-Thomas statistics as the output distribution and a rapid growth of the minimal order-respecting matrix product operator bond dimension. These results thus establish mid-circuit measurement-induced non-commutativity as a route to sampling hardness for noninteracting fermions under reasonable complexity assumptions, without introducing coherent two-body interactions into the FLO evolution.

We compute the one-loop contribution to the semiclassical partition function of near-extremal, asymptotically AdS black holes in five-dimensional Einstein-Gauss-Bonnet gravity. In the absence of an exact analytic rotating solution at finite Gauss-Bonnet coupling $α$, we restrict to static, charged configurations and evaluate the contribution to $Z_{\text{1-loop}}$ arising from tensor, vector, and $U(1)$ gauge fluctuations. The analysis is based on the spectrum of a generalized Lichnerowicz operator governing linearized perturbations on the near-horizon geometry of the extremal solution, including its deformation by the coupling $α$. In the canonical ensemble, the low-temperature behavior of the one-loop partition function leads to logarithmic corrections to the entropy of the form $\log(T/T_0)$, where the scale $T_0$ depends on both the fluctuation sector and the Gauss-Bonnet coupling. These corrections are controlled by the structure of zero modes of the deformed operator and their splitting at small but finite temperature. Our explicit computation yields a universal low-temperature scaling $Z_{\text{1-loop}}\sim 5 \log T$, where the coefficient arises from the combined contributions of tensor, vector, and $U(1)$ gauge modes, reflecting the corresponding counting of zero modes in each sector.

This work investigates curvature-induced modifications to the Yukawa potential in static, spherically symmetric spacetimes described by Tolman metrics, focusing on their implications for compact stellar objects, with particular application to solutions IV and VI. Motivated by the interplay of quantum interactions and strong gravitational fields in systems like neutron stars, we derive explicit corrections to the Yukawa potential for these metrics. Contrary to previous findings suggesting that curvature corrections break the radial symmetry of the interacting potential near a highly charged black hole, we demonstrate that Tolman metric corrections preserve this symmetry in the local inertial frame. Numerical estimates for astrophysical objects reveal energy shifts of the order of $10^{-34}$ MeV for solution IV. The Tolman VI solution, while singular at the center, yields comparable corrections for most of the fluid sphere radius. A detailed analysis of the repulsive or attractive nature of these curvature corrections for a local observer is provided for each scenario. These results highlight the role of spacetime geometry in shaping quantum interactions and provide a foundation for future studies of nuclear interactions within the context of relativistic stars.

Collapse-locality is an untested loophole in the violation of Bell's inequalities. The core of the argument is that the time value of photon detection is delayed by the time Tc required by the collapse of its quantum state. The value of Tc is given by the underlying theory of quantum collapse, and is mostly unknown. Depending on the value of Tc, detections in the performed Bell's experiments may have not been truly space-like separated events. This implies that the inequalities may have been violated as a consequence of (conspiratorial) information propagating at subluminal speed. We report an optical Bell experiment which closes the weaker ('essential') form of this loophole regardless the theory of quantum collapse. This is possible thanks to unique features of the setup. These features are: classical signals sent to the stations to define a time reference, and variable distance between the stations leaving all other parameters constant.

We present profile-likelihood constraints on velocity-independent dark matter-proton scattering, including cases in which only a fraction of dark matter has such non-gravitational interactions. Frequentist profile-likelihood techniques provide prior-independent constraints, circumventing prior-volume effects that we show arise in Bayesian constraints on this model. In the limit where the scattering cross section or the fraction of interacting dark matter approaches zero, the other interacting dark matter model parameters become unconstrained, causing the posterior distribution to favor that region of parameter space. Using Planck 2018 cosmic microwave background anisotropy data, we find a clear impact of prior-volume effects on the posteriors used to place constraints on dark matter scattering. Compared to the frequentist analysis, the Bayesian method consistently overestimates the constraints on the cross section. Given the potentially biased upper limits on models subject to prior-volume effects, such as this one, we recommend supplementing Bayesian constraints with frequentist statistics to better assess the impact of priors.

We present a bottom-up holographic description of the QCD $θ$-vacuum and the $U(1)_A$ anomaly in five dimensions. The multi-branched $θ$-vacuum structure emerges geometrically from a higher-dimensional gauge field, while the axial anomaly is realized through a Stückelberg coupling that is dual to a Chern-Simons term. In this framework, the $η'$ meson appears as a zero mode of bulk fluctuations, and its mass arises from the anomaly-induced Stückelberg term. The construction provides a transparent holographic derivation of the anomaly contribution to the $η'$ mass and naturally reproduces the Witten-Veneziano relation between the $η'$ mass and the Yang-Mills topological susceptibility.

We propose a nonlocal realization of the Stueckelberg portal between the Standard Model and Dark Sector, which decouple in the local limit. This implies that the mediator, $U(1)_{\rm D}$ Dark Photon $A'$ with a Stueckelberg mass, interacts nonlocally with the Standard Model quarks and leptons. We study phenomenological implications of this scenario for the meson decays into semi-invisible and invisible channels. We discuss the experimental limitations on the model parameters, including the nonlocality scale.

Standard Model predictions for Majorana neutrino transition magnetic moments (TMMs) are subject to severe chiral and GIM-like suppressions, rendering them vanishingly small. To dynamically generate a macroscopic TMM, we propose a dark sector framework featuring a $U(1)_D$ gauge symmetry, a vector-like lepton doublet, and two complex dark scalars. We demonstrate that while fermion-radiated loop amplitudes identically cancel due to Majorana self-conjugacy, a chirally enhanced dark TMM is successfully generated exclusively through scalar-radiated loops. This mechanism safely shifts the required chirality flip onto the heavy internal fermion line and utilizes a misaligned double-scalar mixing in flavor space to evade the Majorana antisymmetry prohibition. We systematically confront this tensor portal framework with multi-frontier experimental constraints. Since the dark TMM generation is inextricably linked to charged lepton flavor violation, the internal Yukawa couplings are stringently capped by the latest $μ\to e γ$ limits from MEG II. Concurrently, the visible-dark kinetic mixing portal is heavily bottlenecked by missing energy and mono-photon searches at NA64 and BaBar. Our global phenomenological analysis reveals that the synergistic theoretical upper bound dictated by these indirect high-energy probes completely eclipses the direct scattering constraints from Borexino. This establishes a strict phenomenological hierarchy: high-intensity cLFV probes and accelerator-based dark sector searches jointly possess the overwhelmingly dominant exclusionary power over direct solar neutrino limits for such microscopic magnetic moment models.

We study the coaction of cosmological wavefunction coefficients of conformally coupled scalars in FRW background of a two-site example, which turns out to have an elegant diagrammatic interpretation. We show how the coaction acts on the twisted integrals for wavefunction coefficients, decomposing them into contributions associated with subtopologies and cuts, with the subtopologies admitting an interpretation as time-ordered integrals. This provides a clear interpretation of their analytic structure and suggests a broader applicability to more general cosmological diagrams.

As a unique probe for geophysical research, geoneutrinos can reveal the distribution of internal heat sources in the Earth by detecting electron antineutrinos produced by the radioactive decay of $^{238}$U, $^{232}$Th, and $^{40}$K. However, commercial nuclear power plants continuously produce the same type of electron antineutrinos, which constitute a primary background difficult to eliminate in geoneutrino experiments. As geoneutrino measurements and reactor background modeling approach sub-percent precision, even small matter-induced corrections to reactor antineutrino propagation require quantitative assessment. In this paper, we develop a high-precision prediction framework for reactor neutrino fluxes at underground labs, using global reactor operating data, reactor-to-detector distances, and matter effects (MSW) on neutrino propagation through the Earth. To solve the three-flavor MSW evolution efficiently, we implement a second-order Strang-splitting solver in the vacuum mass basis. Within this framework, we have calculated the reactor neutrino oscillation probabilities, including the MSW effect under one-dimensional (spherically symmetric) and three-dimensional (including lateral inhomogeneities) Earth models, and compared them with the vacuum oscillation scenario, to assess the impact of Earth's structural features on the accuracy of reactor neutrino flux predictions.

An amplitude analysis of the singly Cabibbo-suppressed decay $D^0 \to K^+ K^- π^0 π^0$ is performed, for the first time, to determine the relative magnitudes and phases of different intermediate processes. The analysis uses $e^+e^-$ collision data collected with the BESIII detector at the center-of-mass energy 3.773~GeV corresponding to an integrated luminosity of 20.3 $\rm fb^{-1}$. The absolute branching fraction of $D^0 \to K^+ K^- π^0 π^0$ is measured to be \BF. The dominant intermediate process is $D^0 \to K^{*}(892)^+K^{*}(892)^-$, with a branching fraction of $(2.79 \pm 0.13_{\rm{stat.}} \pm 0.11_{\rm{syst.}}) \times 10^{-3}$. Amplitude analysis reveals that the $D^0 \to K^{*}(892)^+K^{*}(892)^-$ decay is S-wave dominant. The longitudinal polarization fraction of $D^0 \to K^{*}(892)^+ K^{*}(892)^-$ is measured to be $0.468\pm0.046_{\rm{stat.}}\pm0.011_{\rm{syst.}}$.

Type IIB string theory on AdS$_3 \times S^3\times T^4$ with RR flux as the near-horizon limit of the D1-D5 solution is expected to be dual to a (4,4) supersymmetric 2d CFT parametrized by the integers $Q_1,Q_5$ and other moduli. It is related by T-duality to type IIA string theory in the near-horizon limit of the D2-D4 solution which admits an uplift to the 11d AdS$_3 \times S^3\times T^5$ background which is the near-horizon limit of the M2-M5 solution. We point out that this relation allows one to use the quantum M2-brane description to probe ``non-planar'' corrections in the dual 2d CFT, in close analogy with the ABJM theory case (described by M-theory on AdS$_4 \times S^7/\mathbb{Z}_k$). We consider an analog of a supersymmetric Wilson loop (line defect) expectation value represented by type IIA string partition function expanded around AdS$_2\subset $AdS$_3$ minimal surface. Its M-theory analog is the M2 brane partition function expanded near AdS$_2\times S^1$. We compute the 1-loop contribution $Z_1$ to the M2 brane partition function and find that in contrast to the ABJM case in arXiv:2303.15207 (where $Z_1= (2\sin{\frac{2π}{ k}})^{-1} = \frac{k}{ 4 π} +\fracπ{ 6k} +...$ contains an infinite series of higher genera string corrections, $k^{-1} \sim \frac{g_s}{ \sqrt {\rm T}}$), here it is given solely by the leading string-theory contribution $Z_1= \fracκ{ \sqrt{2π}}$ where $κ\sim \sqrt{Q_5}$ plays a role analogous to $k$. We also discuss a generalization to the mixed flux case which is straightforward from the 11d perspective.

We consider the hypothetical possibility of neutral gauge bosons~($Z^\prime$) with flavor-specific leptophilic couplings. For such {\it New Physics}~(NP) interactions, the current experimental constraints are much relaxed in the heavy mass regime, particularly for masses $\geq \mathcal{O}(1)$ TeV. However, in the presence of a leptophilic $Z^\prime$, leptonic decay modes of the electroweak gauge bosons and Higgs can be corrected at the loop level. Using the existing upper bounds on the corresponding decay widths, we find that one can impose stronger exclusion limits on the interactions of a heavy $Z^\prime$. Future updates on the aforesaid decay channels can be used in complementarity with the proposed lepton colliders to probe even weaker leptophilic NP interactions at the TeV scale and beyond.

Within a neural-network-based holographic framework, we incorporate lattice QCD (LQCD) and Hadron Resonance Gas (HRG) data to train the model and predict the location of the QCD critical endpoint (CEP). The training dataset consists of the entropy density, baryon number susceptibility, and baryon density. The metric warp factor $A(z)$ and the gauge kinetic function $f(z)$ are parameterized by neural networks and determined through the training procedure. The resulting model reproduces the equation of state at vanishing chemical potential in good agreement with both LQCD and HRG data. Extending the analysis to finite chemical potential, we solve the equations of motion and obtain thermodynamic observables consistent with LQCD results at finite density. After incorporating the HRG constraints, the predicted position of the CEP shifts toward larger chemical potentials compared to recent studies. We further employ symbolic regression to derive analytic expressions for $A(z)$ and $f(z)$, providing convenient functional forms for future phenomenological applications. Finally, we perform a data-driven validation using synthetic thermodynamic data generated from an existing analytical holographic model. The neural-network framework reproduces the corresponding CEP location with good accuracy, showing close agreement within numerical uncertainties.

Quantum gravity may modify the fundamental symmetries that govern identical particles. In particular, noncommutative spacetime frameworks predict deformations of Bose and Fermi statistics. Here we develop a relativistic quantum field theory based on the most general oscillator algebra compatible with $θ$-deformed Poincaré symmetry. This construction generalizes twisted statistics to a class of quon-like deformations allowing non-involutive particle exchange. We show that the resulting theory is consistent at both the free and interacting levels and derive its implications for atomic systems. Purely twisted statistics predicts Pauli-forbidden atomic transitions at rates incompatible with experiments. By contrast, a class of quon deformations suppresses such processes by powers of the noncommutativity scale, but only if superselection rules between permutation-symmetry sectors are violated. This implies an effective breakdown of particle indistinguishability and provides theoretical motivation for high-precision experimental tests of the Pauli exclusion principle.

The twist-3 transverse spin--dependent nucleon structure function $g_2$ arises in high-energy processes involving a transversely polarized nucleon. Its connection to quark--gluon correlations allows for an interpretation in terms of the average transverse color Lorentz force acting on unpolarized quarks inside a transversely polarized nucleon. In this work, we investigate this force using light-front wave functions obtained by diagonalizing the light-front Hamiltonian with quantum chromodynamics inputs within the Basis Light-front Quantization approach. We evolve our results to a common scale of $5~\mathrm{GeV}^2$ and present the corresponding form factors in momentum space as well as the transverse force components in impact-parameter space. These distributions provide a complementary perspective on the Sivers asymmetry in transversely polarized deep-inelastic scattering. In the forward limit, we extract the twist-3 reduced matrix element $d_2$, and our results are found to be comparable with those from other theoretical calculations and experimental determinations.

We propose a method to accurately calculate the momentum distributions and the phase distributions of the probability amplitude for both boson and fermion pair production in a spatially homogeneous and time-dependent electric field. Applying this method to multiphoton pair production in a circularly polarized electric field rotating around the $z$-axis, we clarify that the topological charge appearing in the phase distribution of the probability amplitude for pair production reflects the orbital angular momentum (OAM) of the produced pairs rather than that of individual particles. On this basis, we demonstrate that, within the semiclassical framework, the $z$-component of the total angular momentum of the field and the particles is conserved, whereas the conservation of total angular momentum cannot be verified. The results also reveal that the pair production is also constrained by $C$-parity conservation, and that pairs with smaller OAM are produced more favorably. These findings provide deeper insight into angular momentum transfer in pair production.

We study wavepackets of exotic excitations after two-dimensional fermions are scattered by the boundary condition constructed by Maldacena and Ludwig, which turns elementary excitations into exotic fractionally-charged objects. They are of interest in the s-wave approximation of the fermion-monopole scattering in four-dimensional QED and of the multi-channel Kondo effect. We in particular give an explicit expression of the outgoing state of a pair of such particles; we then examine its properties, such as the charge density $\langle J(x)\rangle$ and the expectation value $\langle N\rangle$ of the number of fermions and anti-fermions in the state. The charge density $\langle J(x)\rangle$ is found to be localized with its integral finite and fractional, while the expectation value $\langle N\rangle$ diverges when the wavepacket is localized to a point.

Landau analysis in momentum twistor space can be formulated as the study of varieties of lines in three-dimensional projective space, together with their projections and discriminants. Within this framework, we define enumerative invariants (LS degrees) that count leading singularities. Leading Landau singularities (LS discriminants) arise as discriminants detecting the collision of leading singularities. We uncover a recursive mechanism underlying Landau singularities, governed by substitution maps between Grassmannians. Applying this framework, we prove positivity and factorization into cluster variables for the LS discriminant of a large class of Landau diagrams at arbitrary loop order. This provides a first-principles explanation for the emergence of positivity and cluster algebra structures in the singularities of planar N=4 super Yang-Mills theory.

The magnetospheres of magnetars, a class of highly magnetized neutron stars, host magnetic fields exceeding the Schwinger limit, where Quantum Electrodynamics (QED) becomes nonlinear. In such environments, QED scattering processes are strongly modified, which may affect plasma dynamics. In this work, we apply a formalism originally developed for the study of magnetic-field effects in hot quark-gluon plasma to strong-field QED. The method resums interactions between virtual electrons and the external magnetic field, consistently incorporating the finite decay widths of excited Landau levels derived from the fermion self-energy. Using this framework, we perform the first systematic analysis of tree-level QED scattering processes in strong magnetic fields, concentrating on the processes of highest relevance for the plasma dynamics of magnetars. All resulting cross sections are provided in an open-source Python package.

In the context of the Scintillating Fibre (SciFi) Tracker for the LHCb Upgrade 2, radiation-induced damage in silicon photomultipliers (SiPMs) has been studied over a wide temperature range, from room temperature down to 100 K. With the LHCb detector Upgrade 1, installed during the LHC's Long Shutdown 2 (LS2) (2019-2021), the first large-scale SciFi tracker read out by multichannel silicon photomultipliers (SiPMs) was constructed, installed, and has been operated ever since. A major challenge for the SciFi tracker is the neutron radiation at the SiPMs' location. At the end of the lifetime of the Upgrade 1 detector, the expected neutron fluence for the SiPMs will reach 6e11 neq/cm^2. Cryogenic operation is being investigated to mitigate even higher radiation-induced damage for Upgrade 2, where the total neutron fluence is expected to reach 3e12 neq/cm^2. A large set of custom SiPM arrays, varying in pixel size, electric field configuration, and doping implant concentration, developed by FBK and Hamamatsu, were tested after neutron irradiation. Characterisation was performed in a dedicated cryogenic test setup, where the operating temperature was varied over a wide range. Key performance parameters such as breakdown voltage, gain, dark count rate, optical crosstalk, and afterpulsing were characterised as functions of temperature, overvoltage, and neutron fluence. The result is a precise assessment of radiation damage for state-of-the-art technology from two leading SiPM manufacturers, allowing the results to be transferred to other SiPM applications.

The astrophysical $S$-factor for the proton-proton fusion is calculated in the low-energy regime for a variety of nuclear interactions and consistent nuclear currents, derived within chiral effective field theory. We estimate, for the first time, the theoretical uncertainty on the $S$-factor due to the truncation of the chiral expansion of the currents using a Bayesian analysis. In order to reach an accuracy at the percent level in the calculation, the electromagnetic potential includes contributions beyond the leading Coulomb interaction, such as two-photon exchange and vacuum polarization. The initial proton-proton state is expanded in partial waves and only the ${}^1S_0$ contribution is included, as it is known that the other partial-waves effects are negligible. The low-energy constant entering the contact term in the weak axial current operator is calibrated to reproduce the Gamow-Teller matrix element in Tritium $β$-decay. The value $S(0)$ is found to be $S(0)=(4.068 \pm 0.025)\times 10^{-25} \: \text{MeV}\: \text{b}$.

A measurement is presented of dijet angular distributions in proton-proton collisions at $\sqrt{s}$ = 13 TeV, using data collected with the CMS detector at the CERN LHC and corresponding to an integrated luminosity of 138 fb$^{-1}$. For the first time, the dijet angular distributions, corrected for detector effects, are compared with the predictions of perturbative quantum chromodynamics at next-to-next-to-leading order, including next-to-leading-order electroweak corrections. While data are generally found to be in agreement with predictions, a small difference in shape of the normalized distributions is seen for dijet masses ranging from 2.4 to 4.8 TeV and above 6 TeV. The distributions are used to search for proposed signatures of quark compositeness, extra spatial dimensions, quantum black holes, dark-matter mediators, axion-like particles, and anomalous gluon couplings. The most stringent limits to date are set for most of these scenarios. Quark contact interactions are excluded at 95% confidence level (CL) up to a scale of 17 (37) TeV for destructive (constructive) interference in a benchmark scenario, valid to next-to-leading order in quantum chromodynamics, and in which only left-handed quarks participate. The coupling of axion-like particles to the gluon, $c_{\text{g}}/f_{\text{a}}$, is constrained to be lower than 0.42 TeV$^{-1}$ at 95% CL. The anomalous triple-gluon coupling, $C_{\text{G}}/Λ^2$, in a standard model effective field theory is constrained to be lower than 0.0076 TeV$^{-2}$ at 95% CL.

Momentum twistors for scattering amplitudes in particle physics are lines in three-space. We develop Landau analysis for Feynman integrals in this setting. The resulting discriminants and resultants are identified with Hurwitz and Chow forms of incidence varieties in products of Grassmannians. We study their degrees and factorizations, and the kinematic regimes in which the fibers of the Landau map are rational or real. Identifying this map with the amplituhedron map on positroid varieties, and the associated recursions with promotion maps, yields a geometric mechanism for the emergence of positivity and cluster structures in planar N=4 super Yang-Mills theory.

Within the framework of the left-right symmetric model (LRSM) and under the assumption of a diagonal Dirac neutrino mass matrix $M_{\rm D}$, this paper systematically investigates 20 types of three-zero textures in the Majorana neutrino mass matrix $M_{\rm R}$. The study reveals that only five three-zero textures of $M_{\rm R}$ satisfy the constraints of the latest NuFit 6.0 global fit results. Furthermore, we phenomenologically explore the correlations between Majorana phases $ρ$ and $σ$, as well as the relationships between the heavy neutrino mass spectrum $M_{I}~(I=1,2,3)$, ratio of Dirac matrix elements $y_{2}$, $y_{3}$ and the scale factor $r$. The results indicate strong correlations among the model parameters. In particular, the allowed regions for the Majorana CP phases are significantly restricted and depend on the specific texture of $M_{\rm R}$. On this basis, leptogenesis originating from heavy right-handed neutrino decays is investigated. Numerical results demonstrate that the $M_{\nu3}$ pattern can achieve successful leptogenesis within specific $r$ intervals for both the normal ordering (NO) and the inverted ordering (IO) of the light neutrino masses, while the $M_{\nu4}$ and $M_{\nu5}$ patterns possess viable parameter space for successful leptogenesis only in the NO case.

We report the results of the second iteration of an experiment searching for neutron-hidden-neutron oscillations in a beam of ultracold neutrons, conducted at the PF2 facility of the Institut Laue Langevin (ILL). Oscillations were tested via neutron disappearance as a function of an applied magnetic field, in the context of a phenomenological two-parameter model assuming zero hidden potentials. The magnetic field was varied in a step-wise manner in order to resonantly enhance the oscillation probability at different mass splittings ($δm$) across a 60--1550 peV range. No evidence for neutron disappearance is observed and conservative limits on the neutron-hidden-neutron oscillation period ($τ_{nn'}$) have been set at 95 % confidence level: $τ_{nn'} > 200$ms for $|δm| \in [60, 400]$ peV and $τ_{nn'} > 100$ ms for $|δm| \in [400, 1550]$ peV

Continuing our previous study of Deeply Virtual Meson Production (DVMP) at twist-3 accuracy, we derive compact expressions for all helicity amplitudes. We perform a phenomenological analysis of the helicity-amplitude ratio $\mathcal{A}^{11}/\mathcal{A}^{00}$ and of the spin-density matrix element $r_{00}^{04}$ within the Color Glass Condensate framework. Small-$x$ evolution is incorporated by numerically solving the running-coupling-and-collinearly-improved Balitsky-Kovchegov and Balitsky-Fadin-Kuraev-Lipatov equations with the McLerran-Venugopalan model as the initial condition. By capturing a relevant subset of next-to-leading order corrections, we provide the most theoretically accurate description of these observables to date. Our results are compared to HERA data, and predictions are presented for electron-lead collisions at the future Electron-Ion Collider. We discuss the impact of non-linear effects at low photon virtuality and the role of genuine higher-twist contributions associated with light vector meson distribution amplitudes, corresponding to higher-Fock-state components of the projectile.

We develop a theoretical framework for quantum field description of neutrino evolution in a medium, with a focus on quantum decoherence induced by neutrino scattering on fermions. By deriving a generalized Lindblad master equation that accounts for neutrino momentum-changing transitions, we go beyond the standard treatment that assumes fixed neutrino momentum. Our model explicitly connects decoherence parameters to scattering cross sections, offering a bridge between theoretical predictions and experimental constraints on neutrino quantum decoherence. We apply the developed formalism to several scenarios: 1) neutrino scattering on electrons, 2) neutrino scattering on protons and neutrons through non-standard interactions (NSI), 3) neutrino scattering on dark matter fermions. The obtained results demonstrate that decoherence effects can significantly alter neutrino oscillation patterns and provide a new probe for physics beyond the Standard Model.

Applying a rule-based holographic method, we investigate the reconstruction of dual gravity theories from the quantum field theory (QFT) data, specifically entanglement entropy. We first derive a three-dimensional black hole geometry from the entanglement entropy of a two-dimensional thermal system. Using the reconstructed solution, we extract various thermodynamic quantities with small numerical errors. Moreover, we explore how to reconstruct the dual gravity theory beyond the geometry itself. For an undeformed conformal field theory (CFT), we show that the dual gravity theory can be constructed analytically from the analytic form of the entanglement entropy. In particular, we demonstrate how to reconstruct the analytic dual geometry by applying the Abel transformation. Finally, we investigate the numerical reconstruction of the dual gravity theory from numerical entanglement entropy data for a relevantly deformed CFT. After reconstructing the dual gravity, we show that additional information about the renormalization group (RG) flow, for instance, the $\b$-function and the $c$-function, can be extracted for the considered relevantly deformed CFT.

We demonstrate that the pushforward of the product of Haar measures by the lattice Yang-Mills action concentrates as a Gaussian. It is also sketched how, using this fact, one can recover the strong-coupling expansion.

In the present work, the form factors of $B_{(s)}$ to light P-wave axial vector mesons are calculated via the light cone sum rules (LCSR) in the framework of heavy quark effective field theory (HQEFT). Firstly, the expressions of form factors in terms of the light cone distribution amplitudes (DAs) of axial vector mesons are derived via the LCSR at the leading order of heavy quark expansion. It is found that the penguin type form factors can be obtained directly from the corresponding semileptonic ones, which is similar to the case of S-wave mesons. Considering the light axial vector meson DAs to twist-3, we give the numerical results of form factors systematically. As applications, we investigate the branching ratios, longitudinal polarization fractions and forward-backward asymmetries of relevant semileptonic decays induced by charged current. Our results may be tested by more precise experiments in the future.

The scientific potential of X-ray polarimetry has long been recognized, but the challenges in measuring polarization have left it largely unexplored, particularly in the hard X-ray regime. While tremendous advancement has been made in soft X-ray polarimetery, the lack of sensitive hard X-ray polarimeters and polarisation measurements continues to limit our understanding of high-energy astrophysical processes. With the development of hard X-ray mirrors, it is now possible to develop a sensitive focal plane hard X-ray polarimeter. One such effort is CXPOL, a prototype developed at PRL, India, which consists of a plastic scintillator as active scatterer readout by PMT surrounded by CsI(Tl) scintillators in cylindrical array with SiPM readout from one side. First results of the prototype have been demonstrated in 20 to 80 keV energy range. The sensitivity of the instrument can be significantly enhanced using faster and better light yield scintillator like NaI as absorbers. Further, the use of a position-sensitive scatterer and absorbers, can also provide spectroscopic information by measuring the interaction position along the length and from the known energy depositions in the detectors. Position sensitive detectors are also helpful in mitigating the systematic effects introduced by the off-axis events in the polarisation measurements. Here, we demonstrate the detection sensitivity in the 100x20x5 mm^3 NaI(Tl) scintillator absorber readout on both ends by SiPM arrays operating in co-incidence. In this work, we characterize the first prototype of this detector system and investigate the variation in energy and position resolution, and light output with irradiation position along the length of the detector. The two end readout in co-incidence also reduces the overall SiPM background per absorber by an order of magnitude, further enhancing the polarimetric sensitivity of the instrument.

We present an analysis of the light tensor ($J^{PC}=2^{++}$) hybrid meson mass and coupling from QCD Laplace Sum Rules where the next-to-leading order (NLO) perturbative (PT) corrections and the contributions of the non-perturbative (NP) condensates up to dimension-six ($D=6$) are included. NLO leading-logarithms corrections due to the condensates which contribute in the chiral limit are considered. We obtain the mass $M_{2^+}= (2038\pm 190)$ MeV and a relatively small coupling $f_{2^+}=(10.5\pm 2.9)$ MeV normalized as $f_π=93$ MeV. Our results suggest that the $f_2(1950)$ or/and the $f'_2(2010)$ may have a sizeable $\bar qqg$ hybrid component. We also compute the tensor hybrid topological charge (value of the two-point function at zero momentum) and find (for the first time) at NLO: $Π_{qg}(0)=(2.41\pm 0.43) \times 10^{-4}{\rm GeV}^6$ which could be checked from some lattice QCD or/and low energy theorems (LET).

The exploration of exotic hadrons provides a crucial testing ground for quantum chromodynamics in its non-perturbative regime. In this work, we perform a systematic study of the two-body strong decay properties of single-charm molecular pentaquarks in the $Y_c\bar{K}^{(*)}$ systems, where $Y_c = Λ_c$, $Σ_c$, $Ξ_c$, and $Ξ_c'$. Employing an effective Lagrangian approach combined with hadronic molecular wave functions derived from the one-boson-exchange model, we compute the decay widths and branching ratios for a series of predicted states with strangeness $|S| = 1$ and $|S| = 2$. Our calculations reveal distinctive decay patterns that serve as fingerprints for molecular identification. The total decay widths vary dramatically, from less than 1 MeV for the narrow $Σ_c\bar{K}$ $(I(J^P)=1/2(1/2^-))$ state to several tens of MeV for broader coupled-channel molecules like $Λ_c\bar{K}^*/Σ_c\bar{K}^*$. A key finding is the stability of the predicted branching ratios against variations in the binding energy. The decay dynamics are dominated by light meson (particularly pion) exchange, leading to a strong preference for final states containing a charmed baryon and a strange meson. Furthermore, coupled-channel effects and isospin-related interference play essential roles in both the formation and decay mechanisms of specific candidates. The results provide concrete, testable predictions for future experimental searches at facilities such as LHCb and Belle II.

Understanding near-threshold strong interactions is essential for disentangling hadronic molecules and compact multiquark states in heavy-flavor spectroscopy. In this context, the doubly charmed tetraquark candidate $T_{cc}(3875)^+$ serves as a critical benchmark because it lies very close to the $D^*$-$D$ thresholds. Motivated by the interaction ambiguity reported recently [\href{https://doi.org/10.1103/kd4s-9rzr}{Phys.Rev.D 113, L031505 (2026)}], we evaluate the $D^*$-$D$ scattering lengths and femtoscopic correlation functions for the molecular and molecule-compact admixture assignments of the $T_{cc}(3875)^+$. We show that, although these scenarios yield similar invariant-mass line shapes, their corresponding femtoscopic correlation functions differ markedly and remain clearly distinguishable for typical particle-emitting sources created at the LHC. Our results indicate that femtoscopy can serve as a sensitive and complementary probe of the near-threshold dynamics of $T_{cc}(3875)^+$, providing vital theoretical references for future LHC femtoscopy measurements.

In heavy-ion collisions, an excess in photon production, together with a larger than expected positive elliptic flow, has been observed, a phenomenon commonly referred to as the direct photon puzzle. In this work we study the mechanism of photon production arising from gluon splitting and fusion during the pre-equilibrium stage in the presence of magnetic fields in peripheral heavy-ion collisions. We begin by analyzing the general tensor structure of the two-gluon one-photon vertex, computing it at the one-loop level for magnetic fields of arbitrary strength without resorting to additional approximations. Using these expressions, we calculate the contribution of gluon fusion and splitting to the photon yield, revealing that splitting dominates over fusion at low photon energies. Our results are compared with experimental data from the PHENIX collaboration. Finally, we incorporate a longitudinal anisotropy into the initial gluon distribution and find that it does not significantly alter the photon yield compared to an isotropic distribution.

By using Wald formalism, we show that the identification between the bulk and boundary conserved quantities induced by the perturbation of generic non-electromagnetic matter field holds not only on top of the asymptotically flat stationary spacetimes but also on top of the asymptotically AdS stationary ones. We further show that such an identification reduces to the familiar form for the test point particle by viewing it as the limiting case of general matter.

We introduce a minimal two-parameter formulation of the dark energy (DE) density evolution normalized to its present-day value, $f_{\rm DE}(z) \equiv ρ_{\rm DE}(z)/ρ_{\rm DE,0}$, in terms of $f_p\equiv f_{\rm DE}(z_p)$ and the DE equation of state $w_p\equiv w(z_p)$, at a pivot redshift $z_p$. This provides an alternative framework for assessing the evidence for evolving DE, complementary to the established Chevallier-Polarski-Linder (CPL) parameterization. By parameterizing the DE density directly, the $(w_p,\,f_p)$ formulation avoids the approximate degeneracies intrinsic to the $(w_0,\,w_a)$ basis -- in particular the weak sensitivity of the expansion history to $w_a$ -- while reproducing the background evolution of representative quintessence models with equivalent accuracy. Confronting it with the latest baryon acoustic oscillation (BAO) measurements from DESI, a prior on early-universe parameters from Planck cosmic microwave background (CMB) observations, and Type Ia supernovae (SNe) data, we find that the $w_p$ and $f_p$ parameters are both tightly constrained and sensitive to distinct subsets of the data. Specifically, $w_p$ is measured to percent-level precision by BAO and CMB alone, while $f_p$ is pinned down by the independent matter density constraint that only SNe provide. Including the Pantheon+ SNe sample, we obtain $w_p = -1.04 \pm 0.04$ and $f_p = 1.07 \pm 0.04$, with similar results when using the DESY5 SNe sample. The preference for evolving DE over $Λ$CDM remains below $3σ$ across all dataset combinations, comparable to that obtained with CPL. Notably, the proximity of both $w_p$ and $f_p$ to their cosmological constant values of $(-1,1)$ -- precisely at the epoch where the data are most sensitive -- deepens the coincidence previously identified in the CPL framework, reinforcing the case for caution in interpreting the current evidence for dynamical DE.

We present a joint Bayesian analysis to impose constraints on the generalized $α$-Starobinsky inflationary model using the high-precision cosmological datasets: Planck, CMB lensing from ACT DR6, and Baryon Acoustic Oscillations (BAO) from DESI DR2. For the parameter inference, we introduce an alternative sampling approach. Rather than imposing priors on the cosmological parameters of the inflationary potential $(V_0, \, α, \, N_*)$, we place priors directly on the primordial physical observables $(A_s,\, n_s,\, r)$ through using analytical slow-roll consistency relations, our pipeline internally maps these sampled observables to the corresponding $α$-Starobinsky parameters. These values are then passed to a modified version of $\tt{CLASS}$, which solves the exact inflationary dynamics fully numerically. This pipeline ensures that the final reported posteriors for the observables are computed exactly, completely free from the slow-roll approximation. Applying this methodology, we explore the viability of the $α$-Starobinsky model. We show that, when the full combined dataset is considered, the pure Starobinsky model (i.e., the canonical limit $α= 1$) faces an apparent discrepancy: it requires a large number of $e$-folds of inflation after horizon crossing ($N_* > 60$) due to the shift in the scalar spectral index, $n_s$. In contrast, allowing the deformation parameter $α$ as a free parameter yields a clear $1σ$ preference for $\log_{10} α> 0$ present across all datasets. By favoring a broader inflationary plateau, the $α$-Starobinsky model elegantly reconciles theoretically sound expansion histories with empirical data. Notably, we also show that the addition of ACT DR6 lensing data introduces no significant impact on these primordial constraints, confirming that our robust posteriors are primarily driven by Planck and DESI measurements.

We present joint measurements of the pre- and post-reconstruction power spectra, $P_{\rm pre}$ and $P_{\rm post}$, together with their cross-power spectrum, $P_{\rm cross}$, for the Luminous Red Galaxies (LRGs) in the DESI Data Release 1 (DR1). We jointly analyse these observables with an emulator-based full-shape modeling framework, thereby, for the first time, we extract complementary nonlinear information from the galaxy density field before and after reconstruction in real survey data. Specifically, including $P_{\rm post}$ and $P_{\rm cross}$ in addition to $P_{\rm pre}$ (hereafter $P_{\rm all}$) yields an improvement of approximately $18$-$27\%$ in the $σ_8$ constraint in both $Λ$CDM and $w$CDM, depending on the redshift bin, relative to the $P_{\rm pre}$-only analysis with the cosmic microwave background distance priors (hereafter CMB). In $w$CDM, the joint CMB+$P_{\rm all}$ analysis can tighten the constraints on $w$ by approximately $5$-$15\%$ across the two LRG redshift bins, compared to the CMB+$P_{\rm pre}$ combination. Further incorporating the Type Ia supernova dataset and comparing the cosmological constraints in $w$CDM from each individual power-spectrum component with those from the full combination, we find that $P_{\rm all}$ consistently provides the tightest constraints. From the joint CMB+$P_{\rm all}$+DES-Dovekie dataset, we obtain $Ω_m = 0.314 \pm 0.0048$ and $w = -0.988 \pm 0.023$ for the \texttt{LRG1} sample, and $Ω_m = 0.318 \pm 0.0046$ and $w = -0.988 \pm 0.025$ for \texttt{LRG2}. These results demonstrate that combining pre- and post-reconstruction power spectra with their cross-correlation enables DESI to harvest additional nonlinear information, leading to tighter constraints on cosmological parameters.

Observations with the Canadian Hydrogen Intensity Mapping Experiment (CHIME) have been used to measure the 21 cm intensity mapping auto power spectrum, at $z\sim 1$, over a frequency range from 608.2 MHz to 707.8 MHz at wavenumbers $0.4~h~{\rm Mpc}^{-1} \lesssim k \lesssim 1.5~h~{\rm Mpc}^{-1}$. In this paper, we present the results of two different approaches to interpreting this measurement. In the first approach, we use a parametric power spectrum model to constrain an amplitude parameter, defined as $\mathcal{A}^2_{\rm HI} \equiv 10^6 Ω_{\rm HI}^2(b^2_{\rm HI}+\langle f μ^2\rangle)^2$, where $Ω_{\rm HI}$ is the cosmological density parameter for atomic hydrogen ($\rm HI$), $b_{\rm HI}$ is the linear bias for $\rm HI$, and $\langle f μ^2\rangle$ incorporates the dominant large-scale impact of redshift-space distortions on the angle-averaged power spectrum. Imposing an additional prior on either $Ω_{\rm HI}$ or $b_{\rm HI}$, based on values in the literature, allows us to break the pairwise degeneracy between those two parameters. In the second approach, we compare CHIME's measurement with predictions for the power spectrum of $\rm HI$ from the IllustrisTNG simulations, finding that the measurement disagrees with the TNG100 run at $3.1σ$ and the TNG300 run at $4.0σ$. This disagreement is most likely attributable to the strength of nonlinear redshift-space clustering of $\rm HI$ in the simulations, rather than the total abundance of $\rm HI$, and invites further investigation of the physical processes in the simulations that determine the behavior of $\rm HI$ at nonlinear scales. These results exemplify the ability of 21 cm intensity mapping to provide astrophysical information using measurements at nonlinear scales.

The adiabatic equation of state $P \propto n^Γ$ describes the pressure evolution of highly collisional, isotropic plasmas in terms of their density, providing a possible closure of the fluid moment hierarchy in the absence of heat fluxes and dissipation. An analogous closure exists for collisionless, magnetised plasmas, whose pressure tensor is anisotropic with respect to the magnetic field, and the closure is therefore double-adiabatic, prescribing the evolution of the parallel and perpendicular pressures in terms of the magnetic-field strength and density. Here, we present a general first-principle formalism to derive adiabatic laws using the symmetries of the system. With this theory we recover the adiabatic equation of state $P \propto n^Γ$ for isotropic plasmas and the double-adiabatic equations of state for collisionless, magnetised plasmas. We extend the latter to the relativistic regime, finding that their exact functional form depends on the pressure anisotropy and is not a simple power law. Our double-adiabatic equations of state describe simple geometries, like magnetic mirrors or compressed homogeneous plasmas, as well as complex high-energy astrophysical processes, such as the evolution of plasmoid structures formed during magnetic reconnection.

Upcoming radio surveys will probe the sky with unprecedented depth and sky coverage, enabling a broad range of cosmological and astrophysical applications, as well as powerful synergies with experiments at other wavelengths. The preparation and scientific exploitation of these surveys require realistic mock catalogues that capture the complexity of the radio sky and the interplay of its emitting components. We present a modular and extensible algorithm for generating empirical simulations over the full radio sky, i.e. a solid angle of $4π$ steradians ($f_{\rm sky}=1$), down to redshift $z=5$, comprising both radio continuum and line emission. The framework combines a simulated dark-matter light-cone with empirically sampled galaxy populations and a probabilistic galaxy-halo assignment scheme, producing self-consistent mock catalogues including multiple radio populations on the same light-cone. We release two public catalogues: a shallow catalogue, fully constrained by existing observational data and limited to flux thresholds of $S_\text{1.4 GHz}^\text{lim} \sim 8\times10^{-5}\ \text{Jy}$ at $1.4\ \text{GHz}$ and $S_\text{21}^\text{lim} \sim 2\ \text{Jy}\cdot\text{Hz}$ for the HI 21 cm line; and a deep catalogue extending the calibrated empirical model to better sensitivities, broadly matching future SKAO surveys, with flux limits of $S_\text{1.4 GHz}^\text{lim} \sim 4\times10^{-5}\ \text{Jy}$ and $S_\text{21}^\text{lim} \sim 0.3\ \text{Jy}\cdot\text{Hz}$. The catalogues include radio continuum active galactic nuclei and star-forming galaxies, together with HI-emitting galaxies, for a total of more than 260 million sources in the shallow catalogue and more than 1 billion in the deep catalogue. We validate the simulations by analysing their statistical properties: the mocks reproduce the targeted clustering and population statistics while retaining minimal physical assumptions.

As the (JWST) pursues its observing journey, several thousands of icy-grain spectra are expected to be measured and analysed. The inventory of ices in particular, via the observations of background sources, is accessible for hundreds of lines of sight (LOSs) per molecular-cloud region, opening the possibility to add strong constraints on the solid phase chemistry in a vast domain of cloud densities. SynthIceSpec is a synthetic infrared (IR) spectrum generator that has been designed as a tool to support observing proposals and to test the outcome of chemical models. It is based on laboratory measurements of pure and mixed ices, where each vibrational component is fitted by a sum of Gaussian profiles. Given an initial ice chemical composition (either set by the user or the outputs of a chemical model), a full JWST spectrum is generated, to which the contribution of silicates; continuum, stellar photospheric absorption bands; and extinction law can be added. For the continuum, stellar photospheric models for a wide range of spectral types can be selected by the program, or, Spectral Energy Distribution (SEDs). We present a few use cases of SynthIceSpec: we probed the impact of dust temperature on CO_2 ice formation using IR data and gas-grain modelling. Next, we used SynthIceSpec to explore the detectability of the main feature of CH_3CN at 4.45 um in a cold core environment with the JWST, which was previously tentatively detected in YSOs. The detection thresholds we derive are reasonably low and observable, but identification is directly impacted by the photosphere absorptions that can greatly hinder identification. For some photostellar types, it could remain feasible. Coupled with the Estimated Time Calculator of the Space Telescope Science Institute, SynthIceSpec can be used to find the optimum observational setup for new observations.

The circumgalactic medium (CGM) regulates galaxy growth and retains the imprint of feedback from supernovae and supermassive black holes. However, the bulk of the hot CGM produces little X-ray emission and is challenging to study with X-ray telescopes. We propose a novel method for evaluating the CGM mass using resonant scattering of the helium-like oxygen (OVII) resonant line at $E=574$ eV. In a spherically symmetric and static CGM halo with a sharp central X-ray peak, the number of OVII ions within an outer radial shell can be calculated from the ratio of the two directly observable quantities: the OVII flux from the bright inner region and the scattered OVII flux from the shell (where the scattered flux can be much higher than the intrinsic emission). To evaluate the accuracy of this geometric estimate for realistic galaxies -- with satellites, asymmetries, and gas velocities -- we use a sample of galaxies from the TNG50 cosmological simulation. We find that, when the most irregular systems are excluded based on their X-ray observables, we accurately predict the OVII mass in the outer halo (e.g., in an $r=R_{\rm 500c}-R_{\rm 200c}$ shell) from the ratio of the fluxes in the corresponding annulus and the central peak region ($r<0.2R_{\rm 500c}$), with only a 10% bias and an rms scatter of $\sim 0.2$ dex. As OVII mass strongly correlates with the total oxygen and gas mass, this direct OVII-counting method enables indirect estimates of those quantities by future X-ray microcalorimeter missions, such as NewAthena and HUBS.

We revisit the double adiabatic evolution equations and extend them to the relativistic and ultrarelativistic regimes. We analytically solve the relativistic, time-dependent drift kinetic equation for a homogeneous, magnetized, collisionless plasma and obtain a solution explicitly dependent on the magnetic field and density variations. In the case of an initial relativistic Maxwellian distribution, a natural extension to an anisotropic Maxwell-Jüttner is obtained. We calculate the moments of this time-dependent solution and obtain analytical expressions for the evolution of the perpendicular and parallel pressures in the ultrarelativistic case. We numerically solve the moment equations in the relativistic case and obtain general expressions for the double-adiabatic equations in this regime. We confirm our results using fully kinetic particle-in-cell simulations of shearing and compressing boxes. Our findings can be readily applied to relativistic species including cosmic-rays and electron-positron pairs, present in astrophysical plasmas like pulsar wind nebulae, astrophysical jets, black hole accretion flows, and Van Allen radiation belts.

Fuzzy Dark Matter (FDM), composed of ultra-light axions ($m_a \sim 1 \times 10^{-22}$ eV), exhibits wave-like properties that can significantly impact early-universe star formation. Using the arepo code with the axirepo module, we simulate the assembly of haloes across a range of axion masses ($1 \times 10^{-22}$ eV $\le m_a \le 7 \times 10^{-22} $ eV) and halo masses ($3 \times 10^{8} M_\odot \le M_h \le 8 \times 10^{9} M_\odot$). We investigate how small-scale dynamics of the FDM density field affect the accumulation of cold, dense gas. We find that the delay in star formation scales inversely with both halo mass and axion mass. While the static, cored geometry of the soliton primarily sets the timing of the initial collapse, we identify a secondary dynamical barrier driven by stochastic fluctuations that is most potent at the low-mass end of our parameter space. These dynamics dictate the spatial scale of dense gas by injecting kinetic energy and inducing significant angular momentum, which can rotationally stabilize gas out to the soliton radius. This wave-driven stirring leads to the formation of extended $\text{H}_2$ plumes and promotes dynamical mixing, effectively starving the central regions and forcing gas to cool in a more fragmented, diffuse manner. Our results indicate a shift from the monolithic central star formation seen in CDM toward lower-mass, fragmented clusters. These internal inefficiencies provide a physical mechanism for delaying Cosmic Dawn beyond the effects of the power spectrum cut-off, which is essential for refining observational constraints on the axion mass.

Missing channels in radio-interferometric visibility data can introduce systematic artifacts into the estimated 21-cm power spectrum. A common workaround is to first estimate the two-frequency correlation $C(Δν)$ and then Fourier-transform it to obtain the power spectrum $P(k_\parallel)$. This procedure yields an unbiased estimate when the signal is statistically homogeneous (ergodic) along the line-of-sight, but it fails in the presence of non-ergodic foregrounds. Smooth Component Filtering (SCF) has recently been proposed as a solution to this problem, in which the dominant non-ergodic (spectrally smooth) component is removed prior to estimating $C(Δν)$. In existing implementations, the smooth component is estimated by convolving the visibilities with a Hann window along the frequency axis. We demonstrate that this Hann-based SCF performs adequately only when foregrounds are extremely spectrally smooth, i.e., when they possess a long frequency-correlation length. In contrast, it breaks down when foregrounds exhibit short correlation lengths, as is frequently encountered in real observations. We introduce a Bayesian extension, Bayes-SCF, based on Gaussian Process (GP) regression, which overcomes this limitation. Bayes-SCF models the smooth component via a covariance function with a fixed correlation length, enabling a controlled and data-driven filtering. Using simulated data, we show that Bayes-SCF robustly recovers the input model 21-cm power spectrum even in the presence of spectrally unsmooth foregrounds. Bayes-SCF is also effective in a delay-spectrum approach. The primary trade-off introduced by the Bayesian framework is increased computational cost; future work will focus on optimizing the algorithm and applying it to real MWA data.

We present HEALFormer, a transformer-based neural network architecture for weak gravitational lensing mass mapping that reconstructs convergence maps from incomplete and noisy shear observations on the celestial sphere. The model operates directly on the Hierarchical Equal Area isoLatitude Pixelization and employs learnable mask tokens to handle arbitrary survey geometries without requiring preprocessing. Through a progressive training strategy, HEALFormer efficiently processes high-resolution maps up to Nside = 1024 and demonstrates excellent performance across diverse survey footprints including KiDS, DES, DECaLS, and Planck. The model generalizes robustly to cosmological parameters beyond its training set, producing nearly unbiased reconstructions with superior noise suppression compared to traditional Kaiser-Squires and Wiener filter methods. Remarkably, HEALFormer exceeds the theoretical phase recovery limits of linear reconstruction methods at small scales, achieving a fundamental breakthrough in weak lensing analysis. The combination of computational efficiency, reconstruction accuracy, and adaptability to varying survey configurations makes HEALFormer well-suited for current and next-generation cosmological surveys. Code is available at GitHub.

We present a detailed analysis of the neutral atomic gas (HI) in the dwarf galaxy UGCA 320, observed with the MeerKAT telescope as part of the MHONGOOSE (MeerKAT HI Observations of Nearby Galactic Objects: Observing Southern Emitters) programme. In a small group consisting of three dwarf galaxies, all of which contain HI, it is the most massive. Detailed kinematic modelling shows that UGCA 320 contains a substantial amount of (kinematically) anomalous gas (>=20%), at least ~30% of which is likely the result of a tidal interaction with its neighbour UGCA 319. It also reveals that UGCA 320 likely harbours a star-formation driven outflow, and that ~10% of its HI is extra-planar and has a filamentary structure. Although UGCA 320 aligns with established scaling relations from the literature, its neutral hydrogen content is notably complex - shaped by its immediate environment. This underscores the importance of deep, resolved observations and detailed kinematic analyses to capture the nuances of galaxy evolution.

We investigate the impact of observational and theoretical uncertainties in cluster gravitational redshifts as a probe of modified gravity using an end-to-end forecasting pipeline. We use a generative model to build a halo catalogue with $M_{500}\ge 3\times 10^{13}\,M_\odot$, populate haloes with member galaxies via a five-parameter halo occupation distribution (HOD), assign projected positions from radial density profiles, apply survey-like selections, and infer a linear rescaling of the gravitational potential, $α_\mathrm{MG}$, to parameterise modifications to general relativity (GR). We vary redshift uncertainties, radial and mass-redshift completeness, member abundance, minimum mass and maximum redshift, as well as mis-specify the clusters density and velocity profiles, centres, and mass function. We find that the intracluster velocity dispersion sets an effective floor: improving redshift precision beyond $σ_z\sim 10^{-4}(1+z)$ brings no improvement in the precision of $α_\mathrm{MG}$. Realistic redshift and mass cuts primarily remove low-mass haloes and have minimal impact on the $α_\mathrm{MG}$ precision. In this setting, we find that shallow, narrower spectroscopic surveys are preferable to deep, wide photometric ones for precise modified gravity constraints. We further find that mis-centring can mimic significant departures from GR. Baryonic deviations from a Navarro-Frenk-White profile and velocity anisotropies do not introduce appreciable biases. In the high-S/N regime of upcoming surveys, accurate determination of cluster centres will be essential to avoid interpreting systematic effects as new physics. The Spectroscopic Stage-5 Experiment and the Widefield Spectroscopic Telescope provide a clear route toward establishing gravitational redshifts as a competitive probe of modified gravity.

Flat-Spectrum Radio Quasars (FSRQs) are among the most energetic and powerful active galactic nuclei, often exhibiting jet powers comparable to or exceeding the Eddington luminosity. In this work, we performed broadband spectral energy distribution (SED) modeling of two FSRQs PKS 1441+25 and Ton 599, using Swift-XRT/UVOT, NuSTAR, Fermi-LAT and VERITAS observations during 2015 and 2021, respectively. We considered four particle distribution models: a broken power law, a log-parabola, and two energy-dependent models in which either the diffusion or acceleration timescale depends on energy. Our results show that the jet power estimates derived from models with intrinsic curvature, such as the log-parabola and energy-dependent models, are of the same order as those obtained with a broken power-law distribution. This contrasts with the case of High Synchrotron Peaked Blazars (HBLs), where the power estimates can differ by nearly two orders of magnitude between models. We attribute this difference to the lower electron break energies typically observed in FSRQs. Consequently, our findings suggest that, unlike in HBLs, the estimated jet powers in FSRQs are relatively insensitive to the assumed particle energy distribution, reflecting the dominance of external Compton processes and weaker dependence on spectral curvature.

The Vera C. Rubin Observatory Legacy Survey of Space and Time (LSST) will produce unprecedentedly deep and wide photometric catalogs, enabling transformative studies of faint stellar systems such as the research of ultra-faint dwarf galaxies (UFDs). A critical challenge for these studies is reliable star-galaxy separation at faint magnitudes, where compact background galaxies increasingly contaminate stellar samples. This work aims to assess the performance of supervised machine-learning techniques for star-galaxy separation in LSST-like data, quantify the relative importance of morphological and photometric information, and identify the most effective combinations of input features for minimizing galaxy contamination while preserving stellar completeness in the faint regime relevant for UFD searches. We apply a Random Forest classifier to observations of the Extended Chandra Deep Field South from LSST Data Preview 1 (DP1), the deepest field observed within the DP1. We construct a curated sample of bona fide stars and galaxies using spectroscopic data, Gaia DR3, and multi-band photometric catalogs. We train and validate the classifier using several configurations of LSST-based input features, including multi-band colors, the LSST morphological parameter refExtendedness, and photometric uncertainties. We find that LSST multi-band photometry alone delivers a good star-galaxy separation, significantly outperforming morphology-based classification at faint magnitudes. Colors involving the u-band are essential to provide a robust star galaxy separation. Furthermore, explicitly including photometric uncertainties as input features yields the best overall performance. Across all configurations that include all the six LSST filters, galaxy contamination remains negligible almost the whole magnitude range probed in this work (i.e. r < 27.5 mag). [abridged]

The East Asia VLBI Network (EAVN) has recently enabled dual-polarization observations at $22$ and $43\,\mathrm{GHz}$. We present the first systematic verification of its polarimetric performance using EAVN observations of M87, 3C 279, 3C 273, and OJ 287, calibrated with the GPCAL pipeline and evaluated against near-contemporaneous VLBA images at comparable frequencies. Most stations show stable polarimetric leakages with amplitudes of $5$-$10\%$ over monthly timescales. While several VERA stations exhibit D-term phase variations between epochs, we attribute these to field-rotator (FR) offsets and demonstrate that phase stability is restored after applying the analytically derived FR corrections. The resulting linear-polarization morphologies and EVPAs broadly agree with the VLBA results within uncertainties; fractional polarization measured by the EAVN tends to be slightly higher near polarization peaks. Although exact one-to-one comparisons are limited by moderate frequency and epoch differences, the combined evidence indicates robust EAVN polarimetric calibration and imaging capabilities at $22$ and $43\,\mathrm{GHz}$. These results support the scientific capability of EAVN polarimetry and lay the groundwork for expanded, higher-fidelity polarimetric studies in East Asia.

The discovery of the complex stellar populations hosted in two massive stellar systems in the Galactic bulge, namely Terzan5 and Liller 1, posed intriguing questions about their origin. Despite their globular cluster appearance, they host sub-populations with significantly different ages (several Gyrs) and metallicities (about 1 dex) tracing a chemical abundance pattern that is consistent only with that observed in the bulge. These surprising properties can be naturally explained in the context of a self-enrichment scenario, opening the possibility that they could be the remnants of primordial massive structures that contributed to the bulge formation (the so-called Bulge Fossil Fragments, BFFs) capable of retaining supernova ejecta within their potential well. In this paper we present a first attempt to quantify the expected contribution of BFFs to the gravitational wave emission. In particular, by adopting Terzan5 as prototype of BFF, using its chemical evolutionary model, and following a scaling relation derived for globular clusters, we present a first-guess estimate of the number of binary black hole (BH) mergers expected in this stellar system. Within the adopted simplifying assumptions and the uncertainties about the initial conditions of the proto-Terzan 5 system, we find that several hundreds of binary BH mergers are expected, a number that is between 15 and 250 times larger than that produced by a typical globular cluster. Hence, this study identifies in the BFF family a new population of stellar systems potentially able to produce a significant number of gravitational wave emitters, that has not been considered in any previous investigation. Moreover we speculate that they could also be the natural place where BHs with masses above 60 Msun and even intermediate-mass BHs can form via repeated dynamical interactions.

Tidal disruption events (TDEs) are powerful probes of supermassive black hole (SMBH) properties and accretion physics. The existing light curve fitting tools assume that the disrupted stars are on parabolic orbits, which may introduce systematic biases in derived parameters. In this work, we develop a non-parabolic TDE model that incorporates orbital energy of the disrupted star as a free parameter ($\tildeε_{\rm orb}$) to modify the debris mass distribution and mass fallback rate. We apply this model to 30 TDEs from the ZTF-I survey and compare the results with those from a standard parabolic model. We find that neglecting orbital energy leads to biased black hole mass estimates: for eccentric (hyperbolic) orbits, parabolic models systematically underestimate (overestimate) the black hole mass. Additionally, we measure orbital eccentricities ($e$) and penetration factors ($β$) of the disrupted stars in this sample, enabling an investigation of their origins via the $e$-$β$ parameter space. Most events (24/30) are consistent with production via two-body relaxation in spherical nuclear star clusters, but six outliers with high $β$ and $e<1$ suggest alternative mechanisms. Our results highlight the importance of accounting for orbital energy in TDE modeling to improve the accuracy of SMBH mass measurements and to better understand the dynamical origin of the disrupted stars.

Cosmological filaments, galaxy clusters, and galaxies are magnetized reservoirs of cosmic rays (CRs). The exchange of CRs across these structures is usually modeled assuming that they remain charged and magnetically confined. At high energies, hadronic interactions can convert CR protons to neutrons. This physics is routinely included in air-shower and ultra-high-energy (UHE) CR propagation Monte Carlo simulations used for composition studies but is rarely treated explicitly in propagation models of CR transport and exchange between magnetized reservoirs. CR neutrons are not affected by magnetic fields and can propagate ballistically over kpc-Mpc distances before decaying back into protons, with relativistic time dilation extending their effective decay length. We show how such charged-neutral switching modifies CR confinement and escape in four representative environments: a Milky Way-like galaxy, a starburst galaxy, a galaxy cluster, and a cosmological filament. By solving the transport of a confined CR proton population in each structure using a diffusion/streaming propagation approach with hadronic pp and p$γ$ interactions, and treating neutron production and decay as a stochastic Poisson ''jump'' process, we find that neutron-mediated steps can allow additional CR escape from large-scale cosmological structures at energies where charged-particle transport alone would predict strong CR confinement and attenuation in ambient radiation fields. These effects imply a qualitative shift in how ultra-high-energy CRs are transferred from embedded sources into filaments and voids once intermediate neutron propagation is considered, with consequences for the partitioning of CRs across the large-scale structure of the Universe.

AstroSat/LAXPC and SXT observed the persistent neutron star low-mass X-ray binary 4U 1820-30 between 2016 and 2022. During these observations, the hardness-intensity diagram (HID) and color-color diagram (CCD) indicated that the source was in the banana state. We divided the CCD into 11 segments for spectral and timing analyses. For each segment in the CCD, we modeled the spectral data using two distinct approaches over the 0.7-20.0 keV band. A combination of a multi-color-disk component with an inner disk temperature of around 0.6 keV and Comptonized emission from the boundary layer (BL)/ hot corona provided the best description of the X-ray spectral data of this source. The truncation radius was found to be in the range of $\sim$ 19-40 km. The Comptonized component has an optical depth in the range of $\sim 7 - 13$ with electron temperature in the range of $\sim 2.5 - 3.8$ keV. The optical depth of the corona varies significantly along the position on the CCD, while $\sim$ 80\% of the X-ray flux comes from the Comptonized component. We discuss possible physical scenarios to explain the relationship between the spectral evolution and motion of the source along the CCD. The timing analysis revealed kHz QPOs peaks at $\sim 710$ Hz and $\sim 740$ Hz in the lower left banana branch. An energy-dependent study indicates that these QPOs are stronger in the high-energy band.

We present results from a mini-survey of dust spectral features arising in galaxies at redshifts $0.5 < z < 1.2$ in our James Webb Space Telescope (JWST) mid-infrared spectra of physically-unrelated background quasars. We analyze the JWST Mid-infrared Instrument (MIRI) Medium-Resolution Spectrometer (MRS) spectra of five quasars presented in Klimenko, Kulkarni, \& Aller 2025a (Paper I) to determine the best-fit silicate mineralogies. Template profile fits to the 10 $μ$m feature suggest the possible presence of crystalline silicates in three of the galaxies. This contrasts with the predominately amorphous silicate grains in the Milky Way diffuse interstellar medium (ISM). We also measure the extinction curves using existing data from UV to mid-IR. Combining our results with past Spitzer IRS studies, we find that (i) the 10~$μ$m silicate peak optical depth ($τ_{10}$) is about three times stronger than expected for the local diffuse ISM over the range $A_V =0.1-2.0$, with $τ_{10}/A_V$=$0.17\pm0.09$. (ii) The relative strength of the UV bump is similar to that in the local ISM. However, the ratio $τ_{10}/A_{2175}$ is larger ($\sim0.1-1$), and appears to decrease with $A_V$, approaching the Galactic ISM value ($\sim 0.1$) at $A_V\sim1.5-2$. (iii) No significant correlation of $τ_{10}/A_V$ with $R_V$. (iv) $τ_{10}$ is strongly correlated with the gas-phase Mg~II absorption strength for the quasar sightlines. Possible interpretations include that some quasar sightlines probe dust in the circumgalactic medium (CGM), and that dust grains may have been significantly reprocessed in the ISM and CGM under conditions that may differ from those in the local ISM.

We present high-resolution optical spectroscopy of interstellar comet 3I/ATLAS (C/2025 N1) obtained with the High Dispersion Spectrograph mounted on the Subaru Telescope on UT 2026 January 7, when the comet was on its outbound trajectory at a heliocentric distance of $r_{\mathrm{h}} = 2.87$ au. The spectra cover the forbidden atomic oxygen lines, [O~I], at 557.7, 630.0, and 636.4 nm. The [O~I] red-doublet intensity ratio $I_{630.0}/I_{636.4} = 2.91 \pm 0.21$ matches the optically thin branching ratio ($\sim$3; \citealt{StoreyZeippen2000}), indicating that optical-depth effects are small and that our relative flux calibration is reliable. We measure a green-to-red [O~I] intensity ratio of $G/R = I_{557.7}/(I_{630.0} + I_{636.4}) = 0.339 \pm 0.027$. This value is higher than those of most Solar System comets at similar heliocentric distances, but comparable to that of the interstellar comet 2I/Borisov. From the measured $G/R$ ratio in 3I/ATLAS, we estimate the CO$_2$/H$_2$O abundance ratio under the assumption that H$_2$O and CO$_2$ are the dominant parents of O($^1$S) and O($^1$D), with other oxygen-bearing species expected to have a smaller influence under typical conditions (e.g., \citealt{FestouFeldman1981}). The derived ratio is significantly lower than the extremely CO$_2$-rich composition reported from infrared observations on the inbound trajectory at $r_{\mathrm{h}} \sim 3.3$ au, yet higher than typical values measured for comets in the Solar System. Together with published pre- and post-perihelion measurements, our result indicates that the CO$_2$/H$_2$O ratio decreased substantially across perihelion.

We present a search for neutrino counterparts to fast radio bursts (FRBs) using temporal and spatial cross-matching between the Second CHIME/FRB catalogue and the IceCube high-energy alert-track catalogue ICECAT1. Because current FRB--neutrino models do not provide a unique consensus on emission ordering, our primary significance test adopts a two-sided, order-agnostic temporal hypothesis. The analysis accounts for declination-dependent CHIME/FRB exposure and the look-elsewhere effect across multiple trials. No statistically significant FRB--neutrino association is found. The most significant pair is FRB\,20190630C--IC\,190629A, with a post-trial probability of $p=0.076$ ($1.43σ$), consistent with a chance coincidence. Within our statistical framework, a detectable physical association would require a time offset shorter than $\sim256$~s at $3σ$ or $\sim63$~ms at $5σ$. Using a population-level stacking analysis, we derive 90\% upper limits on the neutrino-to-radio luminosity ratio of FRBs, $ξ\lesssim 10^{8}-10^{11}$ for neutrino power-law spectral indices $γ=1.0-3.0$. These limits improve upon previous constraints by approximately two orders of magnitude and represent the most stringent bounds from FRB--neutrino coincidence searches to date. Although the current limits remain above the predictions of most magnetar-based models, they begin to constrain scenarios involving exceptionally efficient hadronic energy dissipation.

A bar-like central feature is commonly observed in both nearby and distant spiral-type galaxies, including the Milky Way. While many methods exist to categorise this morphology, no one method has emerged as the field-wide standard. To develop a rigorous and consistent method for identifying these bars, we investigate a classification scheme based on dynamical actions. In the Gaia era, actions can be estimated for individual stars in both observations and simulations, making this a natural and unifying diagnostic, assuming the intrinsic errors and selection functions are understood. Our approach is straightforward: stars that participate in the bar are subject to a strongly non-axisymmetric potential and, therefore, do not completely conserve their actions. We use this property to define a star-by-star criterion, formulated as an inequality and evaluated within measurement uncertainties, to identify bar members based on the degree to which their total action fails to be conserved. From tests on simulated galaxies, we find that the bar region is indeed characterised by a lower fraction of stars with conserved actions and that stars on bar orbits are represented by larger percentage changes in their actions. We are able to classify the spatial extent of barred region via the standard parameters of bar length and orientation, while also individually separating bar-located from bar-member stars on bar orbits. As proof of concept, our automated method based on dynamical actions robustly identifies bar parameters that closely match the eye's performance (average bar length variation ~9%) in barred snapshots of the test galaxy.

This study investigates the X-ray characteristics of jet-associated radio-quiet AGNs across distinct optical/UV to X-ray correlation phases. Quasi-simultaneous optical/UV/X-ray observations of NGC 4051 from May-June 2009, obtained through Swift and XMM-Newton, reveal a temporal dichotomy: a strong optical/UV to X-ray correlation dominates the initial observation phase (before May 27), followed by an optical/UV flare event concurrent with X-ray flux suppression in the latter period. Our multi-method analysis of XMM-Newton data, incorporating short-term X-ray variability assessment, spectral decomposition, and RGS spectral analysis, identifies significant inter-phase X-ray emission disparities. During optical/UV flaring episodes, compared to the correlated phase, we observe: attenuated short-term X-ray variability amplitudes, enhanced soft X-ray absorption, suppressed intrinsic hard X-ray flux, and more prominent RGS emission-line features. Notably, these X-ray characteristics during optical/UV flaring intervals show no statistically significant deviations from pre-flare low-state X-ray emission patterns. These non-synchronous optical/UV-X-ray variations contradict predictions from both reprocessing models, starburst-driven emission scenarios, and the simplistic absorption models. While potential jet-related mechanisms remain ambiguous, our findings demonstrate strong consistency with predictions from the inhomogeneous accretion disk perturbation framework.

We perform MHD simulations of AGN-blown bubbles in the Intercluster Medium (ICM) containing large-scale coherent magnetic fields. We assume that bubbles, created by the intermittent jets from Active Galactic Nuclei, quickly relax to the Woltjer-Taylor spheromak-like state, with internal plasma beta-parameter $\sim 1$. We demonstrate that such bubbles rising through hydrostatically-stratified atmosphere are magnetically stabilized against fluid interface instabilities, remaining coherent for a long time. Typical velocity is $ v /c_s \sim \sqrt{R/H} \leq 1 $ ($c_s$ is sound speed, $R$ is the bubble size, $H$ is the scale height). Current-driven instabilities (internal kinks) lead to bubble's tilting, but develop on long time scales, and remain unimportant, leading to minor modifications of the internal structure. Our results explain apparent long-term stability of ICM cavities. Subsonically rising stable bubbles dissipate in their wake approximately the energy initially injected by the jet, and may efficiently reheat the clusters cores in a ``gentle'' way.

We reassess the direct collapse black hole (DCBH) interpretation of UHZ1 (UNCOVER-26185), a gravitationally lensed galaxy at $z_\mathrm{spec}=10.054$. That interpretation rests on a hard ($2-7$ keV) X-ray excess detected with Chandra, attributed to a Compton-thick AGN with an inferred $2-10$ keV luminosity of $L_\mathrm{X,int}\sim10^{46}~\mathrm{erg~s^{-1}}$ (Bogdan et al. 2024). The resulting extreme X-ray to rest-frame optical-IR ratio was taken as the hallmark signature of an "outsize black hole galaxy" at cosmic dawn. We analyse the full 2.2 Ms Chandra imaging dataset -- including 0.95 Ms of unpublished observations -- and present new JWST/MIRI photometry at $λ_\mathrm{obs}>5~μ\mathrm{m}$. Across the full range of plausible Chandra data reductions, the $2-7$ keV excess at the position of UHZ1 reaches a significance of only $2.3-2.9σ$; the originally reported $4.2-4.4σ$ detection is sensitive to the specific astrometric alignment adopted and is not robustly reproducible. Moreover, the hard X-ray signal does not grow with the additional exposure, contrary to expectations for a steady source, indicating that any excess is not persistent. UHZ1 is also undetected in all nine MIRI imaging bands. Fitting red/obscured AGN SED templates to the tightest MIRI upper limit, we constrain the bolometric luminosity of any buried AGN to $L_\mathrm{bol}<1.3\times10^{45}~\mathrm{erg~s^{-1}}$. These conclusions are further supported by independent JWST spectroscopy (Alvarez-Marquez et al. 2026), which reveals no AGN signatures in the rest-frame UV or optical. Taken together, the multiwavelength data paint a consistent picture of UHZ1 as a low-mass, metal-poor, star-forming galaxy in the early Universe, with no compelling evidence for a luminous obscured AGN, regardless of its proposed formation channel.