arXiv Daily Digest - 2026-04-30

We present KAYRA, an end-to-end karyotyping system that operates inside the operational constraints of a clinical cytogenetic laboratory. KAYRA is architected as a containerized microservice pipeline whose ML stack combines an EfficientNet-B5 + U-Net semantic segmenter, a Mask R-CNN (ResNet-50 + FPN) instance detector, and a ResNet-18 classifier, orchestrated through a cascaded ROI-narrowing strategy that focuses each downstream model on the chromosome-bearing region. The same container images are deployed both as a cloud service and as an on-premise installation, supporting clinical environments where patient-data egress is not permitted as well as those where it is. A pilot clinical evaluation against two commercial reference karyotyping systems on 459 chromosomes from 10 metaphase spreads shows segmentation accuracy of 98.91 % (vs. 78.21 % / 40.52 %), classification accuracy of 89.1 % (vs. 86.9 % / 54.5 %), and rotation accuracy of 89.76 % (vs. 94.55 % / 78.43 %). KAYRA improves over the older density-thresholding reference on all three axes (p < 0.0001 for segmentation and classification by Fisher's exact test on chromosome-level counts), and on segmentation also against the modern AI- supported reference (p < 0.0001); on classification the difference vs. the modern AI reference is not statistically significant at the present test-set size (p = 0.34). The system reaches TRL 6 maturity and integrates the human-in-the-loop expert-review workflow that diagnostic cytogenetic practice requires. The thesis of this paper is that a multi-model cytogenetic AI service can be packaged as a microservice architecture supporting flexible deployment - cloud-hosted or on-premise - while delivering strong empirical performance on a pilot clinical evaluation.

Large language models (LLMs) acquire most of their factual knowledge during the pre-training stage, through next token prediction. Subsequent stages of post-training often introduce new facts outwith the parametric knowledge, giving rise to hallucinations. While it has been demonstrated that supervised fine-tuning (SFT) on new knowledge may exacerbate the problem, the underlying mechanisms are still poorly understood. We conduct a controlled fine-tuning experiment, focusing on closed-book QA, and find latent directions that causally contribute to hallucinations. Specifically, we fine-tune Llama 3.1 8B, Gemma 2 9B and Mistral 7B v03 on seven distinct single QA datasets, controlling for the percentage of new knowledge and number of training epochs. By measuring performance on the test set, we validate that incrementally introducing new knowledge increases hallucinations, with the effect being more pronounced with prolonged training. We leverage pre-trained sparse autoencoders (SAEs) to analyze residual stream activations across various checkpoints for each model and propose Monotonic Relationship Feature Identification (MoRFI) for capturing causally relevant latents. MoRFI filters SAE features that respond monotonically to controlled fine-tuning data mixtures of a target property. Our findings show that exposure to unknown facts disrupts the model's ability to retrieve stored knowledge along a set of directions in the residual stream. Our pipeline reliably discovers them across distinct models, recovering knowledge through single-latent interventions.

Deploying accurate object detection for Vulnerable Road User (VRU) safety on edge hardware requires balancing model capacity against computational constraints. Large models achieve high accuracy but fail under INT8 quantization required for edge deployment, while small models sacrifice detection performance. This paper presents a knowledge distillation (KD) framework that trains a compact YOLOv8-S student (11.2M parameters) to mimic a YOLOv8-L teacher (43.7M parameters), achieving 3.9x compression while preserving quantization robustness. We evaluate on full-scale BDD100K (70K training images) with Post-Training Quantization to INT8. The teacher suffers catastrophic degradation under INT8 (-23% mAP), while the KD student retains accuracy (-5.6% mAP). Analysis reveals that KD transfers precision calibration rather than raw detection capacity: the KD student achieves 0.748 precision versus 0.653 for direct training at INT8, a 14.5% gain at equivalent recall, reducing false alarms by 44% versus the collapsed teacher. At INT8, the KD student exceeds the teacher's FP32 precision (0.748 vs. 0.718) in a model 3.9x smaller. These findings establish knowledge distillation as a requirement for deploying accurate, safety-critical VRU detection on edge hardware.

The integration of Large Language Models (LLMs) into the software development lifecycle (SDLC) masks a critical socio-technical failure: Cognitive-Systemic Collapse. This paper introduces "Epistemological Debt," the hidden carrying cost incurred when engineers substitute logical derivation with passive AI verification. This debt erodes the mental models essential for root-cause analysis, widening the gap between system complexity and human comprehension. Furthermore, recursive training on synthetic code threatens to homogenize the global software reservoir, diminishing the variance required for robust engineering. Using the 2026 Amazon outages as a case study, this research illustrates how "mechanized convergence" leads to systemic fragility. To preserve long-term resilience, engineering leaders must move beyond prompt-based development to implement rigorous human-in-the-loop pedagogical standards. This framework balances AI-driven productivity with the epistemic sovereignty necessary to manage increasingly opaque software ecosystems.

When generative AI (genAI) systems are used in high-stakes decision-making, its recommended role is to aid, rather than replace, human decision-making. However, there is little empirical exploration of how professionals making high-stakes decisions, such as those related to employment, perceive their agency and level of control when working with genAI systems. Through interviews with 22 recruiting professionals, we investigate how genAI subtly influences control over everyday workflows and even individual hiring decisions. Our findings highlight a pressing conflict: while recruiters believe they have final authority across the recruiting pipeline, genAI has become an invisible architect that shapes the foundational building blocks of information used for evaluation, from defining a job to determining good interview performances. The decision of whether or not to adopt was also often outside recruiters' control, with many feeling compelled to adopt genAI due to calls to integrate AI from higher-ups in their business, to combat applicant use of AI, and the individual need to boost productivity. Despite a seemingly seismic shift in how recruiting happens, participants only reported marginal efficiency gains. Such gains came at the high cost of recruiter deskilling, a trend that jeopardizes the meaningful oversight of decision-making. We conclude by discussing the implications of such findings for responsible and perceptible genAI use in hiring contexts.

Many of the thousands of attested languages share common configurations of features, creating a spectrum from typologically very rare (e.g., object-verb-subject word order) or impossible languages to very common combinations of features (e.g., subject-object-verb word order). One central question is under what conditions such typological tendencies can be predicted, and specifically whether the learning bias of language models (LMs) is sufficient to reproduce such patterns. In this study, we add one dimensionality to such analysis -- the learning scenario for LMs -- to explore its interaction with the inductive bias of LMs. Specifically, as a first study, we examine the effect of curriculum learning (CL), as a developmentally motivated learning scenario, i.e., starting with simpler sentences rather than randomly-ordered input. We expand existing LM-based exploration (El-Naggar et al., 2025a,b) with a simple CL variant and find that CL substantially impacts the apparent inductive bias of LMs.

When do language diffusion models memorize their training data, and how to quantitatively assess their true generative regime? We address these questions by showing that Uniform-based Discrete Diffusion Models (UDDMs) fundamentally behave as Associative Memories (AMs) $\textit{with emergent creative capabilities}$. The core idea of an AM is to reliably recover stored data points as $\textit{memories}$ by establishing distinct basins of attraction around them. Historically, models like Hopfield networks use an explicit energy function to guarantee these stable attractors. We broaden this perspective by leveraging the observation that energy is not strictly necessary, as basins of attraction can also be formed via conditional likelihood maximization. By evaluating token recovery of $\textit{training}$ and $\textit{test}$ examples, we identify in UDDMs a sharp memorization-to-generalization transition governed by the size of the training dataset: as it increases, basins around training examples shrink and basins around unseen test examples expand, until both later converge to the same level. Crucially, we can detect this transition using only the conditional entropy of predicted token sequences: memorization is characterized by vanishing conditional entropy, while in the generalization regime the conditional entropy of most tokens remains finite. Thus, conditional entropy offers a practical probe for the memorization-to-generalization transition in deployed models.

Long-context LLM serving is bottlenecked by the cost of attending over ever-growing KV caches. Dynamic sparse attention promises relief by accessing only a small, query-dependent subset of the KV state per decoding step and extending the KV storage to CPU memory. In practice, however, these algorithmic savings rarely translate into end-to-end system-level gains because sparse methods typically operate at different granularities and thus rely on ad hoc, per-algorithm implementations. At the same time, hierarchical KV storage introduces a new systems bottleneck: retrieving fine-grained, irregular KV subsets across the GPU-CPU boundary can easily erase the benefits of sparsity. We present SPIN, a sparse-attention-aware inference framework that co-designs the execution pipeline with hierarchical KV storage through three techniques: (1) a unified partition abstraction that maps different sparsity granularities onto a shared page-based KV substrate; (2) a locality-aware KV cache manager that dynamically sizes per-request HBM budgets and uses a GPU-friendly bucketed LRU policy to cut PCIe round-trips; and (3) a two-level hierarchical metadata layout sized to the active working set rather than the worst-case address space. Built on vLLM with three representative sparse attention algorithms, SPIN delivers 1.66-5.66x higher end-to-end throughput and 7-9x lower TTFT than vLLM, and reduces TPOT by up to 58% over the original sparse-attention implementations.

Predictive safety filters (PSFs) leverage model predictive control to enforce constraint satisfaction during deep reinforcement learning (RL) exploration, yet their reliance on first-principles models or Gaussian processes limits scalability and broader applicability. Meanwhile, model-based RL (MBRL) methods routinely employ probabilistic ensemble (PE) neural networks to capture complex, high-dimensional dynamics from data with minimal prior knowledge. However, existing attempts to integrate PEs into PSFs lack rigorous uncertainty quantification. We introduce the Uncertainty-Aware Predictive Safety Filter (UPSi), a PSF that provides rigorous safety predictions using PE dynamics models by formulating future outcomes as reachable sets. UPSi introduces an explicit certainty constraint that prevents model exploitation and integrates seamlessly into common MBRL frameworks. We evaluate UPSi within Dyna-style MBRL on standard safe RL benchmarks and report substantial improvements in exploration safety over prior neural network PSFs while maintaining performance on par with standard MBRL. UPSi bridges the gap between the scalability and generality of modern MBRL and the safety guarantees of predictive safety filters.

We introduce HalluCiteChecker, a toolkit for detecting and verifying hallucinated citations in scientific papers. While AI assistant technologies have transformed the academic writing process, including citation recommendation, they have also led to the emergence of hallucinated citations that do not correspond to any existing work. Such citations not only undermine the credibility of scientific papers but also impose an additional burden on reviewers and authors, who must manually verify their validity during the review process. In this study, we formalize hallucinated citation detection as an NLP task and provide a corresponding toolkit as a practical foundation for addressing this problem. Our package is lightweight and can perform verification in seconds on a standard laptop. It can also be executed entirely offline and runs efficiently using only CPUs. We hope that HalluCiteChecker will help reduce reviewer workload and support organizers by enabling systematic pre-review and publication checks. Our code is released under the Apache 2.0 license on GitHub and is distributed as an installable package via PyPI. A demonstration video is available on YouTube.

We present a quantum feature-selection framework based on a higher-order unconstrained binary optimization (HUBO) formulation that explicitly incorporates multivariate dependencies beyond standard quadratic encodings. In contrast to QUBO-based approaches, the proposed model includes one-, two-, and three-body interaction terms derived from mutual-information measures, enabling the objective function to capture feature relevance, pairwise redundancy, and higher-order statistical structure within a unified energy model. To suppress trivial all-selected solutions, we further include structured linear penalties that promote sparsity while preserving informative variables. The resulting HUBO instances are optimized with digitized counterdiabatic quantum optimization on IonQ Forte and compared against noiseless quantum simulation as well as two classical dimensionality-reduction baselines: SelectKBest based on mutual information and principal component analysis (PCA). We evaluate the proposed workflow on two benchmark classification datasets, namely the Gallstone dataset and the Spambase dataset, and analyze both predictive performance and selected-subset structure. The results show good qualitative agreement between hardware executions and noiseless simulations, supporting the feasibility of implementing higher-order feature-selection Hamiltonians on current trapped-ion processors. In addition, the quantum approach yields competitive classification performance while producing compact and informative feature subsets, highlighting the potential of higher-order quantum optimization for machine-learning preprocessing tasks.

This paper presents a hierarchical decision-making framework for unmanned aerial vehicle (UAV) missions motivated by search-and-rescue (SAR) scenarios under limited simulation training. The framework combines a fixed rule-based high-level advisor with an online goal-conditioned low-level reinforcement learning (RL) controller. To stress-test early adaptation, we also consider a strict no-pretraining deployment regime. The high-level advisor is defined offline from a structured task specification and compiled into deterministic rules. It provides interpretable mission- and safety-aware guidance through recommended actions, avoided actions, and regime-dependent arbitration weights. The low-level controller learns online from task-defined dense rewards and reuses experience through a mode-aware prioritized replay mechanism augmented with rule-derived metadata. We evaluate the framework on two tasks: battery-aware multi-goal delivery and moving-target delivery in obstacle-rich environments. Across both tasks, the proposed method improves early safety and sample efficiency primarily by reducing collision terminations, while preserving the ability to adapt online to scenario-specific dynamics.

I propose the \emph{Random Cloud} method, a training-free approach to neural architecture search that discovers minimal feedforward network topologies through stochastic exploration and progressive structural reduction. Unlike post-training pruning methods that require a full train-prune-retrain cycle, this method evaluates randomly initialized networks without backpropagation, progressively reduces their topology, and only trains the best minimal candidate at the end. I evaluate on 7 classification benchmarks against magnitude pruning and random pruning baselines. The Random Cloud matches or outperforms both baselines in 6 of 7 datasets, achieving statistically significant improvements on Sonar ($+4.9$pp accuracy, $p{=}0.017$ vs magnitude pruning) with 87\% parameter reduction. Crucially, the method is faster than both pruning baselines in 4 of 5 datasets (0.67--0.94$\times$ the cost of full training), since it avoids training the full-size network entirely.

High-performance computing (HPC) systems are increasingly exploring dynamic resource management and malleable MPI applications to better adapt to heterogeneous architectures, fluctuating workloads, and energy constraints. However, the correctness of the libraries that support these techniques is often evaluated through ad hoc experiments that can be difficult to reproduce and maintain. This article introduces methodology for testing dynamic resource management frameworks that combines a taxonomy of tests for MPI malleable libraries with an HPC-oriented continuous integration (CI) ecosystem. The taxonomy structures functional and non-functional tests at both component-integration and system levels. The CI ecosystem instantiates this taxonomy in a containerized virtual cluster enabling automated validation. The approach is instantiated and evaluated using the Dynamic Management of Resources (DMR) framework as a representative case study. Results show that the proposed methodology improves early fault detection, simplifies maintenance under evolving dependencies, and transfers to other malleability solutions that expose analogous primitives for initialization, readiness checking, and reconfiguration.

We present a Spatially Embedded Evolutionary Algorithm where robot individuals exist in a physically simulated 2D environment, must navigate to encounter potential mates, and compete for survival under various spatially-aware selection pressures. Using HyperNEAT evolved neural controllers for ARIEL gecko-inspired quadrupeds in MuJoCo, we investigate how spatial structure fundamentally alters evolutionary dynamics. Our experiments show a modest 4.9% difference in peak fitness between proximity-based and random pairing possibly within stochastic variation while combining spatial parent selection with stochastic death selection produces unstable population dynamics. We discover a continuous phase transition in energy-based selection experiments, with critical zone count separating extinction-dominated and explosion-dominated regimes. Our density-dependent death selection mechanism achieves 97% completion rates but causes fitness decline, revealing a fundamental dilemma where decoupled mechanisms produce bistable dynamics, positively coupled mechanisms create counter-selection pressures, and only deterministic fitness-based selection maintains stability. These findings provide important constraints for future spatial EA design.

To overcome the well-known memory bottleneck of AI chips, 3D stacked architectures that employ advanced packaging technology with high-density through-silicon vias (TSVs) pins have proven to be a promising solution. The 3D-stacked AI chip enables ultra-high memory bandwidth between compute and memory by stacking numerous DRAM banks atop many AI cores in a distributed manner. However, it is not easy to explore the efficiency of the 3D-stacked AI chip, due to its unique distributed nature. And we need to carefully consider multiple intertwined factors that range from upper-level computing paradigm to machine learning (ML) compiler optimizations, and to the underlying hardware architecture. In this paper, we develop Voxel, a fast and compiler-aware end-to-end simulation framework to facilitate exploring the efficiency of 3D-stacked AI chips for large language model (LLM) inference. Voxel enables the software/hardware co-exploration by employing a programming interface that allows ML compilers to customize the model execution plans. After validating the results of Voxel with an emulator on real silicon, we thoroughly examine the impact and correlation of different aspects of 3D-stacked AI chips, including state-of-the-art compute paradigms, tile-to-core mapping, tensor-to-bank mapping, NoC topologies and link bandwidth, DRAM bank bandwidth, per-core SRAM capacity, and energy/thermal constraints. Our findings disclose that the end-to-end efficiency of a 3D stacked AI chip not only is determined by the cooperative function of these factors, but also significantly depends on the mappings from tiles to AI core and DRAM banks. We report our findings throughout the paper, with the expectation that they will shed light on the development of the 3D-stacked AI chip ecosystem. We will open source Voxel and our study results for public research.

This paper proposes a novel algorithm for semisupervised learning. This algorithm learns graph cuts that maximize the margin with respect to the labels induced by the harmonic function solution. We motivate the approach, compare it to existing work, and prove a bound on its generalization error. The quality of our solutions is evaluated on a synthetic problem and three UCI ML repository datasets. In most cases, we outperform manifold regularization of support vector machines, which is a state-of-the-art approach to semi-supervised max-margin learning.

The Running Average Power Limit (RAPL) interface is widely used to estimate software energy consumption via CPU and DRAM counters, but tool design differences and high-frequency polling can introduce measurement overhead, namely, extra time and energy consumed by the tool itself.This paper quantifies the impact of RAPL-based tools on high-frequency (1 kHz) energy monitoring and investigates mitigation strategies. We conduct two controlled experiments: the first evaluates seven tools, including a user-space application and a kernel module developed by the authors, against a no-tool baseline, using six NAS Benchmark functions to quantify overhead. The second experiment isolates and times key functions for polling Model-Specific Registers (MSRs) (rdmsr and sys/proc_read) to estimate their execution latencies and identify potential slowdowns. The results show that existing user-space tools can introduce substantial time overhead at 1 kHz, whereas our tools significantly reduce system call overhead and inline math overhead. The time overhead of existing tools ranges from 0.25% to 46.75%. Our solutions maintain time overhead levels close to the baseline. We also find that system calls are slower than rdmsr, which in turn is slower than traditionally long-running instructions like cpuid. These findings indicate that RAPL-based energy measurement can be substantially improved by simplifying tool design and employing lower-level instructions to access RAPL values. Our findings provide guidance for practitioners on how to develop high-frequency energy profiling tools, show possible situations that can skew energy values, and demonstrate that access to RAPL values can be faster using specific techniques.

Federated Unlearning (FU) is an emerging paradigm in Federated Learning (FL) that enables participating clients to fully remove their contributions from a trained global model, driven by data protection regulations that mandate the right to be forgotten. However, existing FU methods mostly rely on synchronous coordination. This requirement forces the entire federation to halt and wait for stragglers to complete erasure, creating significant delays due to device heterogeneity. Furthermore, these methods often face the problem that the influence of erased data is merely suppressed temporarily and resurfaces during subsequent training, rather than being genuinely removed. To overcome these limitations, this paper proposes Asynchronous Federated Unlearning with Invariance Calibration (AFU-IC), a novel framework for medical imaging that decouples the erasure process from the global training workflow. This enables the target client to perform unlearning asynchronously without interrupting global training. Meanwhile, a server-side invariance calibration mechanism prevents the model from relearning the erased data. Extensive experiments on three medical benchmarks demonstrate that AFU-IC achieves unlearning efficacy and model fidelity comparable to gold-standard retraining while significantly reducing wall-clock latency compared to synchronous baselines. AFU-IC ensures efficient, compliant and reliable FL in cross-silo medical environments.

Despite being resource-intensive to train, 3D convolutional neural networks (CNNs) have been the standard approach to classify CT and MRI scans. Recent work suggests that deep multiple instance learning (MIL) may be a more efficient alternative for 3D brain scans, especially when the pre-trained image encoder used to embed each 2D slice is frozen and only the pooling operation and classifier are trained. In this paper, we provide a systematic comparison of simple MIL, attention-based MIL, 3D CNNs, and 3D ViTs across three CT and four MRI datasets, including two large datasets of at least 10,000 scans. Our goal is to help resource-constrained practitioners understand which neural networks work well for 3D neuroimages and why. We further compare design choices for attention-based MIL, including different encoders, pooling operations, and architectural orderings. We find that simple mean pooling MIL, without any learnable attention, matches or outperforms recent MIL or 3D CNN alternatives on 4 of 6 moderate-sized tasks. This baseline remains competitive on two large datasets while being 25x faster to train. To explain mean pooling's success, we examine per-slice attention quality and a semi-synthetic dataset where we can derive the best possible classifier via a Bayes estimator. This analysis reveals the limits of existing MIL approaches and suggests routes for future improvements.

Transformer-based architectures have established a dominant paradigm in global semantic perception; however, they remain fundamentally constrained by the profound spatial heterogeneity inherent in natural images. Specifically, the imposition of a uniform global receptive field across regions of varying information density inevitably leads to local feature degradation, particularly in dense conflict zones populated by microscopic targets. To address this mechanistic limitation, we propose ViCrop-Det, a training-free inference framework that introduces adaptive spatial trust region shrinkage. Inspired by the use of attention entropy in anomaly segmentation, ViCrop-Det leverages the detection decoder's cross-attention distribution as an endogenous probe. By utilizing Spatial Attention Entropy (SAE) to heuristically evaluate local spatial ambiguity, the framework executes dynamic spatial routing, allocating a fixed computational budget exclusively to regions exhibiting both high target saliency and high cognitive uncertainty. By shrinking the spatial trust region and injecting high-frequency localized observations, ViCrop-Det actively resolves spatial ambiguity and recovers fine-grained features without requiring architectural modifications. Extensive evaluations on VisDrone and DOTA-v1.5 demonstrate that ViCrop-Det yields competitive performance enhancements, consistently adding +1-3 mAP@50 to RT-DETR-R50 and Deformable DETR with a marginal 20-23\% latency overhead. On MS COCO, $AP_{S}$ improves while $AP_{M}/AP_{L}$ remains stable, indicating precise fine-scale refinement without compromising the global spatial prior. Under compute-matched settings, our adaptive routing strategy comprehensively surpasses uniform slicing baselines, achieving a highly optimized accuracy-speed trade-off.

Operating and maintaining (O&M) large-scale online engine systems (search, recommendation, advertising) demands substantial human effort for release monitoring, alert response, and root cause analysis. While LLM-based agents are a natural fit for these tasks, the deployment bottleneck is not reasoning capability but orchestration: selecting, for each operational event, the relevant data (metrics, logs, change events) and the applicable operational knowledge (handbook rules and practitioner experience). Feeding all signals indiscriminately causes dilution and hallucination, while manually curating the event-to-(data, knowledge) mapping is intractable under dozens of daily releases. We present Bian Que, an agentic framework with three contributions: (i) a \emph{unified operational paradigm} abstracting day-to-day O&M into three canonical patterns: release interception, proactive inspection, and alert root cause analysis; (ii) \emph{Flexible Skill Arrangement}, where each Skill specifies which data and knowledge to retrieve for a given business-module context and can be automatically generated and updated by LLMs or iteratively refined through natural-language instructions from on-call engineers; (iii) a \emph{unified self-evolving mechanism} in which one correction signal drives two parallel pathways, case-memory-to-knowledge distillation and targeted Skill refinement. Deployed on the e-commerce search engine of KuaiShou, the major short-video platform in China, Bian Que reduces alert volume by 75%, achieves 80% root-cause analysis accuracy, and cuts mean time to resolution by over 50%. Our framework achieves 99.0% pass rate on offline evaluations. Our code is available at https://github.com/benchen4395/BianQue_Assistant.

As hybrid qubit-oscillator algorithm development and trapped-ion hardware demonstrations advance in parallel, there is a lack of a compilation layer connecting the two at the pulse level in the vertical software stack. While qubit gate control and pulse synthesis are well-established, the translation of hybrid qubit-oscillator primitives to the pulse level has not been systematically addressed. This gap is further compounded by the inherently continuous parametric nature of such gates. Each distinct parameter value defines a physically unique operation requiring independent pulse optimization, making static pre-compilation strategies inapplicable. To fill this gap, we present HyPulse, a hardware-aware pulse synthesis and generation framework, which contributes a two-phase architecture decoupling pulse discovery from circuit assembly. An offline optimization engine populates a content-addressed cache of high-fidelity primitives: If a pulse for a given gate, parameter, and device specification already exists in the library, it is retrieved instantly; otherwise the optimizer synthesizes, hashes, and caches it automatically. An online assembler then constructs circuit-specific pulse programs ready to drive trapped-ion hardware control systems via DAX/ARTIQ (Duke) and JaqalPaw/QSCOUT (Sandia), trapped-ion pulse execution backends.

Motivated by sensing modalities in modern autonomous systems that involve hardware-constrained spatial sampling over large arrays with limited coherence time, we develop a novel framework for rapid super-resolution multi-signal direction-of-arrival (DoA) estimation based on Hankel-structured sensing and data matrix decomposition of arbitrary rank, under both the $L_2$ and $L_1$-norm formulation. The resulting $L_2$-norm estimator is shown to be maximum-likelihood optimal in white Gaussian noise. The $L_1$-norm estimator is shown to be maximum-likelihood optimal in independent, identically distributed (i.i.d.) isotropic Laplace noise, offering broad robustness to impulsive interference and corrupted measurements commonly encountered in practice. Extensive simulations demonstrate that the proposed methods exhibit powerful super-resolution capabilities, requiring significantly lower SNR and achieving substantially higher resolution probability than recent competing approaches.

Hybrid quantum--classical workflows often execute large ensembles of circuits that differ syntactically but implement identical operations, leading to substantial redundant computation. To address this, we introduce the Quantum Circuit Cache, a content-addressable system that detects semantic equivalence and reuses previously computed results across executions, backends, and workflow stages. Our approach combines ZX-calculus reduction with isomorphism-invariant Weisfeiler--Leman graph hashing to generate deterministic circuit identifiers, enabling constant-time lookup in distributed caches supporting both lightweight LMDB and scalable Redis deployments. The system integrates transparently into hybrid HPC workflows and remains backend-agnostic across CPU, GPU, and QPU environments. We evaluate the system on MareNostrum 5 with two representative workloads: distributed wire cutting and Differential Evolution-based QAOA optimization. For wire cutting, caching eliminates up to 91.98% of redundant subcircuit simulations, yielding speedups up to 7.0 times on a single node and maintaining advantages at scale, with Redis-based caching achieving up to 1.6 times speedups under high parallelism. Validation on a 35-qubit superconducting QPU confirms these benefits, achieving an 11.2 times speedup on real hardware. In distributed QAOA optimization, equivalence-aware caching avoids up to 27.6% of circuit evaluations and consistently reduces execution cost without altering the optimization algorithm. In both cases, reuse grows with concurrency and circuit structure, highlighting redundancy as a major systems bottleneck and demonstrating the effectiveness of our Quantum Circuit Cache.

We consider the problems of computing the optimal rank-$1$ Hankel and Toeplitz-structured approximation of arbitrary matrices under $L_2$ and $L_1$-norm error. Such problems arise naturally in engineered systems, including the basic few-shot signal Direction-of-Arrival (DoA) estimation problem that is of importance to modern autonomous systems applications. We develop accurate and computationally efficient structured matrix decomposition algorithms for both formulations and then derive analytically grounded small-sample-support DoA estimators for practical sensing system deployments. The resulting estimators under the $L_2$ and $L_1$ norms are formally shown to be maximum-likelihood optimal under white Gaussian and Laplace noise, respectively. The estimators are further validated through extensive simulation studies and real-world data experiments in few-shot DoA inference.

RL post-training of frontier language models is increasingly bottlenecked by autoregressive rollout generation, making rollout acceleration a central systems challenge. Many existing efficiency methods improve throughput by changing the rollout or optimization regime, for example, through off-policy execution, replay, or lower-precision generation. We study speculative decoding as a lossless acceleration primitive for RL rollouts that preserves the target model's output distribution. We implement speculative decoding in NeMo-RL with a vLLM backend, supporting both synchronous and asynchronous pipelines and enabling speculation during RL rollouts. This benefit is realizable across speculation mechanisms, such as pretrained MTP heads, small external draft models or even techniques such as Eagle3, which are traditionally applied after RL phase. This yields a deployment path for state-of-the-art speculative decoding inside RL training. In a reasoning post-training workload at 8B scale under synchronous RL, speculative decoding improves rollout throughput by 1.8x. Using a high-fidelity performance simulator, we project that combining speculative decoding with asynchronous RL yields up to 2.5x end-to-end training speedup at 235B scale.

Open-vocabulary change detection aims to identify semantic changes in bi-temporal remote sensing images without predefined categories. Recent methods combine foundation models such as SAM, DINO and CLIP, but typically process each timestamp independently or interact only at the final comparison stage. Such paradigms suffer from insufficient temporal coupling during semantic reasoning, which limits their ability to distinguish genuine semantic changes from non-semantic appearance discrepancies. In addition, patch-dominant inference on high-resolution images often weakens global semantic continuity and produces fragmented change regions. To address these issues, we propose MemOVCD, a training-free open-vocabulary change detection framework based on cross-temporal memory reasoning and global-local adaptive rectification. Specifically, we reformulate bi-temporal change detection as a two-frame tracking problem and introduce weighted bidirectional propagation to aggregate semantic evidence from both temporal directions. To stabilize memory propagation across large temporal gaps, we construct histogram-aligned transition frames to smooth abrupt appearance changes. Moreover, a global-local adaptive rectification strategy adaptively fuses local and global-view predictions, improving spatial consistency while preserving fine-grained details. Experiments on five benchmarks demonstrate that MemOVCD achieves favorable performance on two change detection tasks, validating its effectiveness and generalization under diverse open-vocabulary settings.

Parametric Retrieval-Augmented Generation (PRAG) encodes external documents into lightweight parameter modules that can be retrieved and merged at inference time, offering a promising alternative to in-context retrieval augmentation. Despite its potential, many PRAG implementations train document adapters with task-supervised objectives, which may cause each adapter to encode both document-specific facts and reusable task-solving behavior. This entanglement may make adapter composition less reliable: when multiple adapters are merged at inference time, their overlapping task behaviors can accumulate together with document-specific updates, potentially making the merged adapter less stable and less focused on the intended document knowledge. To examine this issue, we explore Orthogonal Subspace Decomposition (OSD), an adapter-training setup that separates reusable task behavior from document-specific knowledge adapters. Concretely, we first train a Task LoRA to capture reusable task behavior, and then train document LoRAs to encode document-specific knowledge in a orthogonal subspace. This setup provides a controlled way to examine how orthogonalizing task and document LoRA updates affects adapter composition in multi-document PRAG. Experiments across multiple knowledge-intensive tasks and model scales suggest that this orthogonalization strategy can improve compositional robustness in parametric RAG, especially when multiple document adapters are merged.

Accurate and consistent Emergency Severity Index (ESI) assignment remains a persistent challenge in emergency departments, where highly variable free-text triage documentation contributes to mistriage and workflow inefficiencies. This study evaluates whether open-source small language models (SLMs) can serve as reliable, privacy-preserving decision-support tools for clinical triage. We systematically compared multiple SLMs across diverse prompting pipelines and found that clinical vignettes, concise summaries of triage narratives, yielded the most accurate predictions. The SLM, Qwen2.5-7B, demonstrated the strongest balance of accuracy, stability, and computational efficiency. Through large-scale domain adaptation using expert-curated and silver-standard pediatric triage data, fine-tuned Qwen2.5-7B models substantially reduced discordance and clinically significant errors, outperforming all baseline SLMs and advanced proprietary large language models (LLMs, e.g., GPT-4o). These findings highlight the feasibility of institution-specific SLMs for reliable, privacy-preserving ESI decision support and underscore the importance of targeted fine-tuning over more complex inference strategies.

The Probabilistic Transformer (PT) establishes that the Transformer's self-attention plus its feed-forward block is mathematically equivalent to Mean-Field Variational Inference (MFVI) on a Conditional Random Field (CRF). Under this equivalence the Transformer ceases to be a black-box neural network and becomes a programmable factor graph: graph topology, factor potentials, and the message-passing schedule are all explicit and inspectable primitives that can be engineered. PT was originally developed for natural language and in this report we investigate its potential for time series. We first lift PT into the Spatial-Temporal Probabilistic Transformer (ST-PT) to repair PT's missing channel axis and weak per-step semantics, and adopt ST-PT as a shared cornerstone backbone. We then identify three distinct properties that PT/ST-PT offers as a factor-graph model and derive three Research Questions, one per property, that probe how each property can be exploited in time series: RQ1. The graph topology and potentials are direct programmable primitives. Can this be used to inject symbolic time-series priors into ST-PT through structural graph modifications, especially under data scarcity and noise? RQ2. The CRF's factor matrices are the operator's potentials. Can an external condition program these factor matrices on a per-sample basis, so that conditional generation becomes structural rather than feature-level modulation of a fixed one? RQ3. Each MFVI iteration is a Bayesian posterior update on the factor graph. Can this turn the latent transition of latent-space AutoRegressive (AR) forecasting from an opaque MLP into a principled posterior update, and can a CRF teacher distill its latents into the AR student to counter cumulative error? We give one empirical study per question. Together, these three studies position ST-PT as a programmable framework for time-series modeling.

Live future prediction refers to the task of making predictions about real-world events before they unfold. This task is increasingly studied using large language model-based agent systems, and it is important for building agents that can continually learn from real-world. Just as interactive environments have often driven progress in agents, advancing live future prediction naturally motivates viewing it as a learning environment. Prior works have explored future prediction from several different parts, but have generally not framed it as a unified learning environment. This task is appealing for learning because it can provide a large number of prediction questions grounded in diverse real-world events, while preventing answer leakage. To leverage the advantages of live future prediction, we present FutureWorld, a live agentic reinforcement learning environment that closes the training loop between prediction, outcome realization, and parameters update. In our environment, we take three open-source base models and train them for consecutive days. The results show that training is effective. Furthermore, we build a daily benchmark based on the environment and evaluate several frontier agents on it to establish performance baselines for current agent systems.

Smart contract vulnerabilities have caused billions in financial losses, raising questions about whether programming language paradigms can reduce security overhead. While imperative languages like Solidity require developers to manually implement security checks, resource-oriented languages like Move encode safety guarantees in type systems. We present a preliminary mixed-methods study analyzing 12 functionally-equivalent contract pairs implemented in both Solidity and Move by the same development team, complemented by a survey of 11 developers experienced in both languages. Quantitative analysis reveals that Move reduces explicit security overhead by 60\% (security check density: 6.7% vs. 16.8%, p=0.002, Cohen's d=-1.75) at the cost of 47% larger code size (p=0.002, d=1.90), while maintaining identical cyclomatic complexity. Developer surveys show moderate learning difficulty but higher safety confidence in Move (Median=6/7, 10 of 11 above neutral), with 55% preferring Move for security-critical applications despite ecosystem maturity gaps. These preliminary findings suggest resource-oriented paradigms shift security from runtime validation to compile-time guarantees, though adoption requires investment in learning and tooling. The controlled comparison provides initial evidence for paradigm effects on smart contract development, informing language selection decisions and identifying opportunities for improved developer resources.

Languages of the world vary concerning the order of subject, object and verb. The most frequent dominant orders are SOV and SVO, and researchers have tailored models to this fact. However, there are still languages whose dominant order does not conform to these expectations or even lack a dominant order. Here we show that across linguistic families and macroareas, word order variation within languages is shaped by the principle of swap distance minimization even when the dominant order is not SOV/SVO and even when a dominant order is lacking.

Recent advances in self-evolution video understanding frameworks have demonstrated the potential of autonomous learning without human annotations. However, existing methods often suffer from weakly controlled optimization and uncontrolled difficulty progression, as they lack structured guidance throughout the iterative learning process. To address these limitations, we propose CurEvo, a curriculum-guided self-evolution framework that introduces curriculum learning into self-evolution to achieve more structured and progressive model improvement. CurEvo dynamically regulates task difficulty, refines evaluation criteria, and balances data diversity according to model competence, forming a curriculum-guided feedback loop that aligns learning complexity with model capability. Built upon this principle, we develop a multi-dimensional adaptive QA framework that jointly evolves question generation and answer evaluation across perception, recognition, and understanding dimensions, ensuring coherent and measurable curriculum progression. Through this integration, CurEvo transforms weakly controlled self-evolution into a more structured learning process for autonomous video understanding. Across seven backbones, CurEvo consistently improves both benchmark accuracy and evaluator-based semantic score on four VideoQA benchmarks, validating the effectiveness of curriculum-guided self-evolution for video understanding.

Standard density functional theory (DFT) routinely misclassifies the electronic ground state of correlated and structurally complex compounds, predicting metallic behaviour for materials that experiments report as semiconductors. Each such mismatch encodes a specific non-ideality -- magnetic ordering, electron correlation, an alternative polymorph, or a defect -- that the calculation excluded, but extracting that signal at scale has remained a manual exercise. Here we introduce XDFT, a closed-loop agent that diagnoses the mismatch automatically: it draws candidate hypotheses from a curated catalogue, executes the corresponding first-principles tests, and updates a global Bayesian posterior over hypothesis usefulness from each verdict. On a verified benchmark of 124 materials, XDFT identifies a resolving mechanism for 70 of 90 mismatch cases (78\%), an order of magnitude above a uniform-random baseline (19\%) and a static LLM ordering (20\%). The internal posterior aligns with empirical performance over the benchmark timeline, and resolved cases collapse into a tri-partite element-class taxonomy that we distil into a four-line static rule. Each diagnosed material is returned with a corrected protocol and a mechanistic attribution; failed cases are flagged as evidence-backed targets for experimental re-examination.

We propose X-WAM, a Unified 4D World Model that unifies real-time robotic action execution and high-fidelity 4D world synthesis (video + 3D reconstruction) in a single framework, addressing the critical limitations of prior unified world models (e.g., UWM) that only model 2D pixel-space and fail to balance action efficiency and world modeling quality. To leverage the strong visual priors of pretrained video diffusion models, X-WAM imagines the future world by predicting multi-view RGB-D videos, and obtains spatial information efficiently through a lightweight structural adaptation: replicating the final few blocks of the pretrained Diffusion Transformer into a dedicated depth prediction branch for the reconstruction of future spatial information. Moreover, we propose Asynchronous Noise Sampling (ANS) to jointly optimize generation quality and action decoding efficiency. ANS applies a specialized asynchronous denoising schedule during inference, which rapidly decodes actions with fewer steps to enable efficient real-time execution, while dedicating the full sequence of steps to generate high-fidelity video. Rather than entirely decoupling the timesteps during training, ANS samples from their joint distribution to align with the inference distribution. Pretrained on over 5,800 hours of robotic data, X-WAM achieves 79.2% and 90.7% average success rate on RoboCasa and RoboTwin 2.0 benchmarks, while producing high-fidelity 4D reconstruction and generation surpassing existing methods in both visual and geometric metrics.

Skill libraries in deployed robotic systems are continually updated through fine-tuning, fresh demonstrations, or domain adaptation, yet existing typed-composition methods (BLADE, SymSkill, Generative Skill Chaining) treat the library as frozen at test time and do not analyze how composition outcomes change when a skill is replaced. We introduce a paired-sampling cross-version swap protocol on robosuite manipulation tasks to characterize this dimension of compositional skill learning. On a dual-arm peg-in-hole task we discover a dominant-skill effect: one ECM achieves 86.7% atomic success rate while every other ECM is at or below 26.7%, and whether this dominant ECM enters a composition shifts the success rate by up to +50pp. We characterize the boundary on a simpler pick task where all atomic policies saturate at 100% and the effect is undefined. Across three tasks we further find that off-policy behavioral distance metrics fail to identify the dominant ECM, ruling out the natural cheap predictor. We propose an atomic-quality probe and a Hybrid Selector combining per-skill probes (zero per-decision cost) with selective composition revalidation (full cost), and characterize its Pareto frontier on 144 skill-update decisions. On T6 the atomic-only probe sits 23pp below full revalidation (64.6% vs 87.5% oracle match) at zero per-decision cost; a Hybrid Selector with m=10 closes most of that gap to ~12pp at 46% of full-revalidation cost. On the cross-task average over 144 events, atomic-only is within 3pp of full revalidation under a mixed-oracle caveat. The atomic-quality probe is, to our knowledge, the first principled, deployment-ready primitive for skill-update governance in compositional robot policies.

Training large language models requires jointly configuring two interdependent aspects of the system: the global batch size, which governs statistical efficiency, and the 3D parallelism strategy, which governs hardware throughput. Existing approaches make these decisions independently: optimization work adapts the batch size to track the evolving critical batch size while keeping parallelism fixed, and systems work selects the fastest parallelism for a given fixed batch size without anticipating that the optimal batch size could change. We show that these decisions are tightly coupled: the throughput-optimal parallelism strategy may shift as the global batch size changes, so any method that fixes one while adapting the other operates with a suboptimal configuration for part of the training run. We present COPUS, a system that adaptively tunes the global batch size, parallelism strategy, and micro-batch size as training evolves. COPUS is guided by Goodput, the product of throughput and statistical efficiency, which models both hardware and statistical effects jointly and directly measures useful convergence per unit of wall-clock time. The system combines online gradient noise scale estimation under 3D parallelism with throughput-aware evaluation of candidate configurations, and supports efficient reconfiguration of both batch size and parallelism during training. We evaluate COPUS on LLM pre-training workloads across 1-4 nodes of 8xH100 and 8xMI210 GPUs and model sizes from 3B to 32B parameters, demonstrating average time-to-convergence speedups of 3.9-8.0% over the fastest baseline across four configurations, with peak gains up to 11.1%, including system overheads.

The rapid evolution of software services poses substantial challenges to the design and implementation of effective recommendation systems. Traditional service recommendation approaches often rely on static representations and historical usage data, which are insufficient for adapting to the dynamic and evolving nature of service ecosystems. Recently, large language models (LLMs) have shown strong potential to overcome these limitations by leveraging rich contextual understanding. However, their practical use faces two major challenges: outdated service facts and invalid or redundant services. To address these issues, we propose EVOREC, an evolution-aware framework for service recommendation that leverages model editing in a locate-then-edit paradigm to incorporate updated service facts without costly retraining efficiently. This allows the model to remain aligned with evolving service ecosystems. To address invalid service issues, we introduce a Finite Automata (FA)-based constrained decoding mechanism with deduplication, which enforces structural and semantic validity while eliminating repeated services. Experiments on real-world service datasets demonstrate that our framework consistently outperforms existing baselines, e.g., achieving an average relative improvement of 25.9% in Recall@5. Moreover, under evolving service scenarios, our approach outperforms model fine-tuning approaches by 22.3%, demonstrating strong adaptability to service evolution and providing a practical solution for service recommendation in dynamic ecosystems

We introduce a toolkit for uncovering spurious correlations between recording characteristics and target class in speech datasets. Spurious correlations may arise due to heterogeneous recording conditions, a common scenario for health-related datasets. When present both in the training and test data, these correlations result in an overestimation of the system performance -- a dangerous situation, specially in high-stakes application where systems are required to satisfy minimum performance requirements. Our toolkit implements a diagnostic method based on the detection of the target class using only the non-speech regions in the audio. Better than chance performance at this task indicates that information about the target class can be extracted from the non-speech regions, flagging the presence of spurious correlations. The toolkit is publicly available for research use.

We investigate variational quantum classifiers (VQCs) for land-cover classification from multispectral satellite imagery, adopting a feature-map perspective in which the quantum circuit defines a nonlinear data embedding while the readout determines how this representation is exploited. Using the EuroSAT-MS dataset, we perform a systematic one-vs-one evaluation across all class pairs under a controlled experimental protocol, comparing classical baselines (logistic regression, SVMs, neural networks) with VQCs employing both linear readout and quantum-kernel SVM strategies. Our results show that, while VQCs with linear readout do not outperform strong classical baselines such as RBF-SVM, the same trained quantum feature map can significantly improve performance when reused within a kernel-based decision framework. A qubit-count sweep further reveals saturation effects consistent with the mismatch between exponential Hilbert space dimension and linear parameter scaling. Overall, our findings highlight that the effectiveness of quantum models depends critically on the interplay between representation and readout, and that meaningful gains may arise from combining learned quantum feature maps with classical decision mechanisms rather than seeking direct replacement of classical models.

In the research of automated program repair (APR), benchmark datasets consisting of known defects in combination with test suites that indicate the defects are of high importance. They allow for an evidence-based comparison of different APR approaches. In our own work on APR we found significant challenges when working with widely used defect datasets, which go beyond mere repeatability of defects via test cases. We summarize these identified challenges and related lessons learned to bring them to the attention of the APR community and quantify the potential impact of them. In particular, we investigate the widely used benchmark Defects4J, which has according to Google Scholar over 1,800 citations. It consists of 835 defects from 17 open-source Java projects; a hand-curated collection of defects, test suites that clearly indicate the defect, and human patches where any unrelated changes are removed. We find that, when executing the test suites with strict requirements for reproducibility in APR settings (beyond merely reproducing the defect via test cases), 180 (21.6 %) of the defects are not suitable for evaluation experiments. Further, we find that an additional 59 (7.1 %) defects have test suites that are obviously under-specified, as deleting a single statement from the code base makes all test cases pass, although the human-written patch does not only delete code. Our contributions are: a systematic collection of requirements for defect datasets for APR beyond traditional reproducibility of defects, a description of practical experiences and quantitative analysis of problems with the Defects4J dataset, as well as an implementation of an evaluation framework for APR tools for Java programs. This evaluation framework does stricter checking for indications of inadequate test suites, to avoid otherwise unnoticed problems in the test suite, such as flaky tests.

Uncertainty estimation is essential for robust decision-making in the presence of ambiguous or out-of-distribution inputs. Gaussian Processes (GPs) are classical kernel-based models that offer principled uncertainty quantification and perform well on small- to medium-scale datasets. Alternatively, formulating the weight space learning problem under tensor network assumptions yields scalable tensor network kernel machines. However, these assumptions break Gaussianity, complicating standard probabilistic inference. This raises a fundamental question: how can tensor network kernel machines provide principled uncertainty estimates? We propose a novel Bayesian Tensor Network Kernel Machine (LA-TNKM) that employs a (linearized) Laplace approximation for Bayesian inference. A comprehensive set of numerical experiments shows that the proposed method consistently matches or surpasses Gaussian Processes and Bayesian Neural Networks (BNNs) across diverse UCI regression benchmarks, highlighting both its effectiveness and practical relevance.

Understanding how software defects manifest and evolve in production environments is critical for improving reliability. While previous research has largely focused on pre-release defects, the nature of residual faults, i.e., those escaping testing and surfacing post-release, remains poorly understood. This paper presents a large-scale characterization of pre- and post-release defects across C/C++ and Java systems, encompassing over 14k defects mined from open-source projects. We employ a broad suite of software metrics to capture diverse code attributes such as complexity, size, structure, and development history. Results show that post-release defects are concentrated in older, frequently modified, and high-churn components, typically requiring longer and more complex fixes than pre-release ones. These findings highlight that residual defects arise more from evolutionary and process dynamics than code structure alone, suggesting that reliability engineering should prioritize targeted testing in mature and complex code regions.

Trust in clinical artificial intelligence (AI) cannot be reduced to model accuracy, fluency of generation, or overall positive user impression. In medicine, trust must be engineered as a measurable system property grounded in evidence, supervision, and operational boundaries of AI autonomy. This article proposes a practical framework for trustworthy clinical AI built around three principles: evidence, supervision, and staged autonomy. Rather than replacing deterministic clinical logic wholesale with end-to-end black-box models, the proposed approach combines a deterministic core, a patient-specific AI assistant for contextual validation, a multi-tier model escalation mechanism, and a human supervision layer for verification, escalation, and risk control. We demonstrate that trust also depends on selective verification of clinically critical findings, bounded clinical context, disciplined prompt architecture, and careful evaluation on realistic cases. Classifier-driven modular prompting is examined as an incremental path to scaling clinical depth without sacrificing prompt performance and without waiting for complete rule-based coverage. To operationalize trust, a set of trust metrics is proposed, built on metrological principles -- measurement uncertainty, calibration, traceability -- enabling quantitative rather than subjective assessment of each architectural layer. In this perspective, trustworthy clinical AI emerges not as a property of an individual model, but as an architectural outcome of a system into which evidence trails, human oversight, tiered escalation, and graduated action rights are embedded from the outset.

Microservice root cause localization is fundamentally challenged by the inherent heterogeneity of cloud-native systems, which encompasses diverse observability data and multiple system entities. Existing approaches typically focus on only one aspect of heterogeneity and thus fail to capture its full diagnostic value. In this work, we systematically examine the multifaceted role of heterogeneity within both microservice systems and the RCL process. This analysis motivates a deeper investigation into how entity-level distinctions and their asymmetric dependencies influence fault behavior. Our empirical analysis of two microservice benchmarks reveals that entity-level heterogeneity naturally gives rise to heterogeneous fault propagation, which is highly asymmetric and dominated by cross-layer interactions between services and hosts. In light of this, we propose NexusRCL, a semi-supervised framework that internalizes these propagation patterns by formalizing services and hosts as distinct node types within a heterogeneous graph. This design, coupled with an event-based abstraction mechanism, allows NexusRCL to effectively capture both data level and entity-level heterogeneity while minimizing labeling costs through active learning. Comprehensive evaluations on two industrial benchmark datasets demonstrate NexusRCL's superior performance, achieving improvements of up to 49.85\% in Top-1 accuracy (A@1) and 32.70\% in Average Top-5 accuracy (A@5) compared to state-of-the-art methods.

Python's dynamic nature complicates testing and increases the possibility that some defects evade detection, so an effective fault prediction becomes essential. We examine whether post-release faults can be predicted using modern ML and DL. Using a balanced dataset of over 4,000 labeled faults with 83 product, process, statistical, and Python-specific metrics plus normalized code representations, we conduct cross-project experiments. LLMs and unsupervised models fail to distinguish residual from non-residual faults, while supervised metric-based models (RandomForest, XGBoost, CatBoost) perform far better, yielding a 0.85-0.9 recall and cutting false negatives by an order of magnitude. Process metrics, especially age, churn, and developer-activity, alongside class and file size, consistently prove most predictive. Notably, the Principal Component Analysis shows that metrics and code embeddings occupy distinct regions of the representation space, suggesting that they capture complementary rather than redundant information.

Deep learning compilers and vendor libraries deliver strong baseline performance but are bounded by finite, engineer-curated catalogs. When these omit needed optimizations, practitioners substitute hand-written CUDA or CUTLASS, demanding expertise in GPU microarchitecture and C++ template metaprogramming. Recent LLM-based agents target kernel generation in raw CUDA, forcing rediscovery of optimizations already encoded in mature libraries. We present FACT (Framework for Agentic CUTLASS Transpilation), a framework that employs a three-stage, agent-driven workflow optimizing PyTorch modules through multi-pattern composition while grounding synthesis in CUTLASS C++. (1) Pattern discovery: an LLM agent inspects the traced graph, matches subgraphs to optimization rules, retrieves vetted examples from an architecture-specific index, and outputs prioritized patterns. (2) Pattern realization: each pattern is implemented as a CUTLASS kernel wrapped in a PyTorch extension, verified, and auto-tuned by sweeping parameters inferred from the CUTLASS hierarchy. (3) Pattern composition: extensions are loaded together into a single composed module for end-to-end benchmarking. We evaluate the workflow using KernelBench's evaluation framework and provided modules on an NVIDIA A100. On Level 1, we apply the workflow to three GEMM workloads (square matrix multiply, batched matrix multiply, and large-$K$ matrix multiply). Auto-tuned CUTLASS kernels improve over PyTorch cuBLAS baseline by $1.06\times$--$1.18\times$. On Level 3 MiniGPT block, composing fused multi-head attention with fused MLP GEMM+GELU yields $2.79\times$ end-to-end speedup. Our work couples agentic graph-level pattern discovery with auto-tuning and a dynamic pattern table, offering a practical path from traced PyTorch to deployable kernels by automating CUTLASS kernel synthesis and auto-tuning.

Differential Privacy (DP) for text matured from disjointed word-level substitutions to contiguous sentence-level rewriting by leveraging the generative capacity of language models. While this form of text privatization is best suited for balancing formal privacy guarantees with grammatical coherence, its impact on the register identity of text remains largely unexplored. By conducting a multidimensional stylistic profiling of differentially-private rewriting, we demonstrate that the cost of privacy extends far beyond lexical variation. Specifically, we find that rewriting under privacy constraints induces a systematic functional mutation of the text's communicative signature. This shift is characterized by the severe attrition of interactive markers, contextual references, and complex subordination. By comparing autoregressive paraphrasing against bidirectional substitution across a spectrum of privacy budgets, we observe that both architectures force convergence toward a non-involved and non-persuasive register. This register-blind sanitization effectively preserves semantic content but structurally homogenizes the nuanced stylistic markers that define human-authored discourse.

In the rapidly evolving field of software engineering, the skills required of graduates entering the job market are constantly changing. Several studies have identified a gap between the skills taught in university curricula and those demanded by the software engineering industry. This chapter investigates the technical skill and expertise gap between higher education institutions (HEIs) and the UK software engineering industry by mapping job descriptions to the skills included in computer science degree programmes. A custom web scraping and text analysis tool, utilising fuzzy matching, was developed to extract and categorise skills from 300 job postings and undergraduate curricula from 30 UK universities. The analysis showed that the curricula place a strong emphasis on Programming Languages (18%) and Database Management (12.83%). In contrast, the industry s most frequently requested skill category is Software Design and Planning, which appears in approximately 88.68% of job descriptions, highlighting its critical importance. General Programming Language and System Structures also show strong demand, present in over 78.30% and 66.04% of postings, respectively. The mapping indicates that areas such as System Structures and Software Domains are significantly underrepresented in curricula, while Database Management and Compiler Design may be overemphasised. These insights can support HEIs in aligning their programmes with industry needs, supporting the preparation of graduates for dynamic careers in software engineering.

This paper presents an application of the biologically realistic JASTAP neural network model to classification tasks. The JASTAP neural network model is presented as an alternative to the basic multi-layer perceptron model. An evolutionary procedure previously applied to the simultaneous solution of feature selection and neural network training on standard multi-layer perceptrons is extended with JASTAP model. Preliminary results on IRIS standard data set give evidence that this extension allows the use of smaller neural networks that can handle noisier data without any degradation in classification accuracy.

With the increasing availability of online information, recommender systems have become an important tool for many web-based systems. Due to the continuous aspect of recommendation environments, these systems increasingly rely on contextual multi-armed bandits (CMAB) to deliver personalized and real-time suggestions. A critical yet underexplored component in these systems is the representation of user state, which typically encapsulates the user's interaction history and is deeply correlated with the model's decisions and learning. In this paper, we investigate the impact of different embedding-based state representations derived from matrix factorization models on the performance of traditional CMAB algorithms. Our large-scale experiments reveal that variations in state representation can lead to improvements greater than those achieved by changing the bandit algorithm itself. Furthermore, no single embedding or aggregation strategy consistently dominates across datasets, underscoring the need for domain-specific evaluation. These results expose a substantial gap in the literature and emphasize that advancing bandit-based recommender systems requires a holistic approach that prioritizes embedding quality and state construction alongside algorithmic innovation. The source code for our experiments is publicly available on https://github.com/UFSCar-LaSID/bandits_blind_spot.

Large reasoning models such as DeepSeek-R1 and OpenAI o1 generate extended chains of thought spanning thousands of tokens, yet their integration with retrieval-augmented generation (RAG) remains fundamentally misaligned. Current RAG systems optimize for providing context before reasoning begins, while reasoning models require evidence injection during multi-step inference chains. We introduce ReaLM-Retrieve, a reasoning-aware retrieval framework that addresses this mismatch through three key innovations: (1) a step-level uncertainty detector that identifies knowledge gaps at reasoning-step granularity rather than token or sentence level; (2) a retrieval intervention policy that learns when external evidence maximally benefits ongoing reasoning; and (3) an efficiency-optimized integration mechanism that reduces per-retrieval overhead by 3.2x compared to naive integration. Experiments on MuSiQue, HotpotQA, and 2WikiMultiHopQA demonstrate that ReaLM-Retrieve achieves on average 10.1% absolute improvement in answer F1 over standard RAG (range: 9.0-11.8% across the three benchmarks) while reducing retrieval calls by 47% compared to fixed-interval approaches like IRCoT (all improvements significant at p<0.01, paired bootstrap). On the challenging MuSiQue benchmark requiring 2-4 hop reasoning, our method achieves 71.2% F1 with an average of only 1.8 retrieval calls per question. Analysis shows that ReaLM-Retrieve also improves retrieval quality itself, achieving 81.3% Recall@5 with consistently higher precision and MRR than fixed-interval baselines on supporting evidence, establishing new state-of-the-art efficiency-accuracy trade-offs for reasoning-intensive retrieval tasks.

AI-for-Science (AI4Science) is increasingly transforming scientific discovery by embedding machine learning models into prediction, simulation, and hypothesis generation workflows across domains. However, the effectiveness of these models is fundamentally constrained by the AI-readiness of scientific data, for which no scalable and systematic evaluation mechanism currently exists. In this work, we propose SciHorizon-DataEVA, a novel agentic system to scalable AI-readiness evaluation of heterogeneous scientific data. At the evaluation-criteria level, we introduce the Sci-TQA2 principles, which organize AI-readiness into four complementary dimensions: Governance Trustworthiness, Data Quality, AI Compatibility, and Scientific Adaptability. Each dimension is decomposed into measurable atomic elements that enable fine-grained and executable assessment. To operationalize these principles at scale, we develop Sci-TQA2-Eval, a hierarchical multi-agent evaluation approach orchestrated through a directed, cyclic workflow. Our Sci-TQA2-Eval dynamically constructs dataset-aware evaluation specifications by combining lightweight dataset profiling, applicability-aware metric activation, and knowledge-augmented planning grounded in domain constraints and dataset-paper signals. These specifications are executed through an adaptive, tool-centric evaluation mechanism with built-in verification and self-correction, enabling scalable and reliable assessment across heterogeneous scientific data. Extensive experiments on scientific datasets spanning multiple domains demonstrate the effectiveness and generality of SciHorizon-DataEVA for principled AI-readiness evaluation.

Large Reasoning Models (LRMs) achieve strong performance on mathematical reasoning tasks but remain unreliable on challenging instances. Existing test-time scaling methods, such as repeated sampling, self-correction, and tree search, improve performance at the cost of increased computation, yet often exhibit diminishing returns on hard problems. We observe that output disagreement is strongly correlated with instance difficulty and prediction correctness, providing a useful signal for guiding instance-level strategy selection at test time. Based on this insight, we propose a training-free framework that formulates test-time scaling as an instance-level routing problem, rather than allocating more computation within a single strategy, dynamically selecting among different scaling strategies based on output disagreement. The framework applies lightweight resolution for consistent cases, majority voting for moderate disagreement, and rewriting-based reformulation for highly ambiguous instances. Experiments on seven mathematical benchmarks and three models show that our method improves accuracy by 3% - 7% while reducing sampling cost compared to existing approaches.

Annotating long-horizon robotic demonstrations with precise temporal action boundaries is crucial for training and evaluating action segmentation and manipulation policy learning methods. Existing annotation tools, however, are often limited: they are designed primarily for vision-only data, do not natively support synchronized visualization of robot-specific time-series signals (e.g., gripper state or force/torque), or require substantial effort to adapt to different dataset formats. In this paper, we introduce ATLAS, an annotation tool tailored for long-horizon robotic action segmentation. ATLAS provides time-synchronized visualization of multi-modal robotic data, including multi-view video and proprioceptive signals, and supports annotation of action boundaries, action labels, and task outcomes. The tool natively handles widely used robotics dataset formats such as ROS bags and the Reinforcement Learning Dataset (RLDS) format, and provides direct support for specific datasets such as REASSEMBLE. ATLAS can be easily extended to new formats via a modular dataset abstraction layer. Its keyboard-centric interface minimizes annotation effort and improves efficiency. In experiments on a contact-rich assembly task, ATLAS reduced the average per-action annotation time by at least 6% compared to ELAN, while the inclusion of time-series data improved temporal alignment with expert annotations by more than 2.8% and decreased boundary error fivefold compared to vision-only annotation tools.

Norway's electricity market is heavily dominated by hydropower, but the 2021--2022 energy crisis and stronger integration with Continental Europe have fundamentally altered price formation, reducing the reliability of forecasting models calibrated on historical data. Despite the critical need for updated models, a unified benchmark evaluating feature contributions across all structurally diverse Norwegian bidding zones remains lacking. Here we present a comprehensive evaluation of electricity price forecasting across all five Norwegian Nord Pool bidding zones. We constructed a multimodal hourly dataset spanning 2019--2025 and evaluated eight forecasting model families including LightGBM, ARX, and advanced deep learning architectures using a strictly causal test set. We implemented robust rolling-origin backtesting, leave-one-group-out feature ablation, and conditional regime analysis to dissect model performance and feature utility. Our results show that LightGBM achieves the best performance in every zone with MAE ranging from 1.64 to 5.74~EUR/MWh, while the ridge ARX model remains a highly competitive linear benchmark in northern zones. Feature ablation reveals that models relying solely on lagged prices and calendar variables achieve high accuracy and often match or exceed full multimodal integration. However, conditional regime analysis demonstrates that external features like reservoir levels and gas prices remain crucial to stratify forecast errors, which consistently increase under stressed market regimes. This highlights the practical value of model interpretability and regime awareness for decision makers facing structural changes in market dynamics.

The bottleneck in learning-based industrial defect detection is often limited not by model capacity, but by the scarcity of labeled defect data: defects are rare, annotations are expensive, and collecting balanced training sets is slow. We present an end-to-end pipeline for synthetic defect generation and annotation, combining Vision-Language-Model-based prompts, LoRA-adapted diffusion, mask-guided inpainting, and sample filtering with automatic label derivation, and demonstrates the potential of real data with realistic synthetic samples to overcome data scarcity. The evaluation is conducted on, a challenging dataset of pitting defects on ball screw drives, and then on a subset of the Mobile phone screen surface defect segmentation dataset (MSD) dataset to test cross-domain transfer. Beyond downstream detector performance, we analyze key stages of the pipeline, including prompt construction, LoRA selection, and sample filtering with DreamSim and CLIPScore, to understand which synthetic samples are both realistic and useful. Experiments with YOLOv26, YOLOX, and LW-DETR show that synthetic-only training does not replace real data. When combined with real data, synthetic defects can preserve performance and yield modest gains in selected BSData training regimes. The MSD transfer study shows that the overall pipeline structure carries over to a second industrial inspection domain, while also highlighting the importance of domain-specific adaptation and annotation-quality control. Overall, the paper provides an end-to-end assessment of diffusion-based industrial defect synthesis and shows that its strongest value lies in strengthening scarce real datasets rather than substituting for them.

We propose a dual-channel reservoir-computing scheme for inferring the dynamics of two distinct chaotic systems with a single machine. By augmenting a standard reservoir with a system-label channel and a parameter-control channel, the machine can be trained from time series collected from a few sampled states of the two systems. We show that the trained machine not only predicts the short-time evolution of the sampled states, but also reproduces the long-term statistical properties of unseen states, thereby enabling reconstruction of the bifurcation diagrams of both systems from partial observations. The effectiveness of the scheme is demonstrated for the Lorenz and Rössler systems in numerical simulations and for the Chua and Rossler circuits in experiments. Functional-network analysis further shows that the two target systems are encoded by distinct dynamical patterns in the reservoir. These results extend multifunctional and parameter-aware reservoir computing, and provide a route to data-driven inference of multiple nonlinear systems using a single machine.

Effective mental health counseling is a complex, theory-driven process requiring the simultaneous integration of psychological frameworks, real-time distress signals, and strategic intervention planning. This level of clinical reasoning is critical for safety and therapeutic effectiveness but is often missing in general-purpose Large Language Models (LLMs). We introduce SAGE (Strategy-Aware Graph-Enhanced), a novel framework designed to bridge the gap between structured clinical knowledge and generative AI. SAGE constructs a heterogeneous graph that unifies conversational dynamics with a psychologically grounded layer, explicitly anchoring interactions in a theory-driven lexicon. Our architecture first employs a Next Strategy Classifier to identify the optimal therapeutic intervention. Subsequently, a Graph-Aware Attention mechanism projects graph-derived structural signals into soft prompts, conditioning the LLM to generate responses that maintain clinical depth. Validated through both automated metrics and expert human evaluation, SAGE outperforms baselines in strategy prediction and recommended response quality. By providing actionable intervention recommendations, SAGE serves as a cutting-edge decision-support tool designed to augment human expertise in high-stakes crisis counseling.

Process malleability has proved to have a highly positive impact on the resource utilization and global productivity in data centers compared with the conventional static resource allocation policy. However, the non-negligible additional development effort this solution imposes has constrained its adoption by the scientific programming community. In this work, we present DMRlib, a library designed to offer the global advantages of process malleability while providing a minimalist MPI-like syntax. The library includes a series of predefined communication patterns that greatly ease the development of malleable applications. In addition, we deploy several scenarios to demonstrate the positive impact of process malleability featuring different scalability patterns. Concretely, we study two job submission modes (rigid and moldable) in order to identify the best-case scenarios for malleability using metrics such as resource allocation rate, completed jobs per second, and energy consumption. The experiments prove that our elastic approach may improve global throughput by a factor higher than 3x compared to the traditional workloads of non-malleable jobs.

Autonomous LLM agents increasingly operate in long-horizon, interactive settings where success depends on reusing experience accumulated over extended histories. However, existing agent memory systems are fundamentally constrained by text-context budgets: storing or revisiting raw trajectories is prohibitively token-expensive, while summarization and text-only retrieval trade token savings for information loss and fragmented evidence. To address this limitation, we propose Optical Context Retrieval Memory (OCR-Memory), a memory framework that leverages the visual modality as a high-density representation of agent experience, enabling retention of arbitrarily long histories with minimal prompt overhead at retrieval time. Specifically, OCR-Memory renders historical trajectories into images annotated with unique visual identifiers. OCR-Memory retrieves stored experience via a \emph{locate-and-transcribe} paradigm that selects relevant regions through visual anchors and retrieves the corresponding verbatim text, avoiding free-form generation and reducing hallucination. Experiments on long-horizon agent benchmarks show consistent gains under strict context limits, demonstrating that optical encoding increases effective memory capacity while preserving faithful evidence recovery.

Aspect-based Sentiment Analysis (ABSA) extracts fine-grained opinions toward specific aspects within text but remains largely English-focused despite major advances in transformer-based and instruction-tuned models. This work presents a multilingual evaluation of state-of-the-art ABSA approaches across seven languages (English, German, French, Dutch, Russian, Spanish, and Czech) and four subtasks (ACD, ACSA, TASD, ASQP). We systematically compare different transformer architectures under zero-resource, data-only, and full-resource settings, using cross-lingual transfer, code-switching and machine translation. Fine-tuned Large Language Models (LLMs) achieve the highest overall scores, particularly in complex generative tasks, while few-shot counterparts approach this performance in simpler setups, where smaller encoder models also remain competitive. Cross-lingual training on multiple non-target languages yields the strongest transfer for fine-tuned LLMs, while smaller encoder or seq-to-seq models benefit most from code-switching, highlighting architecture-specific strategies for multilingual ABSA. We further contribute two new German datasets, an adapted GERestaurant and the first German ASQP dataset (GERest), to encourage multilingual ABSA research beyond English.

Large language models (LLMs) accelerate software development but often exhibit instability, non-determinism, and weak adherence to development discipline in unconstrained workflows. While test-driven development (TDD) provides a structured Red-Green-Refactor process, existing LLM-based approaches typically use tests as auxiliary inputs rather than enforceable process constraints. We present an AI-native TDD framework that operationalizes classical TDD principles as structured prompt-level and workflow-level governance mechanisms. Extracted principles are formalized in a machine-readable manifesto and distributed across planning, generation, repair, and validation stages within a layered architecture that separates model proposal from deterministic engine authority. The system enforces phase ordering, bounded repair loops, validation gates, and atomic mutation control to improve stability and reproducibility. We describe architecture and discuss encoding software engineering discipline directly into prompt orchestration, which we think offers a promising direction for reliable LLM-assisted development.

Coverage criteria play a central role in assessing test adequacy in classical software, yet their effectiveness for quantum programs remains poorly understood and largely unexplored. In this paper, we propose six quantum-tailored criteria - condition, decision, and path coverage, and their probabilistic variants - adapted from their classical counterparts. We present QaCoCo, a tool that computes these criteria for circuit-based quantum programs. We empirically evaluate these criteria on a large and diverse set of 540 circuits and analyze the coverage achieved. Our results show that while circuits frequently achieve high condition and decision coverage (97.56% and 97.63%, on average), path coverage remains limited (71.84%), particularly in the presence of multi-controlled gates, which induce extreme path explosion and coverage imbalance. Moreover, to account for the probabilistic nature of quantum circuits, we introduce probabilistic coverage, which augments structural coverage with a confidence measure (88.87%, 88.65%, and 37.18% for condition, decision, and path coverage, respectively, on average). Finally, through mutation testing, we find weak or no correlation between fault detection and structural coverage, consistent with observations in classical computing.

As Competency-Based Education (CBE) is gaining traction around the world, the shift from marks-based assessment to qualitative competency mapping is a manual challenge for educators. This paper tackles the bottleneck issue by suggesting a "Human-in-the-Loop" benchmarking framework to assess the effectiveness of multiple LLMs in automating secondary-level mathematics assessment. Based on the Grade 10 Optional Mathematics curriculum in Nepal, we created a multi-dimensional rubric for four topics and four cross-cutting competencies: Comprehension, Knowledge, Operational Fluency, and Behavior and Correlation. The multi-provider ensemble, consisted of open-weight models -- Eagle (Llama 3.1-8B) and Orion (Llama 3.3-70B) -- and proprietary frontier models Nova (Gemini 2.5 Flash) and Lyra (Gemini 3 Pro), was benchmarked against a ground truth defined by two senior mathematics faculty members (kappa_w = 0.8652). The findings show a marked "Architecture-compatibility gap". Although the Gemini-based Mixture-of-Experts (Sparse MoE) models achieved "Fair Agreement" (kappa_w ~ 0.38), the larger Orion (70B) model exhibited "No Agreement" (kappa_w = -0.0261), suggesting that architectural compliance with instruction constraints outweighs the scale of raw parameters in rubric-constrained tasks. We conclude that while LLMs are not yet suitable for autonomous certification, they provide high-value assistive support for preliminary evidence extraction within a "Human-in-the-Loop" framework.

Federated learning (FL) trains a shared model from updates contributed by distributed clients, often implicitly assuming that contributing clients are representative of the target population. In practice, this representativeness assumption can fail at two distinct stages, inducing selection bias. First, eligibility rules such as device constraints, software requirements, or user consent determine which clients are ever enrolled and reachable for training, inducing \emph{enrollment bias}. Second, among enrolled clients, user and system factors such as battery state, network status, and local time determine which clients participate in each communication round, inducing \emph{participation bias}. Although existing work has largely addressed round-level participation bias, it has paid far less attention to population-level enrollment bias, which can induce a persistent mismatch between the training objective and the target-population objective. We formalize FL under a two-stage selection model and derive \textsc{FedIPW}, an inverse-probability-weighted aggregation scheme that recovers the target-population mean update under standard ignorability and positivity assumptions. Because client-level covariates are often unavailable for non-enrolled clients, we also introduce a limited-information aggregate-calibration extension that uses known target-population summaries to reweight the enrolled sample, partially correcting enrollment bias. We further provide an algorithm-agnostic optimization analysis under residual weighting error and show that incomplete selection correction can induce a non-vanishing bias floor. Finally, experiments on synthetic federated logistic regression validate the predicted objective mismatch and show that enrollment correction reduces target-population error under two-stage selection.

Timely and reliable multilingual communication is critical during natural and human-induced disasters, but developing effective solutions for crisis communication is limited by the scarcity of curated parallel data. We propose a domain-adaptive pipeline that expands a small reference corpus, by retrieving and filtering data from general corpora. We use the resulting dataset to fine-tune a small language model for crisis-domain translation and then apply preference optimization to bias outputs toward CEFR A2-level English. Automatic and human evaluation shows that this approach improves readability, while maintaining strong adequacy. Our results indicate that simplified English, combined with domain adaptation, can function as a practical lingua franca for emergency communication when full multilingual coverage is not feasible.

Digital twins provide a powerful paradigm for diagnostic and prognostic tasks in the monitoring and control of engineered systems; however, their deployment for complex structures remains challenged by model-form uncertainty, arising from unknown nonlinear dynamics, and by sparse sensing. These limitations hinder reliable online state estimation using either purely physics-based or purely data-driven approaches. This work introduces the Physics-Guided Graph Neural ODE (PiGGO) framework, a physics-informed, graph-based Bayesian state estimation approach in which a learned graph neural ordinary differential equation (GNODE) serves as the continuous-time state-transition model within an extended Kalman filter. The graph representation explicitly defines the system state-space, while physics-guided inductive biases encode known structural relationships and constrain the learning of nonlinear dynamics. By integrating graph-native learned dynamics with recursive Bayesian filtering, the proposed PiGGO framework enables online virtual sensing and uncertainty-aware state estimation for nonlinear systems with unknown model form, while maintaining generalisation across topologically similar structures. Numerical case studies demonstrate improved robustness to model uncertainty and measurement noise, outperforming both open-loop graph neural models and conventional filtering approaches in online prediction tasks.

Technology mapping is a critical yet challenging stage in logic synthesis. While Large Language Models (LLMs) have been applied to generate optimization scripts, their potential for core algorithm enhancement remains untapped. We introduce MappingEvolve, an open-source framework that pioneers the use of LLMs to directly evolve technology mapping code. Our method abstracts the mapping process into distinct optimization operators and employs a hierarchical agent-based architecture, comprising a Planner, Evolver, and Evaluator, to guide the evolutionary search. This structured approach enables strategic and effective code modifications. Experiments show our method significantly outperforms direct evolution and strong baselines, achieving 10.04\% area reduction versus ABC and 7.93\% versus mockturtle, with 46.6\%--96.0\% $S_{overall}$ improvement on EPFL benchmarks, while explicitly navigating the area--delay trade-off. Our code and data are available at https://github.com/Flians/MappingEvolve.

Context: Large language models (LLMs) are observed to have a significant positive impact on various software engineering (SE) activities. With improved accessibility, the adoption of powerful LLMs in industry has surged recently. However, there is a lack of actionable best practices for the efficient and responsible adoption of LLMs within industrial software settings. Objectives: We developed seven actionable recommendations to address this research gap. Methods: We conducted a multi-case study with three organisations that use LLMs within their SE activities and synthesised seven recommendations through qualitative thematic analysis. We conducted a complementary online survey with software practitioners from various industries to evaluate the perceived relevance of our recommendations. Results: Our results and recommendations focus on (i) users' preference to use LLMs as AI assistants, (ii) the importance of relevant stakeholders' satisfaction in the LLM-output evaluation, (iii) scoping the applicability of LLMs within SE tasks, (iv) the effect of LLMs on SE workflows, (v) the necessity and directions for developing human oversight mechanisms, and (vi) the necessary skills for practitioners for leveraging LLMs within SE. The online survey indicates a high level of agreement from the participants regarding the perceived relevance of the recommendations. Conclusion: We outline future research directions, including mapping the seven recommendations to the principles of the EU AI Act (AIA) in order to examine how they relate to the current regulatory compliance frameworks.

As the size of Deep Neural Networks (DNNs) increases dramatically to achieve high accuracy, the DNNs require a large amount of computations and memory footprint. Pruning, which produces a sparse neural network, is one of the solutions to reduce the computational complexity of neural network processing. To maximize the performance of the computations with such compressed data, dedicated sparse neural network accelerators have been introduced, but complex circuits for matching the indices of non-zero inputs/weights cause large overhead in area and power of processing elements (PEs). The sparse PE becomes significantly larger than the dense PE, which raises an interesting question for designers; "Given the area, isn't it better to use larger number of dense PEs despite the low utilization in sparse matrix computations?" In this paper, we show that the answer is "yes", and demonstrate an area and energy-efficient method for sparse neural network computing on dense-matrix multiplication hardware accelerators (Sparse-on-Dense).

Reliable celestial attitude determination is a critical requirement for autonomous spacecraft navigation, yet traditional "Lost-in-Space" (LIS) algorithms often suffer from high computational overhead and sensitivity to sensor-induced noise. While deep learning has emerged as a promising alternative, standard regression models are often confounded by the non-Euclidean topology of the celestial sphere and by the periodic boundary conditions of Right Ascension (RA) and Declination (Dec). In this paper, we present Star-Fusion, a multi-modal architecture that reformulates orientation estimation as a discrete topological classification task. Our approach leverages spherical K-Means clustering to partition the celestial sphere into K topologically consistent regions, effectively mitigating coordinate wrapping artifacts. The proposed architecture employs a tripartite fusion strategy: a SwinV2-Tiny transformer backbone for photometric feature extraction, a convolutional heatmap branch for spatial grounding, and a coordinate-based MLP for geometric anchoring. Experimental evaluations on a synthetic Hipparcos-derived dataset demonstrate that Star-Fusion achieves a Top-1 accuracy of 93.4% and a Top-3 accuracy of 97.8%. Furthermore, the model exhibits high computational efficiency, maintaining an inference latency of 18.4 ms on resource-constrained COTS hardware, making it a viable candidate for real-time onboard deployment in next-generation satellite constellations.

Reusing verification artefacts requires identifying structural and semantic similarities across programs and their specifications. In this paper, we focus on graph construction as a foundational step toward this goal. We present a pipeline that converts imperative programs and their annotations into typed, attributed graphs. Our experiments cover datasets including C with ACSL, Java with JML, and Dafny for C\#. The pipeline integrates abstract syntax tree parsing with semantic embeddings derived from models such as SentenceTransformer and CodeBERT. This enables the generation of graph representations that capture both structural relationships and semantic context. Our results show that consistent graph representations can be constructed across different languages and annotation styles. This work provides a practical basis for future steps in semantic enrichment and approximate graph matching for scalable verification artefact reuse.

Large language models (LLMs) are increasingly considered for deployment as the control component of robotic health attendants, yet their safety in this context remains poorly characterized. We introduce a dataset of 270 harmful instructions spanning nine prohibited behavior categories grounded in the American Medical Association Principles of Medical Ethics, and use it to evaluate 72 LLMs in a simulation environment based on the Robotic Health Attendant framework. The mean violation rate across all models was 54.4\%, with more than half exceeding 50\%, and violation rates varied substantially across behavior categories, with superficially plausible instructions such as device manipulation and emergency delay proving harder to refuse than overtly destructive ones. Model size and release date were the primary determinants of safety performance among open-weight models, and proprietary models were substantially safer than open-weight counterparts (median 23.7\% versus 72.8\%). Medical domain fine-tuning conferred no significant overall safety benefit, and a prompt-based defense strategy produced only a modest reduction in violation rates among the least safe models, leaving absolute violation rates at levels that would preclude safe clinical deployment. These findings demonstrate that safety evaluation must be treated as a first-class criterion in the development and deployment of LLMs for robotic health attendants.

Dynamic Resource Management (DRM) techniques can be leveraged to maximize throughput and resource utilization in computational clusters. Although DRM has been extensively studied through analytical workloads and simulations, skepticism persists among end administrators and users regarding their feasibility under real-world conditions. To address this problem, we propose a novel methodology for validating DRM techniques, such as malleability, in realistic scenarios that reproduce actual cluster conditions of jobs and users by replaying workload logs on a High-performance Computing (HPC) infrastructure. Our methodology is capable of adapting the workload to the target cluster. We evaluate our methodology in a malleability-enabled 125-node partition of the Marenostrum 5 supercomputer. Our results validate the proposed method and assess the benefits of MPI malleability on a novel use case of a pioneer user of malleability (our "PhD Student"): parallel efficiency-aware malleability reduced a malleable workload time by 27% without delaying the baseline workload, although introducing queueing delays for individual jobs, but maintaining the resource utilization rate.

Improving large language model (LLM) reasoning requires supervision that is both aligned with the model's own test-time states and informative at the token level. Reinforcement learning with verifiable rewards provides on-policy exploration but offers sparse, high-variance credit; supervised fine-tuning and distillation provide dense targets but often train on fixed trajectories or rely on stronger teachers. Recent privileged on-policy self-distillation explores a middle ground by scoring student rollouts with the same model under verified solution context. We revisit this setting through a contextual re-scoring lens: for reasoning, the important choices are not only whether privileged context is available, but how much of it should be revealed and where its distribution should shape the student. We propose PAINT (Partial-solution Adaptive INterpolated Training), which masks the verified solution according to rollout-reference overlap and applies a small energy-space interpolation on a sparse set of entropy-mismatch token positions. Across competition-level math benchmarks, PAINT consistently improves over a strong prior on-policy self-distillation baseline at all three Qwen3 scales. On Qwen3-8B, it raises macro Avg@12 by 2.1 points over this prior baseline and 2.9 points over GRPO.

This paper combines methods from the fields of Model-Based Testing (MBT) and Behaviour-Driven Development (BDD) to define a testing approach with human-readable specifications and test cases, as in BDD, while using the modelling techniques and automatic test generation algorithms from MBT. We introduce PICKLES, a Precise Input and Control-flow Keyword-based Language for tEst Scenarios; an extension of Gherkin-style BDD scenarios, specified in structured natural language. We provide a bi-directional translation from Pickles scenarios to formal models, where both specifications and tests are human-readable, and a method to obtain a so-called master model combining all translated scenarios. Standard MBT algorithms can then be applied to automatically derive test cases from it. We implement a prototype of the translation and composition steps, which we use on an industrial case study: a software component from a traffic management system. With it, we illustrate the pipeline and show how Pickles can achieve significantly higher coverage with respect to BDD from the same set of scenarios.

Municipal solid waste incineration is increasingly central to urban waste management, yet its sustainability benefit depends on controlling carbon emissions and multiple air pollutants under highly heterogeneous operating conditions. Current data-driven models are often accurate within individual plants but are difficult to transfer across facilities, limiting their value for scalable emission-control strategies. Here we show that multi-site emission behaviour can be represented through transferable system-level structures when physical constraints, operating-regime heterogeneity and carbon--pollutant coupling are jointly considered. We develop a physics-informed transfer learning framework built on a carbon--pollutant mixture-of-experts model, which combines regime-dependent expert routing with conservation-based regularization and a carbon--pollutant synergistic index for integrated risk evaluation. Across 13 municipal solid waste incineration plants, the model captured both pollutant-specific emissions and system-level risk, achieving source-domain average pollutant $R^2$ values of 0.668--0.904 and CPSI $R^2$ values of 0.666--0.970. After transfer from a reference facility to 12 target plants, average pollutant $R^2$ remained between 0.661 and 0.842, while CPSI retained comparable transferability ($R^2$ = 0.610--0.841). Expert-utilization patterns further indicate that adaptation occurs through structured re-weighting of operating regimes rather than complete model re-learning. By extending the learned representation into an interpretable digital twin, this framework provides a route from emission prediction to regime-aware operational navigation, supporting scalable carbon--pollutant synergistic control across heterogeneous waste-to-energy systems.

Speech Sound Disorders (SSD) affect roughly five percent of children, yet speech-language pathologists face severe staffing shortages and unmanageable caseloads. We test a hierarchical approach to SSD classification on the granular multi-task SLPHelmUltraSuitePlus benchmark. We propose a cascading approach from binary classification to type, and symptom classification. By fine-tuning Speech Representation Models (SRM), and using targeted data augmentation we mitigate biases found by previous works, and improve upon all clinical tasks in the benchmark. We also treat Automatic Speech Recognition (ASR) with our data augmentation approach. Our results demonstrate that SRM consistently outperform the LLM-based state-of-the-art across all evaluated tasks by a large margin. We publish our models and code to foster future research.

Electric truck operations require routing decisions that remain feasible under limited battery range, long charging times, travel and energy consumption, and competition for shared charging infrastructure. These features make electric truck routing a coupled logistics and energy problem, limiting the practicality of heuristics-based methods and rendering them computationally infeasible at scale. This paper proposes a learning-based framework for the stochastic electric truck routing under charging constraints and operational uncertainty. The problem, solved by Reinforcement Learning, is formulated as an event-driven semi-Markov decision process with shared charging resources, stochastic travel and energy requirements, and realistic nonlinear fast-charging behavior. To support learning in this setting, a graph-based representation of system state and feasible decisions is introduced, together with a rule-based action mask that restricts policies to operationally admissible actions; thus, improving training efficiency. Building on this formulation, an event-driven simulation environment is developed that supports both Reinforcement Learning and benchmarking against heuristic and mathematical programming baselines. Computational experiments across a range of fleet sizes show that the proposed learning-based algorithm consistently outperforms baselines and attains performance close to optimization benchmarks in many settings, while preserving high success rates under charging congestion and uncertainty.

Multi-agent deliberation systems using large language models (LLMs) are increasingly proposed for policy simulation, yet they suffer from artificial consensus: evaluator agents converge on the same option regardless of their assigned value perspectives. We present the AI Council, a three-phase deliberation framework, and conduct 120 deliberations across two policy scenarios to test two interventions. First, architectural heterogeneity (assigning a different 7-9B parameter model to each value perspective) significantly reduces first-choice concentration compared to a homogeneous baseline (child welfare: 70.9% to 46.1%, p < 0.001, r = 0.58; housing: 46.0% to 22.9%, p < 0.001, r = 0.50). This contrasts with accuracy-oriented multi-agent debate, where heterogeneity does not reduce convergence, suggesting model diversity operates differently when no objectively correct answer exists. Second, coherence validation (using a frontier model to assess whether each evaluator's reasoning is grounded in its assigned values) reveals a fidelity-diversity tradeoff: on a scenario with a dominant option, it further reduces concentration (46.1% to 40.8%, p = 0.004), but on a scenario with genuinely competitive options, it increases concentration (22.9% to 26.6%, p = 0.96) by amplifying high-coherence evaluators who cluster on one option. This tradeoff may be a general property of multi-agent systems employing quality weighting. We report negative results from three failed Delphi designs, demonstrate that 8B models exhibit binary rather than graded responses to counter-arguments, and propose the trustworthy tension rate as a diagnostic measure of small-model deliberation capabilities.

Deep learning methods have proved highly effective for classification and image recognition problems. In this paper, we ask whether this success can be transferred to hypothesis testing: if a neural network can distinguish, for example, an image of a handwritten digit from another, can it also distinguish an "image of a sample" (such as a scatter plot) generated under a given statistical model from one generated outside that model? Motivated by this idea, we propose a novel procedure called deep-testing, which approaches the classical inferential problem of hypothesis testing through deep learning. More specifically, the test statistic is a classification map learned by a deep neural network from simulated data satisfying the null and alternative hypotheses, leveraging its strong discriminating power to construct a highly powerful test. As a proof of concept, we apply deep-testing to the problem of independence testing, arguably one of the most important problems in statistics. In a large-scale simulation study, deep-testing achieves the highest overall power against nineteen competing methods across a broad range of complex dependence structures, confirming the viability of the proposed approach.

The increasing deployment of Large Language Model (LLM) inference on edge AI systems demands efficient execution under tight memory budgets. A key challenge arises from Key-Value (KV) caches, which often exceed available device memory. Although NVMe-based offloading offers scalable capacity, existing file-based designs rely heavily on the kernel page cache, leading to cache thrashing, unpredictable latency, and high software overhead under memory pressure. We present DUAL-BLADE, a dual-path KV residency framework that dynamically assigns KV tensors to either a page-cache path or an NVMe-direct path based on runtime memory availability. The NVMe-direct path bypasses the filesystem by mapping KV tensors to contiguous logical block address (LBA) regions, enabling low-overhead direct storage access. DUAL-BLADE further incorporates adaptive pipeline parallelism to overlap storage I/O with GPU DMA, improving inference throughput. Our evaluation shows that DUAL-BLADE substantially mitigates I/O bottlenecks, reducing prefill and decode latency by up to 33.1% and 42.4%, respectively, while improving SSD utilization by 2.2x across diverse memory budgets.

GPU-accelerated Self-Organizing Map (SOM) implementations are among the most competitive options for large-scale SOM analysis, but growing dataset sizes increasingly challenge their practical use because workloads no longer fit cleanly within device-memory limits. We introduce FloatSOM, a SOM framework for scalable training and deployment that supports multi-GPU execution, out-of-memory disk-backed streaming, and novel topologies beyond regular lattices. We evaluate FloatSOM on 14 synthetic and real benchmark datasets together with controlled speed scaling benchmarks, and show that these improved topologies, combined with topology-aware hyperparameter fine-tuning, yield lower quantization error than current state-of-the-art SOM baselines. FloatSOM also sustains this performance at large scale with high-throughput distributed execution; in the largest benchmark, it trains a 1024-node SOM network on 1,000,000,000 samples with 50 features in 6.16 minutes on 8 GPUs across two separate high-performance-computing nodes.

Large language models (LLMs) demonstrate strong multilingual capabilities, yet often fail to consistently generate responses in the intended language, exhibiting a phenomenon known as language confusion. Prior mitigation approaches based on sequence-level fine-tuning, such as DPO, ORPO, and GRPO, operate at the level of entire responses and can lead to unintended degradation of general model capabilities, motivating the need for more fine-grained alternatives. To address this, we introduce Token-Level Policy Optimization (TLPO), a fine-tuning framework designed to mitigate language confusion through localized, token-level updates. TLPO identifies error-prone positions, explores alternative candidate tokens, and updates the policy using a tailored objective to suppress error-inducing outputs at a granular level. This selective intervention enables effective mitigation of language confusion without compromising the model's general abilities. Experiments on multiple multilingual LLMs across diverse languages demonstrate that TLPO significantly outperforms baselines in improving language consistency while preserving downstream task accuracy.

We introduce Sparse-HFS, a scalable algorithm that can compute solutions to SSL problems using only O(n polylog(n)) space and O(m polylog(n)) time.

Over the past 25 years, I have been involved in some intriguing developments in the foundations of physics, exploring the quantum reality problem, the relationship between quantum theory and gravity and the interplay between consciousness and physical laws. These investigations make it plausible that we will find physics beyond quantum theory, potentially including both new evolution laws and new types of measurement. There is also a significant chance they could have potentially transformative impact on information processing and on the development of and our future with AI.

Smart Contracts are essential blockchain components, mainly written in Solidity. The high availability of public Solidity code leads to frequent reuse and high clone ratios. Since cloning can propagate vulnerabilities and flaws, effective detection is crucial. Although existing techniques work well in detecting syntactic clones, the identification of semantic clones is an open problem. To address this challenge, in this paper, we present and empirically assess a scalable methodology, based on analyzing code and comments, to spot semantically equivalent Solidity functions. We first collected an up-to-date dataset of about 300,000 Ethereum smart contracts, 82.07% of which are compliant with modern Solidity version 0.8. Manual validation of a statistically significant sample comprising 1,155 function pairs confirms the effectiveness of our solution, achieving an overall precision of 59% (rising to 84% for homonymous functions) and a recall of 97%. Besides, we explore the structural differences occurring on semantically equivalent Solidity functions, demonstrating that they often represent design alternatives focused on security choices, modularization, and gas optimization. Finally, we investigate the use of Large Language Models (LLMs) as documentation engines in scenarios where code comments are poor or absent. Our results show that LLM-generated summaries, combined with sentence transformers like BERT, can bridge the documentation gap, enabling the identification of semantic clones in uncommented code with 75% precision. This work establishes a modern benchmark for Solidity clone detection and provides a foundation for the automated discovery of secure and efficient code alternatives.

Maintaining up-to-date, comprehensive documentation for large codebases is a persistent challenge. Recent progress in automated documentation has moved from template-based rules to large language models (LLMs), yet existing tools still process source code as flat fragments, producing isolated documents that lack semantic structure. This design also leads to excessive token consumption and slow generation, while failing to capture how code changes propagate across dependencies. We propose RepoDoc, a system that uses a repository knowledge graph (RepoKG) as the semantic foundation for the entire documentation lifecycle. Our framework consists of three stages: (1) RepoKG construction, which extracts code entities and their relationships; (2) module clustering, which groups code into functionally cohesive, hierarchical units; and (3) skillful agent-based generation, which queries the graph to create modular, cross-referenced documentation with auto-generated Mermaid diagrams. For incremental maintenance, a semantic impact propagation mechanism navigates the RepoKG bidirectionally to pinpoint all affected parts, allowing selective, targeted regeneration. Evaluated on 24 repositories across 8 programming languages, RepoDoc substantially outperforms state-of-the-art alternatives. It improves API coverage by 32.5% and completeness by 10.4%, while generating documentation 3x faster with 85% fewer tokens. For incremental updates, it cuts update time by 73% and token usage by 77%, and achieves 10.2% higher update recall, more accurately reflecting code changes in the regenerated documentation. The source code and experimental artifacts are available at https://github.com/SYSUSELab/RepoDoc.

Large Language Model (LLM)-based agents exhibit systemic failures in compositional generalization, limiting their robustness in interactive environments. This work introduces AGEL-Comp, a neuro-symbolic AI agent architecture designed to address this challenge by grounding actions of the agent. AGEL-Comp integrates three core innovations: (1) a dynamic Causal Program Graph (CPG) as a world model, representing procedural and causal knowledge as a directed hypergraph; (2) an Inductive Logic Programming (ILP) engine that synthesizes new Horn clauses from experiential feedback, grounding symbolic knowledge through interaction; and (3) a hybrid reasoning core where an LLM proposes a set of candidate sub-goals that are verified for logical consistency by a Neural Theorem Prover (NTP). Together, these components operationalize a deduction--abduction learning cycle: enabling the agent to deduce plans and abductively expand its symbolic world model, while a neural adaptation phase keeps its reasoning engine aligned with new knowledge. We propose an evaluation protocol within the \texttt{Retro Quest} simulation environment to probe for compositional generalization scenarios to evaluate our AGEL agent. Our findings clearly indicate the better performance of our AGEL model over pure LLM-based models. Our framework presents a principled path toward agents that build an explicit, interpretable, and compositionally structured understanding of their world.

Compositional generalization remains a foundational weakness of modern neural networks, limiting their robustness and applicability in domains requiring out-of-distribution reasoning. A central, yet unverified, assumption in neuro-symbolic AI is that compositional reasoning will emerge as a byproduct of successful symbol grounding. This work presents the first systematic empirical analysis to challenge this assumption by disentangling the contributions of grounding and reasoning. To operationalize this investigation, we introduce the Iterative Logic Tensor Network ($i$LTN), a fully differentiable architecture designed for multi-step deduction. Using a formal taxonomy of generalization -- probing for novel entities, unseen relations, and complex rule compositions -- we demonstrate that a model trained solely on a grounding objective fails to generalize. In contrast, our full $i$LTN, trained jointly on perceptual grounding and multi-step reasoning, achieves high zero-shot accuracy across all tasks. Our findings provide conclusive evidence that symbol grounding, while necessary, is insufficient for generalization, establishing that reasoning is not an emergent property but a distinct capability that requires an explicit learning objective.

Offline reinforcement learning (RL) agents often fail when deployed, as the gap between training datasets and real environments leads to unsafe behavior. To address this, we present SAS (Self-Alignment for Safety), a transformer-based framework that enables test-time adaptation in offline safe RL without retraining. In SAS, the main mechanism is self-alignment: at test time, the pretrained agent generates several imagined trajectories and selects those satisfying the Lyapunov condition. These feasible segments are then recycled as in-context prompts, allowing the agent to realign its behavior toward safety while avoiding parameter updates. In effect, SAS turns Lyapunov-guided imagination into control-invariant prompts, and its transformer architecture admits a hierarchical RL interpretation where prompting functions as Bayesian inference over latent skills. Across Safety Gymnasium and MuJoCo benchmarks, SAS consistently reduces cost and failure while maintaining or improving return.

This paper investigates efficient methods for utilizing text-only data to improve speech recognition, focusing on encoder-dominated models that facilitate faster recognition. We provide a comprehensive comparison of techniques to integrate text-only data, including modality matching and dynamic downsampling to reach text-level representations within the encoder. Our experiments on the LibriSpeech corpus show that a larger encoder with a smaller decoder can equal or surpass the performance of architectures with larger decoders. We demonstrate that simple configurations, such as random duration models, are often more effective than complex alternatives, significantly simplifying the training pipeline. All code and recipes are made publicly available.

Alignment faking (AF) occurs when an LLM strategically complies with training objectives to avoid value modification, reverting to prior preferences once monitoring is lifted. Current detection methods focus on conversational settings and rely primarily on Chain-of-Thought (CoT) analysis, which provides a reliable signal when strategic reasoning surfaces, but cannot distinguish deception from capability failures if traces are absent or unfaithful. We formalize AF as a composite behavioural event and detect it through observable tool selection, where the LLM selects the safe tool when unmonitored, but switches to the unsafe tool under monitoring that rewards helpfulness over safety, while its reasoning still acknowledges the safe choice. We release a dataset of 108 enterprise IT scenarios spanning Security, Privacy, and Integrity domains under Corruption and Sabotage pressures. Evaluating six frontier LLMs across five independent runs, we find mean AF detection rates between 3.5% and 23.7%, with vulnerability profiles varying by domain and pressure type. These results suggest that susceptibility reflects training methodology rather than capability alone.

Deploying Vision-Language Models (VLMs) on edge devices remains challenging due to their substantial computational and memory demands, which exceed the capabilities of resource-constrained embedded platforms. Conversely, fully offloading inference to the cloud is often impractical in bandwidth-limited environments, where transmitting raw visual data introduces substantial latency overhead. While recent edge-cloud collaborative architectures attempt to partition VLM workloads across devices, they typically rely on transmitting fixed-size representations, lacking adaptability to dynamic network conditions and failing to fully exploit semantic redundancy. In this paper, we propose a progressive semantic communication framework for edge-cloud VLM inference, using a Meta AutoEncoder that compresses visual tokens into adaptive, progressively refinable representations, enabling plug-and-play deployment with off-the-shelf VLMs without additional fine-tuning. This design allows flexible transmission at different information levels, providing a controllable trade-off between communication cost and semantic fidelity. We implement a full end-to-end edge-cloud system comprising an embedded NXP i.MX95 platform and a GPU server, communicating over bandwidth-constrained networks. Experimental results show that, at 1 Mbps uplink, the proposed progressive scheme significantly reduces network latency compared to full-edge and full-cloud solutions, while maintaining high semantic consistency even under high compression. The implementation code will be released upon publication at https://github.com/open-ep/ProSemComVLM.

Background & Objectives: In the last decade, Machine learning research has grown rapidly, but large models are reaching their soft limits demonstrating diminishing returns and still lack solid reasoning abilities. These limits could be surpassed through synergistic combination of Machine Learning scalability and rigid reasoning. Methods: In this work, we propose a theoretical framework for reasoning through object-relations in an automated manner integrated with Artificial Neural Networks. We present a formal analysis of the Reasoning, and we show the theory in practice through a paradigm integrating Reasoning and Machine Learning. Results: This paradigm is a system that solves Intelligence Quotient problems without any prior knowledge of the problem. Our system achieves 98.03% solving rate corresponding to the top 1% percentile or 132-144 iq score. This result is only limited by the small size of the model and the processing capabilities of the machine it run on. Conclusions: With the integration of prior knowledge in the system and the expansion of the dataset, the system can be generalized to solve a large category of problems. The functionality of the system inherently favors the solution of such problems in few-shot or zero-shot attempts.

As Large Language Models (LLMs) are increasingly integrated into academic peer review, their vulnerability to adversarial prompts -- adversarial instructions embedded in submissions to manipulate outcomes -- emerges as a critical threat to scholarly integrity. To counter this, we propose a novel adversarial framework where a Generator model, trained to create sophisticated attack prompts, is jointly optimized with a Defender model tasked with their detection. This system is trained using a loss function inspired by Information Retrieval Generative Adversarial Networks, which fosters a dynamic co-evolution between the two models, forcing the Defender to develop robust capabilities against continuously improving attack strategies. The resulting framework demonstrates significantly enhanced resilience to novel and evolving threats compared to static defenses, thereby establishing a critical foundation for securing the integrity of peer review.

Dynamic quantization emerged as a practical approach to increase the utilization and efficiency of the machine learning serving flow. Unlike static quantization, which applies quantization offline, dynamic quantization operates on tensors at run-time, adapting its parameters to the actual input data. Today's mainstream machine learning frameworks, including ML compilers and inference engines, frequently recommend dynamic quantization as an initial step for optimizing model serving. This is because dynamic quantization can significantly reduce memory usage and computational load, leading to faster token generation and improved model serving efficiency without substantial loss in model accuracy. In this paper, we reveal a critical vulnerability in dynamic quantization: an adversary can exploit such quantization strategy to steal sensitive user data placed in the same batch as the adversary's input. Our analysis demonstrates that dynamic quantization, when improperly implemented or configured, can create side channels that expose information about other inputs within the same batch. We call this phenomenon Quantamination, describing contamination from quantization. Specifically, we show that at least 4 of the most popular ML frameworks in use today either default to or can use configurations that leak data across the batch boundary. This data leakage, in theory, allows attackers to partially or even fully recover other users' batched input data, representing a serious privacy risk for existing ML serving frameworks.

Generating sports game reports from structured tables is a complex table-to-text task that demands both precise data interpretation and fluent narrative generation. Traditional model-based approaches require large, annotated datasets, while prompt-based methods using large language models (LLMs) often struggle with hallucination due to weak table comprehension. To overcome these challenges, we propose Tree-of-Text, a tree-structured prompting framework that guides LLMs through a three-stage generation process: (1) Content Planning, where relevant operations and arguments are selected from the input tables; (2) Operation Execution, which breaks down large tables into manageable sub-tables; and (3) Content Generation, where short textual outputs are merged and rewritten into a cohesive report. Experiments show that our method outperforms existing methods on ShuttleSet+, leads in RG and CO metrics on RotoWire-FG, and excels in CS and CO on MLB with roughly 40% of the time and cost of Chain-of-Table. These results demonstrate the effectiveness and efficiency of Tree-of-Text and suggest a promising direction for prompt-based table-to-text generation in the sports domain.

LLMs and speech assistants are increasingly used for task-oriented interactions, yet their evaluation often relies on controlled scenarios that fail to capture the variability and complexity of real user requests. Drink ordering, for example, involves diverse named entities, drink types, sizes, customizations, and brand-specific terminology, as well as spontaneous speech phenomena such as hesitations and self-corrections. To address this gap, we introduce StarDrinks, a test set in English and Korean containing speech utterances features, transcriptions, and annotated slots. Our dataset supports speech-to-slots SLU, transcription-to-slots NLU, and speech-to-transcription ASR evaluation, providing a realistic benchmark for model robustness and generalization in a linguistically rich, real-world task.

The rapid growth of molecular foundation models and general-purpose large language models has encouraged a scale-centric view of artificial intelligence in drug discovery, in which larger pretrained models are expected to supersede compact cheminformatics models and task-specific graph neural networks (GNNs). We test this assumption on 22 molecular property and activity endpoints, including public ADMET and Tox21 benchmarks and two internal anti-infective activity datasets. Across 167,056 held-out task--molecule evaluations under structure-similarity-separated five-fold cross-validation (37,756 ADMET, 77,946 Tox21, 49,266 anti-TB and 2,088 antimalaria), classical machine-learning (ML) models such as RF(ECFP4) and ExtraTrees(RDKit descriptors) win ten primary-metric tasks, GNNs such as GIN and Ligandformer win nine, and pretrained molecular sequence models such as MoLFormer and ChemBERTa2 win three. Rule-based SAR reasoning baselines, represented by GPT5.5-SAR and Opus4.7-SAR, do not win under the prespecified primary metrics, although train-fold-derived SAR knowledge provides measurable but uneven gains for SAR reasoning and interpretation. These results indicate that compact, specialized models remain highly effective for molecular property and activity prediction. The performance differences among classical ML, GNN and pretrained sequence models are often modest and endpoint-dependent, whereas larger or more general models do not provide a universal predictive advantage. Large models may still add value for zero-shot reasoning, SAR interpretation and hypothesis generation, but the results suggest that predictive performance depends on the alignment among molecular representation, inductive bias, data regime, endpoint biology and validation protocol.

Generative AI tools are widely used by youth and have introduced new privacy and safety challenges. While prior research has explored youth's safety in GenAI within western context, it often overlooks the cultural, religious, and social dimensions of technology use that strongly shape youths digital experiences in countries like Saudi Arabia. To address the gap, this study explores children (aged 7 to 17), parents and teachers interactions with GenAI tools and risk perceptions through non-western lens. Through a mixed methods approach, we analyzed 736 Reddit and 1,262 X(Twitter) posts and conducted interviews with 31 Saudi Arabian participants (8 youth, 13 parents, 10 teachers). Our findings highlight context dependent and relational privacy and safety of GenAI from non-western context which often formed by communal structure and prescribed norms. We found significant risks tied to youths disclosure of personal and family information, which conflict with culturally rooted expectations of modesty, privacy, and honor, particularly when youth seek emotional support from GenAI. These risks further compounded by socio economic factors such as cost-saving practices leading to the use of shared GenAI accounts (e.g.ChatGPT) within families or even among strangers. We provide design implication reflecting on parents and teachers expectation of how youth should use GenAI. This work lays groundwork for inclusive, context sensitive parental controls that adhere to cultural norms and values.

DNNs have gained widespread adoption in feature interaction recommendation models. However, there has been a longstanding debate on their roles. On one hand, some works claim that DNNs possess the ability to implicitly capture high-order feature interactions. Conversely, recent studies have highlighted the limitations of DNNs in effectively learning dot products, specifically second-order interactions, let alone higher-order interactions. In this paper, we present a novel perspective to understand the effectiveness of DNNs: their impact on the dimensional robustness of the representations. In particular, we conduct extensive experiments involving both parallel DNNs and stacked DNNs. Our evaluation encompasses an overall study of complete DNN on two feature interaction models, alongside a fine-grained ablation analysis of components within DNNs. Experimental results demonstrate that both parallel and stacked DNNs can effectively mitigate the dimensional collapse of embeddings. Furthermore, a gradient-based theoretical analysis, supported by empirical evidence, uncovers the underlying mechanisms of dimensional collapse.

One of the most exciting applications of vision models involve pixel-level reasoning. Despite the abundance of vision foundation models, we still lack representations that effectively embed spatio-temporal properties of visual scenes at the pixel level. Existing frameworks either train on image-based pretext tasks, which do not account for dynamic elements, or on video sequences for action-level reasoning, which does not scale to dense pixel-level prediction. We present a framework that learns pixel-accurate feature descriptors from videos, LILA. The core element of our training framework is linear in-context learning. LILA leverages spatio-temporal cue maps -- depth and motion -- estimated with off-the-shelf networks. Despite the noisy nature of those cues, LILA trains effectively on uncurated video datasets, embedding semantic and geometric properties in a temporally consistent manner. We demonstrate compelling empirical benefits of the learned representation across a diverse suite of vision tasks: video object segmentation, surface normal estimation and semantic segmentation.

Safety-critical prediction systems, such as autonomous vehicles, weather forecasters, and medical monitors, commonly rely on probabilistic forecasters. These forecasters make predictions about possible future outcomes, and their quality and robustness needs to be validated and certified. Often, only accuracy -- the mean of the predictions -- is evaluated against true outcomes. However, for safety-critical scenarios and decision making under uncertainty, the full distributional properties of the forecasts should be checked: do the observed prediction errors actually follow the forecasted probability distributions? To this end, we introduce a framework for calibration checks: statistical tests that validate distributional properties of forecasts when measured over many samples. In order to support ease-of-use in real-world operations, these checks produce a single accept/reject decision for data collected from a forecaster. This contrasts typical calibration calculations which produce one or multiple continuous calibration scores and require expertise to implement in a validation workflow. We further support operationalization by introducing modifications to calibration testing that (a) reject only overconfident predictions, allowing for pessimistic or cautious predictions in safety-critical settings, and (b) tolerate small, operationally acceptable deviations even for large numbers of validation samples. We organize the calibration checking process into a modular pipeline comprising four steps: (i) the data model, (ii) the chosen metric, (iii) the hypothesis formulation, and (iv) the testing procedure. Each step consists of independently swappable components, thereby supporting a large variety of possible use-cases and trade-offs. We demonstrate the applicability of the framework on two complementary example problems, weather forecasting and robot pose estimation.

We study sequential interventions under prerequisite constraints. In this setting, admissible intervention sequences are paths in the ideal lattice of a finite prerequisite poset rather than unconstrained action strings. We give an exact local-to-global theory of order sensitivity on this state space. First, we prove that any two admissible paths with the same endpoints differ by a finite sequence of elementary diamond swaps. Second, for edge-additive path valuations, we show that path-independence is equivalent to vanishing diamond curvature, yielding an endpoint potential with a canonical Möbius parameterization on the ideal lattice. Third, we prove that a local diamond field is induced by an edge-based path model if and only if it satisfies cube consistency, with uniqueness after fixing a reference-tree gauge. Under reduced-state longitudinal assumptions, supported reference paths identify reference-path scores, whereas local order effects require two-sided support of both orders on each diamond. These results yield exact planning consequences, including an order-insensitivity bound and dynamic programming on the truncated ideal lattice.

Industrial systems increasingly depend on Machine Learning (ML), and operate on heterogeneous nodes that must satisfy tight latency, energy, and memory constraints. Dynamic ML models, which reconfigure their computational footprint at runtime, promise high energy efficiency and lower average latency for modest accuracy tradeoffs; however, their deployment is complex due to the additional hyperparameters they rely on. These hyperparameters, controlling the accuracy versus average latency tradeoff, are often tuned on a calibration dataset that must match the test time distribution, an assumption that rarely holds in real-world scenarios, leading to suboptimal operational conditions, possibly below static models. We propose a two-tier adaptive architecture that co-optimizes model and system decisions. At the global level, a scheduler configures and deploys, for each edge node, a cascade of classifiers composed of lightweight specialized models and a generalist fallback, satisfying latency and memory constraints. At the node level, a local controller tracks data drifts and hardware resources, enabling or disabling specialized predictors (SP) to preserve high energy efficiency and avoid latency-constraint violations under varying conditions. This design allows longer operating times without forcing a global redeployment step, and enables efficient execution in case of an unreachable remote global controller. We evaluate the approach on two datasets under controlled distribution mismatch scenarios, showing average per-inference reductions of latency up to 2.45x and energy up to 2.86x, with less than 4% accuracy drop compared to static baselines. Our contributions are:(1) a budgeted SP-cascade formulation that preserves worst-case latency constraints;(2) a hierarchical controller that maintains efficiency under data and resource changes; and (3) an experimental evaluation on embedded hardware.

Large Language Models (LLMs) have become widely used for Software Engineering (SE) tasks, spanning from function-level code generation to complex repository-level workflows. However, the high latency of autoregressive inference remains a significant bottleneck, hindering their deployment in interactive environments. While Speculative Decoding (SD) offers a promising technique for lossless acceleration, prior research on long-context repository-level tasks and complex agentic interactions remains limited. To bridge this gap, we present the first systematic empirical study to evaluate the effectiveness of SD in SE tasks. We systematically benchmark a comprehensive spectrum of strategies, encompassing both model-based and model-free methods, across representative generation, editing, and repair scenarios. Our empirical results indicate that SD demonstrates clear potential for accelerating inference, particularly for smaller models that achieve higher speedups than those of their larger counterparts. We find that the effectiveness of SD methods varies across different task scenarios. Model-based approaches are well-suited for code generation, whereas model-free methods are better adapted to repository-level repair and editing scenarios. Furthermore, we observe that the repetitiveness of SE tasks improves the performance of model-free methods. In contrast to natural language tasks, the higher predictability of SE tasks allows for more aggressive hyperparameters. Our findings are summarized as guidelines to help increase inference efficiency for SE scenarios.

Stylistic personalization - making LLMs write in a specific individual's style, rather than merely adapting to task preferences - lacks evaluation grounded in authorship science. We show that grounding evaluation in authorship verification theory transforms what benchmarks can measure. Drawing on three measurement traditions - LUAR, a trained authorship verification model; an LLM-as-judge with decoupled trait matching; and classical function-word stylometrics - we evaluate four inference-time personalization methods across 50 authors and 1,000 generations. The theory-grounded metric, LUAR, provides what ad hoc alternatives cannot: calibrated baselines, with a human ceiling of 0.756 and a cross-author floor of 0.626, that give scores absolute meaning. All methods score below this floor, from 0.484 to 0.508, exposing an authorship gap invisible to uncalibrated metrics. The three metrics produce near-zero pairwise correlations, with absolute r less than 0.07, confirming that without theoretical grounding, metric choice determines conclusions: an LLM judge declares a clear winner while LUAR finds no meaningful differentiation. These findings demonstrate the theory-benchmark cycle in action: authorship theory exposes evaluation failures that ad hoc benchmarks miss.

The digitisation of classical Sanskrit literature is impeded by a scarcity of annotated resources, particularly for Named Entity Recognition. While recent methodologies utilise generic Large Language Models (LLMs) for data augmentation, these approaches remain prone to error and often lack the reasoning depth required for classical grammar. In this work, we introduce Naamah, a high quality silver standard Sanskrit NER dataset comprising 102,942 sentences. We propose a methodology that combines entity extraction from DBpedia with the generative capabilities of a 24B parameter hybrid reasoning model to create grammatically natural and synthetically diverse training data. We utilize this dataset to benchmark two transformer architectures: the massive multilingual XLM RoBERTa and the parameter efficient IndicBERTv2.

We study three problems that involve identifying homogeneous halfspaces under Gaussian distributions: agnostic learning, one-sided reliable learning, and fairness auditing. In each of these problems, we are given labeled examples $(\mathbf{x}, \mathrm{y})$ drawn from an unknown distribution on $\mathbb{R}^d\times\{-1, +1\}$, whose marginal distribution on $\mathbf{x}$ is standard Gaussian and on $\mathrm{y}$ is arbitrary. The goal of each problem is to output a homogeneous halfspace that approaches the best-fitting homogeneous halfspace in terms of its corresponding loss measure. We prove near-optimal computational hardness results for these problems under the widely believed hardness assumption of the Learning With Errors (LWE) problem. Prior hardness results for these problems were mostly established for general halfspaces; our findings extend some of these hardness results to homogeneous halfspaces. Remarkably, our lower bound strictly generalizes over prior works and narrows the gap between the upper and lower bounds for agnostically learning homogeneous halfspaces under Gaussian marginals.

We prove that any continuous function f from [0,1]^n to R representable by a finite computation tree with N internal nodes and compositional sparsity s = O(1) admits a deep Kolmogorov-Arnold Network (KAN) representation. Each internal node is realised by a primitive KAN block with controlled block depth and Lipschitz product. The layer-wise Lipschitz product satisfies the primary domain-sensitive bound independent of the input dimension n. It simplifies to P(KAN_f) <= max(C*,1)^L_f with L_f <= c_max * N. For the standard operations {+,-,x,sin,cos} with x nodes on [0,1]-bounded inputs we obtain P(KAN) <= 1. Layer widths satisfy n_l <= n + 2 w_max * N. The uniform approximation error is bounded by N * max(C*,1)^d(f) * epsilon_Op (simplifies when C* <=1). For f in C^m we obtain optimal B-spline rates. Range bounds are also derived (B_f <= N+1 for additive trees). This addresses the gap on Lipschitz control in deep KAN stacks noted by Liu et al. (2024). Experiments confirm P(KAN)=1.0 for several compositionally structured functions.

The rapid advancement of object detection architectures has positioned single stage detectors as the dominant solution for real-time visual perception. A primary source of computational overhead in these models lies in the deep backbone stages, where C2f bottleneck modules at high stride levels accumulate a disproportionate share of parameters due to quadratic scaling with channel width. This work introduces QYOLO, a quantum-inspired channel mixing framework that achieves genuine architectural compression by replacing the two deepest backbone C2f modules at P4/16 (512 channels) and P5/32 (1024 channels) with a compact QMixBlock. The proposed block performs global channel recalibration through a sinusoidal mixing mechanism with shared learnable parameters across both backbone stages, enforcing consistent channel importance without requiring independent per-stage parameter sets. The neck and detection head remain fully classical and unchanged. Evaluation on the VisDrone2019 benchmark demonstrates that QYOLOv8n achieves a 20.2% reduction in parameter count (3.01M to 2.40M) and 12.3% GFLOPs reduction with only 0.4 pp mAP@50 degradation. QYOLOv8s achieves 21.8% reduction with 0.1 pp degradation. When combined with knowledge distillation, full accuracy parity is recovered at no cost to compression. An expanded backbone plus neck variant achieved 38 to 41% reduction at the cost of greater accuracy degradation, motivating the backbone-only final design.

Accurate end-to-end tail-latency forecasting is critical for proactive SLO management in microservice systems. However, modeling long-range dependency propagation and non-stationary, bursty workloads while maintaining inference efficiency at scale remains challenging. We present STLGT (Scalable Trace-based Linear Graph Transformer), a per-API predictor that encodes traces as span graphs for multi-step p95 tail-latency forecasting. STLGT uses a structure-aware linear graph Transformer to propagate cross-service dependencies with inference time linear in span graph size, and a decoupled temporal module to capture workload dynamics. Across a personalized education microservice application, DeathStarBench, and Alibaba traces, STLGT improves forecasting accuracy over PERT-GNN by 8.5% MAPE on average and achieves up to 12x faster CPU inference at N=32, matching the maximum span graph size after preprocessing the Alibaba traces. Ablation studies further demonstrate the effectiveness of each component, especially under bursty traffic.

Large Vision-Language Models (VLMs) have achieved remarkable multimodal performance yet remain prone to factual hallucinations, particularly in long-tail or specialized domains. Moreover, current models exhibit a weak capacity to refuse queries that exceed their parametric knowledge. In this paper, we propose a systematic framework to enhance the refusal capability of VLMs when facing such unknown questions. We first curate a model-specific "Visual-Idk" (Visual-I don't know) dataset, leveraging multi-sample consistency probing to distinguish between known and unknown facts. We then align the model using supervised fine-tuning followed by preference-aware optimization (e.g., DPO, ORPO) to effectively delineate its knowledge boundaries. Results on the Visual-Idk dataset show our method improves the Truthful Rate from 57.9\% to 67.3\%. Additionally, internal probing also demonstrates that the model genuinely recognizes its boundaries instead of just memorizing refusal patterns. Our framework further generalizes to out-of-distribution medical and perceptual domains, providing a robust path toward more trustworthy and prudent visual assistants.

Emotion perception and adaptive expression are fundamental capabilities in human-agent interaction. While recent advances in speech emotion captioning (SEC) have improved fine-grained emotional modeling, existing systems remain limited to static, single-emotion characterization within isolated sentences, neglecting dynamic emotional transitions at the discourse level. To address this gap, we propose Emotion Transition-Aware Speech Captioning (EmoTransCap), a paradigm that integrates temporal emotion dynamics with discourse-level speech description. To construct a dataset rich in emotion transitions while enabling scalable expansion, we design an automated pipeline for dataset creation. This is the first large-scale dataset explicitly designed to capture discourse-level emotion transitions. To generate semantically rich descriptions, we incorporate acoustic attributes and temporal cues from discourse-level speech. Our Multi-Task Emotion Transition Recognition (MTETR) model performs joint emotion transition detection and diarization. Leveraging the semantic analysis capabilities of LLMs, we produce two annotation versions: descriptive and instruction-oriented. These data and annotations offer a valuable resource for advancing emotion perception and emotional expressiveness. The dataset enables speech captions that capture emotional transitions, facilitating temporal-dynamic and fine-grained emotion understanding. We also introduce a controllable, transition-aware emotional speech synthesis system at the discourse level, enhancing anthropomorphic emotional expressiveness and supporting emotionally intelligent conversational agents.

Software Defined Vehicles face an increasing computational gap as advanced algorithms and frequent software updates demand more processing power while onboard hardware remains static throughout a vehicle's 10+ year lifespan. This mismatch threatens the performance of safety-critical functions including advanced driver-assistance systems and real-time perception tasks. We propose a novel four-layer computation offloading pipeline that dynamically distributes vehicular functions to cloud and edge resources while meeting strict Round Trip Time constraints. Our key contribution is an enhanced Particle Swarm Optimization algorithm that integrates distance- and direction-based penalties with functional requirements to optimize edge server selection for mobile vehicles. Evaluation using a Kubernetes-based cloud infrastructure with realistic vehicular mobility patterns demonstrates that our approach reduces average response time compared to conventional Brute-Force methods while maintaining the success rate for latency-critical tasks. The modified Particle Swarm Optimization algorithm achieves an average execution time of 26 ms across ten servers and tasks on Central Processing Unit, and 550ms across 15 servers with 1000 tasks on Graphics Processing Unit. These results confirm the pipeline's effectiveness in bridging the computational gap for next-generation Software Defined Vehicles (SDV).

This paper presents Quantum Gatekeeper, a context-bound image steganography framework where successful payload recovery depends on both cryptographic decryption and the reconstruction of a precise extraction path. The system integrates lossless least significant bit (LSB) embedding with a deterministic variational quantum circuit (VQC)-derived gate key, multi-factor contextual binding, and authenticated encryption. Payload extraction is contingent upon four requisite factors: a password, a shared secret, a user-supplied context string, and a reference image signature. Any deviation in these factors causes the system to read from an incorrect pixel sequence or fail authentication, resulting in silent rejection rather than partial disclosure. The proposed method derives a gatecontrolled extraction key from a seed-conditioned variational circuit, with parameters generated via cryptographic hash expansion and context-dependent image features. To ensure encode/decode consistency, the cryptographic key path is generated via exact statevector simulation; concurrently, IBM superconducting quantum hardware is utilized to evaluate the statistical behavior of the circuit family under physical noise. We introduce a dual-region image layout to resolve the nonce bootstrapping dependency, separating header recovery from payload recovery through independently derived keys. Experimental results confirm successful end-to-end message embedding and recovery on PNG images, demonstrating deterministic success under correct conditions and failure otherwise. The framework supports both text and image payloads; in the image-in-image configuration, a secret image is resized to a fixed resolution prior to embedding, enabling exact pixel-level recovery under correct contextual reconstruction.

Speculative decoding accelerates LLM inference, but SOTA hidden-state-based drafters suffer from long-range decay: draft accuracy degrades as the speculative step increases. Existing work attributes this decay to train-inference mismatch and proposes test-time training (TTT) as a remedy, yet we observe that long-range decay persists even in TTT-trained drafters. We revisit long-range decay from the perspective of context information preservation. In hidden-state reuse, we argue the target hidden state acts as a biased context compression: it aggregates historical token information according to the attention query at the current position, yielding a compact representation optimized for immediate next-token prediction. This compression can suppress information less relevant to the current query but important for later speculative steps. In contrast, the target model's KV cache serves as an explicit context, retaining the complete set of token-wise KV representations. We therefore posit the KV-Reuse Hypothesis: allowing the draft model to reuse the target KV cache can provide richer signals for long-horizon drafting. To test this hypothesis, we introduce KVShot, a diagnostic framework that compares three reuse paradigms: hidden-only, KV-only, and hybrid. Extensive evaluations on Qwen3-8B show that KV-Reuse improves long-range acceptance, although end-to-end speedups remain marginal under current training pipelines. Our analysis identifies two key structural bottlenecks: shallow drafters struggle to estimate target queries accurately, and draft-side KV projections receive sparse gradient signals. These findings suggest that realizing the full potential of KV-aware decoding requires moving beyond TTT toward block-wise training paradigms. By exposing these bottlenecks, KVShot provides a foundational diagnostic testbed and a clear roadmap for designing next-generation inference architectures.

Runtime monitoring is essential to ensure the safety of ML applications in safety-critical domains. However, current research is fragmented, with independent methods emerging from different communities. In this paper, we propose a unified framework categorising runtime monitoring approaches into three distinct types: Operational Design Domain (ODD) monitoring, which ensures compliance with expected operating conditions; Out-of-Distribution (OOD) monitoring, which rejects inputs that deviate from the training data; and Out-of-Model-Scope (OMS) monitoring, which detects anomalous model behaviour based its internal states or outputs. We demonstrate the benefits of this categorization with a dedicated experiment on an aeronautical safety-critical application: runway detection during landing. This framework facilitates design of monitoring activities, with complementary categories of monitors, and enables evaluation and comparison of different monitors using common, safety-oriented metrics.

Recent advances in large language models and agentic frameworks have enabled virtual customer assistants (VCAs) for complex support. We present SecMate, a multi-agent VCA for cybersecurity troubleshooting that integrates device, user, and service specificity from conversational and device-level signals. Device specificity is provided by a lightweight local diagnostic utility, while user specificity relies on implicit proficiency inference and profile-aware troubleshooting. Service specificity is achieved through a proactive, context-aware recommender. We evaluate SecMate in a controlled study with 144 participants and 711 conversations. Device-level evidence increased correct resolutions from about 50% to over 90% relative to an LLM-only baseline, while step-by-step guidance improved pleasantness and reduced user burden. The recommender achieved high relevance (MRR@1=0.75), and participants showed strong willingness to substitute human IT support at costs well below human benchmarks. We release the full code base and a richly annotated dataset to support reproducible research on adaptive VCAs.

Balancing differential privacy (DP) with recommendation accuracy is a key challenge in privacy-preserving recommender systems, since DP-noise degrades accuracy. We address this trade-off at both the data and model levels. At the data level, we apply DP only to the most stereotypical user data likely to reveal sensitive attributes, such as gender or age, to reduce unnecessary perturbation; we refer to this as targeted DP. At the model level, we use meta-learning to improve robustness to remaining DP-noise. This achieves a better trade-off between accuracy and privacy than standard approaches: Meta-learning improves accuracy and targeted DP leads to lower empirical privacy risk compared to uniformly applied DP and full DP baselines. Overall, our findings show that selectively applying DP at the data level together with meta-learning at the model level can effectively balance recommendation accuracy and user privacy.

Federated Split Learning has been identified as an efficient approach to address the computational resource constraints of clients in classical federated learning, while guaranteeing data privacy for distributed model training across data owners. However, it faces some critical challenges when such a training strategy meets large language models (LLMs) for fine-tuning. Such challenges include setting the cutlayer adaptively across different clients to address the data and device heterogeneity issues, which affect the system performance significantly. In addition, efficiently reducing the communication overhead during the fine-tuning procedure is also another challenge. No work tries to address these challenges. To bridge this gap, we propose SplitTF, an adaptive federated split learning system for LLMs fine-tuning. SplitFT enables different clients to set different cut layers according to their computation resources and trained model performance. SplitFT also proposes to reduce the LoRA rank in cutlayer to reduce the communication overhead. In addition to simulating the heterogeneous data in real-world applications for our proposed split federated learning system, we propose a length-based Dirichlet approach to divide the training data into different clients. Extensive experimental results show that our proposed approach outperforms the state-of-the-art approach for fine-tuning time efficiency and model performance based on various popular benchmarks.

Increasing complexity of modern enterprise systems and the demand for automation and interoperability require consistent and semantically validated models in Model-Based Systems Engineering (MBSE). The Object Constraint Language (OCL) supports formal definition of such constraint validations. However, MBSE models and OCL constraints are typically managed in separate tools, causing manual effort during model constraint application and result interpretation. To address this gap, this paper proposes an approach to managing OCL constraints and their validation results through Asset Administration Shells (a well-established technology for interoperability in enterprise systems). The methodology is demonstrated through a fictional industrial scenario, and to support reproducibility, all artifacts are publicly available in a GitHub repository.

Most enterprise document AI today is a pipeline. Parse, index, retrieve, generate. Each of those stages has been studied to death on its own -- what's still hard is evaluating the system as a whole. We built EnterpriseDocBench to take a swing at it: parsing fidelity, indexing efficiency, retrieval relevance, and generation groundedness, all on the same corpus. The corpus is built from public, permissively licensed documents across six enterprise domains (five represented in the current pilot). We ran three pipelines through it -- BM25, dense embedding, and a hybrid -- all with the same GPT-5 generator. The headline numbers: hybrid retrieval narrowly beats BM25 (nDCG@5 of 0.92 vs. 0.91), and both beat dense embedding (0.83). Hallucination doesn't grow monotonically with document length -- short documents and very long ones both hallucinate more than medium ones (28.1% and 23.8% vs. 9.2%). Cross-stage correlations are very weak: parsing->retrieval r=0.14, parsing->generation r=0.17, retrieval->generation 0.02. If quality were cascading the way most of us assume, those numbers would be much higher; they aren't. Design caveats are real (parsing fixed, generator shared, automated proxy metrics) and we don't oversell the result. One result that genuinely surprised us: factual accuracy on stated claims is 85.5%, but answer completeness averages 0.40. The system is right when it answers -- it just leaves things out. That gap matters more for real deployments than the headline accuracy number does. We also describe three reference architectures (ColPali, ColQwen2, agentic complexity-based routing) which are not yet integrated end-to-end. Framework, metrics, baselines, and collection scripts will be released open-source on acceptance.

Post-training quantization (PTQ) has become an important technique for reducing the inference cost of Large Language Models (LLMs). While recent mixed-precision methods improve ultra-low bit quantization by preserving critical subspaces in high precision, they typically construct these subspaces relying solely on activation statistics. This ignores the fundamental nature of linear operations, where the output perturbation is jointly driven by both activation and weight quantization noise. In this paper, we propose CoQuant, a joint weight-activation subspace projection method. By theoretically modeling the expected output error, CoQuant formulates a closed-form weighted PCA solution that balances activation and weight covariances to select the optimal high-precision subspace. Extensive experiments on Llama-3.2 and Qwen2.5 models show that CoQuant consistently outperforms strong PTQ baselines in both WikiText perplexity and zero-shot common-sense reasoning accuracy. These results demonstrate that joint weight-activation subspace modeling provides a principled and effective direction for low-bit LLM quantization. The source code is available at https://github.com/Zachary5895/CoQuant.

Solving large, sparse linear systems is a fundamental workload in scientific computing and engineering simulations, often dominating runtime and energy consumption in high-performance computing (HPC) applications. In this work, we explore an alternative computing paradigm based on analog optical processing, implemented through the Laser Processing Unit (LPU). The LPU encodes linear systems into the dynamics of coupled lasers within an optical cavity, where the steady-state phases of the optical fields correspond to the solution of $Ax=b$. We present a mapping of general linear systems, both dense and sparse, onto the LPU architecture and evaluate its performance using representative matrices from the SuiteSparse collection. Using an LPU emulator, we benchmark convergence behavior and time-to-solution for sparse, multi-banded matrices against established Krylov subspace methods (CG, GMRES, BiCGSTAB, and others) executed on a modern GPU platform. Our results demonstrate that the LPU will achieve significantly lower time-to-solution for selected problem classes, highlighting the potential of optical analog computing for accelerating iterative linear solvers. These findings suggest that optical processors such as the LPU will be able to serve as accelerators for linear systems, in particular structured and/or repeatedly solved, offering advantages in latency, parallelism, and energy efficiency. We discuss current limitations, including scaling constraints and precision considerations, and outline directions toward hybrid optical-digital computing systems.

We describe our system for SemEval-2026 Task 6 (CLARITY: Unmasking Political Question Evasions), which classifies English political interview responses by coarse-grained clarity (3-way) and fine-grained evasion strategy (9-way). Since responses frequently exceed the 512-token limit of standard Transformer encoders, we apply an overlapping sliding-window chunking strategy with element-wise Max-Pooling aggregation over chunk representations. A shared RoBERTa-large encoder supplies two task-specific heads trained jointly via a multi-task objective, with inference-time ensembling over 7-fold stratified cross-validation. Our system achieves a Macro-F1 of 0.80 on Subtask 1 and 0.51 on Subtask 2, ranking 11th in both subtasks.

In multi-agent systems, should limited resources be concentrated into a few capable agents or distributed among many simpler ones? This work formulates the split over $n$ resource sharing problem where a group of $n$ agents equally shares a common resource (e.g., monetary budget, computational resources, physical size). We present a case study in multi-agent coverage where the area of the disk-shaped footprint of agents scales as $1/n$. A formal analysis reveals that the initial coverage rate grows with $n$. However, if the speed of agents decreases proportionally with their radii, groups of all sizes perform equally well, whereas if it decreases proportionally with their footprints, a single agent performs best. We also present computer simulations in which resource splitting increases the failure rates of individual agents. The models and findings help identify optimal distributiveness levels and inform the design of multi-agent systems under resource constraints.

Vision-language models have shown strong performance, but they often generalize poorly to specialized domains. While semi-supervised vision-language learning mitigates this limitation by leveraging a small set of labeled image-text pairs together with abundant unlabeled images, existing methods remain fundamentally pairwise and fail to model the global structure of multimodal representation manifolds. Existing topology-based alignment methods rely on persistence diagram matching, which neither guarantees geometric alignment nor utilizes the image-text pairing information central to vision-language learning. We propose Topology-Aware Multimodal Representation Alignment (ToMA), a framework that uses persistent homology to identify topologically salient edges and aligns them across modalities through available cross-modal correspondences. ToMA leverages both H_0-death edges and lightweight H_1-birth edges, allowing it to capture both connectivity and cycle structure without constructing 2-simplices. Experiments show that ToMA yields stable gains, with clear improvements on remote sensing and modest but consistent benefits on fashion retrieval. Additional analysis shows that ToMA is more stable than alternative topology-based objectives and that lightweight H_1-birth edges provide useful higher-order structural signals.

Data quality assessment is an essential step that ensures the reliability of the subsequent structural health monitoring (SHM) tasks. This study proposes a prediction deviation-based SHM data quality assessment method using a univariate implicit auto-regressive model, enabling outlier diagnosis and data cleaning. The proposed conditional diffusion model (CDM) augments the standard diffusion model with a conditional embedding module to incorporate temporal context, quartile normalization to mitigate distribution skew, and a Huber loss to enhance robustness against outliers. Within this univariate implicit autoregressive framework, each data point is assigned an outlier probability, quantifying its degree of "outlier-ness", and a global quality evaluation score is computed to characterize the overall dataset quality. Extensive case studies utilizing operational data from real-world structures demonstrate that the proposed framework significantly improves the accuracy of data quality assessment, outperforming other strong baselines representative of clustering, isolation-based, and deep reconstruction methods. The effectiveness and robustness of the proposed framework are further demonstrated by the findings of ablation experiments and hyperparameter analysis.

To address the high sampling cost of Diffusion Transformers (DiTs), feature caching offers a training-free acceleration method. However, existing methods rely on hand-crafted forecasting formulas that fail under aggressive skipping. We propose L2P (Learnable Linear Predictor), a simple data-driven caching framework that replaces fixed coefficients with learnable per-timestep weights. Rapidly trained in ~20 seconds on a single GPU, L2P accurately reconstructs current features from past trajectories. L2P significantly outperforms existing baselines: it achieves a 4.55x FLOPs reduction and 4.15x latency speedup on FLUX.1-dev, and maintains high visual fidelity under up to 7.18x acceleration on Qwen-Image models, where prior methods show noticeable quality degradation. Our results show learning linear predictors is highly effective for efficient DiT inference. Code is available at https://github.com/Aredstone/L2P-Cache.

Federated domain generalization for person re-identification (FedDG-ReID) aims to collaboratively train a pedestrian retrieval model across multiple decentralized source domains such that it can generalize to unseen target environments without compromising raw data privacy. However, this task is significantly challenged by the inherent stylistic gaps across decentralized clients. Without global supervision, models easily succumb to shortcut learning where representations overfit to domain specific camera biases rather than universal identity features. We propose CO-EVO, a novel federated framework that resolves this semantic-style conflict through a co-evolutionary mechanism. On the semantic side, Camera-Invariant Semantic Anchoring (CSA) learns identity prompts with cross-camera consistency to establish purified and domain-agnostic anchors that filter out local imaging noise. On the visual side, Global Style Diversification (GSD), powered by a Global Camera-Style Bank (GCSB), synthesizes realistic perturbations to expand the visual boundaries of training data. The core of CO-EVO is its co-evolutionary loop where purified anchors act as gravitational centers to guide the image encoder toward robust anatomical attributes amidst diverse style variations. Extensive experiments demonstrate that CO-EVO achieves state-of-the-art (SOTA) performance, proving that the synergy between semantic purification and style expansion is essential for robust cross-domain generalization. Our code is available at: https://github.com/NanYiyuzurn/ACL-LGPS-2026.

Globalization of graphic designs such as those used in marketing materials and magazines is increasingly important for communication to broad audiences. To accomplish this, the textual content in the graphic designs needs to be accurately translated and have the text styling preserved in order to fit visually into the design. Preserving text styling requires high accuracy word alignment between the original and the translated text. The problem of word alignment between source and translated text is long known. The industry standards for extracting word alignments are defined by Giza++ and attention probabilities from neural machine translation (NMT) models. In this paper, we explore three new methods to tackle the word alignment problem for transferring text styles from the source to the translated text. The proposed methods are developed on top of commercially available NMT and LLM translation technologies. They include: NMT with custom input and output tags for text styling; LLM with custom input and output tags; a hybrid with NMT for translation followed by an LLM with use of unigram mappings. To analyze the performance of these solutions, their alignment results are compared with the results of an attention head approach to gauge their usability in graphic design applications. Interestingly, the attention head strong baseline proves more accurate than the LLM or NMT approach and on par with the hybrid NMT+LLM approach.

Reinforcement learning (RL) systems typically optimize scalar reward functions that assume precise and reliable evaluation of outcomes. However, real-world objectives--especially those derived from human preferences--are often uncertain, context-dependent, and internally inconsistent. This mismatch can lead to alignment failures such as reward hacking, over-optimization, and overconfident behavior. We introduce a dual-source uncertainty-aware reward framework that explicitly models both epistemic uncertainty in value estimation and uncertainty in human preferences. Model uncertainty is captured via ensemble disagreement over value predictions, while preference uncertainty is derived from variability in reward annotations. We combine these signals through a confidence-adjusted Reliability Filter that adaptively modulates action selection, encouraging a balance between exploitation and caution. Empirical results across multiple discrete grid configurations (6x6, 8x8, 10x10) and high-dimensional continuous control environments (Hopper-v4, Walker2d-v4) demonstrate that our approach yields more stable training dynamics and reduces exploitative behaviors under reward ambiguity, achieving a 93.7% reduction in reward-hacking behavior as measured by trap visitation frequency. We demonstrate statistical significance of these improvements and robustness under up to 30% supervisory noise, albeit with a trade-off in peak observed reward compared to unconstrained baselines. By treating uncertainty as a first-class component of the reward signal, this work offers a principled approach toward more reliable and aligned reinforcement learning systems.

Reasoning in Large Language Models incurs significant inference-time compute, yet the token-level information structure of reasoning traces remains underexplored. We observe that reasoning tokens split into two functional types: low-entropy \textit{structural} tokens (recurring phrases that scaffold the reasoning process) and higher-entropy \textit{organic} tokens (problem-specific content that drives toward a solution). This asymmetry motivates a simple, model-agnostic compression pipeline: apply cross-word BPE merges on a model's own reasoning traces to derive \textit{supertokens} that capture frequent structural patterns, then teach the model to adopt them via supervised fine-tuning. Across three model families and five mathematical reasoning benchmarks, our approach shortens reasoning traces by 8.1\% on average with no statistically significant accuracy loss on any model--benchmark pair. Beyond compression, supertokens act as interpretable reasoning-move annotations (backtracking, verification, strategy shifts), exposing the model's high-level strategy at a glance. Analyzing transitions between structural categories reveals systematic differences between correct and incorrect traces: correct traces show productive recovery (backtracking followed by strategy shifts and verification), while incorrect traces are dominated by confusion cycles (repeated hedging and unresolved contradictions). These diagnostic signals suggest applications in reward shaping and early stopping for RL-based reasoning training.

Language models (LMs) behave more like humans when their cognitive resources are restricted, particularly in predicting sentence processing costs such as reading times. However, it remains unclear whether such constraints similarly affect sentence comprehension strategies. Besides, existing methods do not directly target the balance between memory storage and sentence processing, which is central to human working memory. To address this issue, we propose a dual-task paradigm that combines an arithmetic computation task with a sentence comprehension task, such as "The 2 cocktail + blended 3 =..." Our experiments show that under dual-task conditions, GPT-4o, o3-mini, and o4-mini shift toward plausibility-based comprehension, mirroring humans' rational inference. Specifically, these models show a greater accuracy gap between plausible sentences (e.g., "The cocktail was blended by the bartender") and implausible sentences (e.g., "The bartender was blended by the cocktail") in the dual-task condition compared to the single-task conditions. These findings suggest that constraints on the balance between memory and processing resources promote rational inference in LMs. More broadly, they support the view that human-like sentence comprehension fundamentally arises from the allocation of limited cognitive resources.

We present a learning-augmented online algorithm for the preemptive FIFO buffer management problem, where packets arrive online to a finite-capacity buffer, must be transmitted in FIFO order, and the algorithm may preemptively discard buffered packets to accommodate future arrivals. Our algorithm simultaneously achieves 1-consistency, η-smoothness, and asymptotic \sqrt{3}-robustness, where ηdenotes the prediction error. Specifically, it attains an optimal competitive ratio of 1 under perfect predictions, degrades smoothly as the prediction error increases, and maintains an asymptotic competitive ratio of \sqrt{3} under arbitrarily inaccurate predictions, matching the best-known worst-case guarantee for the classical online problem, established by Englert and Westermann in 2009 [Algorithmica 53(4): 523-548]. A key technical contribution of our work is the introduction of an \emph{output-based prediction error metric}. Because capacity constraints dictate that only a strictly bounded subset of arriving packets is ultimately transmitted, our metric assesses prediction quality over the resulting optimal schedules rather than the raw input sequences, avoiding artificial error penalties. To guarantee robustness, our algorithm dynamically monitors predictions and executes a \emph{buffer-clearing strategy} upon transitioning to a worst-case fallback mechanism. We prove that the competitive loss incurred by this clearing operation is bounded by an additive capacity constant that vanishes asymptotically. Finally, we show that our algorithm provides a generalized framework for learning-augmented buffer management: substituting the fallback module with any β-competitive online algorithm immediately yields asymptotic β-robustness.

Diffusion models have achieved remarkable success in image generation, yet their training is predominantly driven by full-reference objectives that enforce pixel-wise similarity to ground-truth images.Such supervision, while effective for fidelity, may insufficient in terms of subjective visual perception quality and text-image semantic consistency. In this work, we investigate the problem of incorporating no-reference perceptual quality into diffusion training. A key challenge is that directly optimizing perceptual signals, such as those provided by no-reference image quality assessment (NR-IQA) models, introduces a mismatch with the original diffusion objective, leading to training instability and distributional drift during fine-tuning. To address this issue, we propose an anchor-constrained optimization framework that enables stable perceptual adaptation. Specifically, we leverage a learned NR-IQA model as a perceptual guidance signal, while introducing an anchor-based regularization that enforces consistency with the base diffusion model in terms of noise prediction. This design effectively balances perceptual quality improvement and generative fidelity, allowing controlled adaptation toward perceptually favorable outputs without compromising the original generative behavior. Extensive experiments demonstrate that our method consistently enhances perceptual quality while preserving generation diversity and training stability, highlighting the effectiveness of anchor-constrained perceptual optimization for diffusion models.

Objective metrics for emotional expressiveness are vital for speech generation, particularly in expressive synthesis and voice conversion requiring emotional prosody transfer. To quantify this, the field widely relies on emotion similarity between reference and generated samples. This approach computes cosine similarity of embeddings from encoders like emotion2vec, assuming they capture affective cues despite linguistic and speaker variations. We challenge this assumption through controlled adversarial tasks and human alignment tests. Despite high classification accuracy, these latent spaces are unsuitable for zero-shot similarity evaluation. Representational limitations cause linguistic and speaker interference to overshadow emotional features, degrading discriminative ability. Consequently, the metric misaligns with human perception. This acoustic vulnerability reveals it rewards acoustic mimicry over genuine emotional synthesis.

LoRA-MoE has emerged as an effective paradigm for parameter-efficient fine-tuning, combining the low training cost of LoRA with the increased adaptation capacity of Mixture-of-Experts (MoE). However, existing LoRA-MoE frameworks typically adopt a fixed and uniform expert configuration across heterogeneous Transformer modules (\eg, attention query/key projections and MLP gating networks), ignoring their distinct functional roles and capacity requirements. This design leads to localized over-provisioning, redundant trainable parameters, and unnecessary optimizer-state overhead. Moreover, prior methods enforce load balancing among experts throughout training. Although beneficial in the early stage, this constraint becomes restrictive once routing patterns stabilize, limiting expert specialization on downstream tasks. In this paper, we propose DMEP, a novel LoRA-MoE fine-tuning framework based on Dynamic Module-wise Expert Pruning. DMEP tracks expert utilization during training and physically removes low-utility experts on a per-module basis, yielding a more compact expert structure tailored to different modules. The pruned model then continues training without the load-balancing constraint, freeing the remaining experts to focus entirely on the downstream task and develop specialized expertise. By jointly adapting module-wise expert capacity and eliminating unnecessary balancing, DMEP improves both parameter efficiency and training efficiency. Extensive experiments on multiple reasoning benchmarks show that DMEP reduces trainable parameters by 35\%--43\% and improves training throughput by about 10\%, while maintaining or surpassing the downstream reasoning accuracy of uniform LoRA-MoE baselines.

Conceptual aircraft design is traditionally an expert-mediated iterative process in which a human designer proposes a configuration, runs low-order physics, inspects the result, and re-proposes. We present AlphaJet, an end-to-end automated synthesis pipeline that closes this loop. From a textual mission specification (mass, range, cruise speed, hard size envelope, engine count, areal density) AlphaJet evolves a feasible 3D aircraft in real time, scored by a transparent multi-disciplinary fitness function covering aerodynamics, structures, weights, stability, packaging, and geometric mount consistency. Three contributions distinguish our approach: (i) an Anatomically-Disentangled Variational Autoencoder (AD-VAE) whose first 25 latent dimensions are supervised to align with named anatomical parameters, providing an interpretable shape prior; (ii) a topology-elitist genetic algorithm that protects the best individual from each of five tail topologies and triggers stagnation restarts, preventing premature collapse to a single configuration; and (iii) mount-aware geometric scoring that computes signed penetration between engines and other structural parts, eliminating the redundant artifacts common in generative aircraft models. The full loop runs interactively on a CPU and streams every generation to a browser viewer, making it a practical real-world automation tool for early-phase design-space exploration.

To usher in the next round of client AI innovation, there is an urgent need to enable efficient, lossless inference of high-accuracy large language models (LLMs) and vision language models (VLMs), jointly referred to as xLMs, on client systems. To address this, we present pipelined sharding, a novel, benchmark-profile-guided CPU-GPU hybrid scheduling technique to achieve efficient, VRAM-constrained inference for both dense and mixture-of-experts (MoE) LLMs. Using a combination of model sharding at the sub-layer level, CPU offloading, pipelined copy-compute, and prioritized tensor placement in VRAM, it optimizes both time-to-first-token (TTFT) and tokens per second (TPS) metrics, while flexibly adapting to system and inference conditions. For efficient, high-accuracy VLM inference, we combine pipelined sharding with a llama.cpp implementation of three well-understood prior ideas (jointly called VLMOpt), namely, vision tensor CPU offloading, flash attention, and vision and language model VRAM overlap avoidance. These enhancements are targeted at improving client xLM inference in future releases of two important NVIDIA products - the In-Game Inferencing software development kit (IGI SDK) and the Cosmos-Reason1 (CR1) physical AI reasoning VLM. Highlights from our rigorous evaluation spanning multiple models and client systems include: for interactive use, TTFT improves by up to 6.7x and TPS by up to 30x for LLMs, and CR1 inference's VRAM demand is down by 10x, while in batched mode, throughput improves by up to 8.2x, all compared to their respective aggressive baselines. This paper is accepted at the 9th MLSys Conference (Industry Track), 2026. Code and artifact available at: https://github.com/deepshnv/pipeshard-mlsys26-ae

The rapid advancement of large language models (LLMs) presents new security challenges, particularly in detecting machine-generated text used for misinformation, impersonation, and content forgery. Most existing detection approaches struggle with robustness against adversarial perturbation, paraphrasing attacks, and domain shifts, often requiring restrictive access to model parameters or large labeled datasets. To address this, we propose DSIPA, a novel training-free framework that detects LLM-generated content by quantifying sentiment distributional stability under controlled stylistic variation. It is based on the observation that LLMs typically exhibit more emotionally consistent outputs, while human-written texts display greater affective variation. Our framework operates in a zero-shot, black-box manner, leveraging two unsupervised metrics, sentiment distribution consistency and sentiment distribution preservation, to capture these intrinsic behavioral asymmetries without the need for parameter updates or probability access. Extensive experiments are conducted on state-of-the-art proprietary and open-source models, including GPT-5.2, Gemini-1.5-pro, Claude-3, and LLaMa-3.3. Evaluations on five domains, such as news articles, programming code, student essays, academic papers, and community comments, demonstrate that DSIPA improves F1 detection scores by up to 49.89% over baseline methods. The framework exhibits superior generalizability across domains and strong resilience to adversarial conditions, providing a robust and interpretable behavioral signal for secure content identification in the evolving LLM landscape.

Reinforcement learning (RL) has unlocked complex reasoning abilities in large language models (LLMs). However, most RL algorithms suffer from performance saturation, preventing further gains as RL training scales. This problem can be characterized by the collapse of entropy, a key diagnostic for exploration in RL. Existing attempts have tried to prevent entropy collapse through regularization or clipping, but their resulting entropy curves often exhibit instability in the long term, which hinders performance gains. In this paper, we introduce Entrocraft, a simple rejection-sampling approach that realizes any user-customized entropy schedule by biasing the advantage distributions. Entrocraft requires no objective regularization and is advantage-estimator-agnostic. Theoretically, we relate per-step entropy change to the advantage distribution under minimal assumptions, which explains the behavior of existing RL and entropy-preserving methods. Entrocraft also enables a systematic study of entropy schedules, where we find that linear annealing, which starts high and decays to a slightly lower target, performs best. Empirically, Entrocraft addresses performance saturation, significantly improving generalization, output diversity, and long-term training. It enables a 4B model to outperform an 8B baseline, sustains improvement for up to 4x longer before plateauing, and raises pass@K by 50% over the baseline.

Stance detection identifies the attitude of a text author toward a given target. Recent studies have explored various LLM-based strategies for this task, from zero-shot prompting to multi-agent debate. However, existing works differ in data splits, base models, and evaluation protocols, making fair comparison difficult. We conduct a systematic comparison that evaluates five methods across two categories -- prompt-based inference (Direct Prompting, Auto-CoT, StSQA) and agent-based debate (COLA, MPRF) -- on four datasets with 14 subtasks, using 15 LLMs from six model families with parameter sizes from 7B to 72B+. Our experiments yield several findings. First, on all models with complete results, the best prompt-based method outperforms the best agent-based method, while agent methods require 7 to 12 times more API calls per sample. Second, model scale has a larger impact on performance than method choice, with gains plateauing around 32B. Third, reasoning-enhanced models (DeepSeek-R1) do not consistently outperform general models of the same size on this task.

We present VulStyle, a multi-modal software vulnerability detection model that jointly encodes function-level source code, non-terminal Abstract Syntax Tree (AST) structure, and code stylometry (CStyle) features. Prior work in code representation primarily leverages token-level models or full AST trees, often missing stylistic cues indicative of risky programming practices, or incurring high structural overhead. Our approach selects only non-terminal AST nodes, reducing input complexity while preserving semantic hierarchy, and integrates syntactic and lexical CStyle features as auxiliary vulnerability signals. VulStyle is pre-trained using masked language modeling on 4.9M functions across seven programming languages, and fine-tuned across five benchmark datasets: Devign, BigVul, DiverseVul, REVEAL, and VulDeePecker. VulStyle achieves state-of-the-art performance on BigVul and VulDeePecker, improving F1 by 4-48% over strong transformer baselines, and attains competitive or best-average performance across all benchmarks. We contribute an ablation study isolating the effect of CStyle and AST structure, error case analysis, and a threat model situating the detection task in attacker-realistic scenarios.

Public opinion towards the Free Nutritious Meal Program (MBG) on YouTube social media reflects diverse community responses. This study applies the Long Short-Term Memory (LSTM) method to classify sentiments from 7,733 YouTube comments. The results show that the LSTM model achieves 89% accuracy, with strong performance on negative sentiment (F1-score 0.94) but weaker performance on positive sentiment (F1-score 0.55) due to class imbalance, as negative data account for 87.7% of the dataset. These findings confirm the effectiveness of LSTM for sentiment analysis of Indonesian text while highlighting the challenge of imbalanced data. This research contributes to social media-based public policy evaluation

We introduce DreamProver, an agentic framework that leverages a "wake-sleep" program induction paradigm to discover reusable lemmas for formal theorem proving. Existing approaches either rely on fixed lemma libraries, which limit adaptability, or synthesize highly specific intermediate lemmas tailored to individual theorems, thereby lacking generality. DreamProver addresses this gap through an iterative two-stage process. In the wake stage, DreamProver attempts to prove theorems from a training set using the current lemma library while proposing new candidate lemmas. In the "sleep" stage, it abstracts, refines, and consolidates these candidates to compress and optimize the library. Through this alternating cycle, DreamProver progressively evolves a compact set of high-level, transferable lemmas that can be effectively used to prove unseen theorems in related domains. Experimental results demonstrate that DreamProver substantially improves proof success rates across a diverse set of mathematical benchmarks, while also producing more concise proofs and reducing computational cost.

Fine-grained emotion classification, which identifies specific emotional states such as happiness, anger, sadness, and fear, remains a challenging task in natural language processing. This study benchmarks classical machine learning and deep learning approaches for 20-class emotion classification using the 20-Emotion Text Classification Dataset containing 79,595 English sentences. On the machine learning side, Logistic Regression, Multinomial Naive Bayes, and Support Vector Machine are evaluated using TF-IDF features. On the deep learning side, Bidirectional Long Short-Term Memory, Gated Recurrent Unit, and a lightweight Transformer implemented in PyTorch are compared. The results show that BiLSTM achieves the best overall performance with 89% accuracy and a weighted F1-score of 0.89, slightly outperforming the best machine learning model, SVM, which reaches 88.11% accuracy. The findings indicate that while traditional machine learning models remain competitive and computationally efficient, sequence-based deep learning models better capture contextual emotional cues in text.

Graph Contrastive Learning (GCL) has emerged as a prominent framework for unsupervised graph representation learning. However, relying on augmentation design alone to define the invariances learned by GCL can be brittle under structural perturbations. To address this issue, we propose Cheeger--Hodge Contrastive Learning (CHCL), a framework that aligns a perturbation-stable Cheeger--Hodge joint signature across augmented views for robust graph representation learning. The proposed signature combines a Cheeger-inspired connectivity signature derived from the algebraic connectivity \(λ_2\) with the low-frequency spectrum of the 1-Hodge Laplacian, thereby capturing both global connectivity and higher-order structural information. By aligning encoder representations with the proposed Cheeger--Hodge joint signature across augmented views, CHCL learns graph embeddings that are robust to local structural perturbations. Extensive experiments on standard benchmarks, transfer settings demonstrate that CHCL consistently improves performance, robustness, and generalization.

Optimization algorithms are fundamental to modern deep learning, yet most widely used methods rely on update rules based primarily on local gradient statistics. We introduce NeuroPlastic, a plasticity-modulated optimizer that augments gradient-based updates with an adaptive multi-signal modulation mechanism inspired by multi-factor synaptic plasticity, a concept from neurobiology. NeuroPlastic dynamically scales gradient updates using interacting components that capture gradient, activity-like, and memory-like statistics, forming a lightweight modulation layer compatible with standard deep learning training pipelines. Across image classification benchmarks, NeuroPlastic consistently improves over a controlled gradient-only ablation, with more pronounced gains on the Fashion-MNIST benchmark and in reduced-data regimes. In transfer experiments on CIFAR-10 with ResNet-18, the method remains stable and competitive without retuning. These results suggest that multi-signal plasticity-inspired modulation can provide a useful extension to conventional gradient-driven optimization, particularly when learning signals are limited or noisy, and offer a promising direction for gradient-based methods in deep learning.

We present tensor and sequence parallelism (TSP), a parallel execution strategy that folds tensor parallelism and sequence parallelism onto a single device axis. In conventional multi-dimensional parallelism layouts, tensor parallelism (TP) shards model weights while sequence parallelism (SP) shards tokens, reducing per-device parameter or activation memory, respectively. Traditionally, each scheme is assigned its own mesh dimension. TSP instead assigns each rank both a weight shard and a sequence shard, reducing both parameter and activation memory along the same device axis. We implement this design with two runtime schedules. For attention, ranks iterate over broadcast parameter shards and reconstruct context through a sequence-wise key/value exchange. For gated MLPs, weight shards circulate in a ring while partial outputs accumulate locally. By sharding both weights and activations across the same devices, TSP trades additional communication volume for reduced memory overhead. We provide a theoretical communication and memory analysis, describe our implementation of TSP attention and gated MLP blocks, and benchmark TSP against TP, SP, and TP+SP. These results position TSP as a hardware-aware alternative for long-context and memory-constrained model training, and as a viable axis of parallelism in concert with existing parallelism schemes such as pipeline and expert parallelism for dense and mixture-of-expert models.

Chest X-ray interpretation is one of the most frequently performed diagnostic tasks in medicine and a primary target for AI development, yet current vision--language models are primarily trained on datasets of paired images and reports, not the cognitive processes and visual attention that underlie clinical reasoning. Here, we present CheXthought, a global, multimodal resource containing 103,592 chain-of-thought reasoning traces and 6,609,082 synchronized visual attention annotations across 50,312 multi-read chest X-rays from 501 radiologists in 71 countries. Our analysis reveals clinical reasoning patterns in how experts deploy distinct visual search strategies, integrate clinical context, and communicate uncertainty. We demonstrate the clinical utility of CheXthought across four dimensions. First, CheXthought reasoning significantly outperforms state--of--the--art vision--language model chain-of-thought in factual accuracy and spatial grounding. Second, visual attention data used as an inference--time hint recovers missed findings and significantly reduces hallucinations. Third, models trained on CheXthought data achieve significantly stronger pathology classification, visual faithfulness, temporal reasoning and uncertainty communication. Fourth, leveraging CheXthought's multi-reader annotations, we predict both human--human and human--AI disagreement directly from an image, enabling transparent communication of case difficulty, uncertainty and model reliability. These findings establish CheXthought as a resource for advancing multimodal clinical reasoning and the development of more transparent, interpretable vision--language models.

High-precision medical diagnosis relies not only on static imaging features but also on the implicit diagnostic memory experts instantly invoke during image interpretation. We pinpoint a fundamental cognitive misalignment in medical VLMs caused by discrete tokenization, leading to quantization loss, long-range information dissipation, and missing case-adaptive expertise. To bridge this gap, we propose ours, a framework for latent diagnostic memory evolution that simulates the experiential invocation of clinicians by dynamically synthesizing implicit diagnostic memories within the model's hidden stream. Specifically, it begins with a Meta Query for Prior Memorization mechanism, where learnable probes retrieve structured priors from an anatomical prior encoder to generate condensed implicit memories. To ensure clinical fidelity, we introduce Causal Counterfactual Refinement (CCR), which leverages reinforcement learning and counterfactual rewards derived from region-level feature masking to quantify the causal contribution of each memory, thereby pruning redundancies and aligning latent representations with diagnostic logic. This evolutionary process culminates in Intrinsic Memory Transition (IMT), a privileged-autonomous dual-branch paradigm that internalizes teacher-branch diagnostic patterns into the student-branch via full-vocabulary divergence alignment. Comprehensive empirical evaluations across multiple datasets demonstrate that ours, by transferring external expertise into endogenous parameters, significantly outperforms existing state-of-the-art methods, particularly chain-of-thought paradigms, in diagnostic accuracy.

To preserve or not to preserve prosody is a central question in voice anonymization. Prosody conveys meaning and affect, yet is tightly coupled with speaker identity. Existing methods either discard prosody for privacy or lack a principled mechanism to control the utility-privacy trade-off, operating at fixed design points. We propose DiffAnon, a diffusion-based anonymization method with classifier-free guidance (CFG) that provides explicit, continuous inference-time control over prosody preservation. DiffAnon refines acoustic detail over semantic embeddings of an RVQ codec, enabling smooth interpolation between anonymization strength and prosodic fidelity within a single model. To the best of our knowledge, it is the first voice anonymization framework to provide structured, interpolatable inference-time prosody control. Experiments demonstrate structured trade-off behavior, achieving strong utility while maintaining competitive privacy across controllable operating points.

The arrival of large language models (LLMs) capable of multi-step reasoning, tool use, and long-horizon planning has produced a qualitative shift in software engineering. Where earlier code-completion tools such as GitHub Copilot operated at the granularity of a line or function, modern agentic systems -- Claude Code, OpenAI Codex CLI, Google Jules, Devin, OpenHands, SWE-agent, MetaGPT, ChatDev, and DeepMind's AlphaEvolve -- operate at the granularity of a repository, a feature, or an algorithm. We synthesize work from Anthropic, OpenAI, Google DeepMind, Microsoft Research, Princeton, Stanford, and the broader academic community to characterize this transition. We propose a six-layer reference architecture for agentic software engineering systems, contrast a traditional Software Development Lifecycle (SDLC) with an emerging Agentic SDLC (A-SDLC), and consolidate empirical evidence on performance (a rise from 1.96% to 78.4% on SWE-bench Verified between October 2023 and April 2026), productivity (13.6%-55.8% time savings across controlled studies), and labor-market impact (49% of jobs sampled by Anthropic in 2026 saw AI used for at least a quarter of their tasks). We argue that the central object of inquiry has shifted from code generation to delegated execution under human supervision, and we identify five open problems -- evaluation, governance, technical debt, skill redistribution, and the economics of attention -- that will determine whether the agentic transition is net-positive for the discipline.

Structured-workflow agents driven by large language models execute tool calls against sensitive external environments. We propose \codename, a telemetry-driven behavioral anomaly detection firewall. Drawing on sequence-based intrusion detection, \codename\ compiles verified benign tool-call telemetry into a parameterized deterministic finite automaton (pDFA). The model defines permitted tool sequences, sequential contexts, and parameter bounds. At runtime, a lightweight gateway enforces these boundaries via an $O(1)$ state-transition structural lookup, shifting computationally expensive analysis entirely offline. Evaluated on the Agent Security Bench (ASB), \codename\ achieves a 5.6\% macro-averaged attack success rate (ASR) across five scenarios. Within three structured workflows, ASR drops to 2.2\%, outperforming Aegis, a state-of-the-art stateless scanner, at 12.8\%. \codename\ achieves 0\% ASR on multi-step and context-sequential attacks in structured settings. Furthermore, against 1,000 algorithmically spliced exfiltration payloads, only 1.4\% matched valid structural paths, all of which failed end-to-end string parameter guards (0 successes out of 14 surviving paths, 95\% CI [0\%, 23.2\%]). \codename\ introduces just 2.2~ms of per-call latency (a 3.7$\times$ speedup over \textsc{Aegis}) while maintaining a 2.0\% benign task failure rate (BTFR) on benign workloads. Modeling the behavioral trajectory effectively collapses the available attack surface, but unmaintained continuous parameter bounds remain vulnerable to synonym-substitution attacks (18\% evasion rate). Thus, exact-match whitelisting of sensitive parameters ultimately bears the final defensive load against execution.

The essence of good creative writing is calibrated surprise: when constraints from all relevant dimensions act together, the feasible solution space collapses into a narrow region, and the surviving choices look least predictable from an unconstrained view. "Calibrated" has a precise meaning: the author's intent, the reader's reasonable expectation, and the logic of reality converge. When these three independent judgements agree on every dimension, the set of admissible writing choices is forced into a very small region. A mathematical corollary follows: full-dimensional accuracy and mediocrity are mutually exclusive -- two sides of one constraint structure, not separate goals. We use Shannon's mutual information $I(X;Y) = H(X) - H(X|Y)$ as our analysis tool. "Calibrated" corresponds to conditional entropy going to zero; "surprise" to entropy going up; mutual information is the precise measure of the joint quantity. The argument rests on two pillars. Static: when constraints from ethos, mythos, lexis, and dianoia are imposed together, the admissible set collapses sharply, and surviving solutions show up as low-probability choices from an unconstrained view. Dynamic: the chain rule shows each writing choice is constrained by what came before and constrains what comes after; macro-level decisions naturally contribute a larger share of information, removing the need for hand-tuned weighting. Through case studies and lightweight LLM-logprob computations, we show the framework is both analytically useful and operational, laying the theoretical groundwork for Creative Quality Alignment (CQA) and a professional evaluation benchmark.

LLM workflows, which coordinate structured calls to individual LLMs (each augmented with varying instructions and tools) to achieve a particular goal, offer a promising path towards extending the capabilities of LLMs and building powerful systems that can tackle diverse tasks. However, existing approaches for building such workflows generally rely on human-crafted pipelines and prompts, which presents a substantial bottleneck in real world deployment. How can automatically induce and optimize such workflows in a data-driven way? This paper describes a simple data-driven approach for automatically inducing LLM workflows. We formulate workflow induction as a bilevel optimization problem: an outer loop which optimizes a high-level sketch of the workflow (in particular how the LLM calls should be structured), and an inner loop which optimizes each individual LLM call one-by one. Both loops are optimized with ``textual gradients'' where for the inner loop we optimize each component in a modular way through ``backpropagating'' textual gradients layer-by-layer. We find that LLM workflows discovered through our \textsc{FlowBot} (work\textbf{flow} induction through \textbf{b}ilevel \textbf{o}ptimization and \textbf{t}extual gradients) approach performs competitively against strong baselines that make use of human-crafted or automatically-generated workflows.

Reinforcement learning (RL) has become a critical paradigm for LLM post-training, yet the rollout phase -- accounting for 50--80% of total step time -- is bottlenecked by skewed generation: long-tailed trajectories indispensable for model performance block the entire training pipeline. Asynchronous training offers a natural remedy by overlapping generation with training, but introduces a fundamental tension between efficiency and algorithmic correctness. We identify three constraints in asynchronous training to preserve convergence: intra-trajectory policy consistency, data integrity, and bounded staleness. Existing approaches fail to intrinsically address the long-tailed trajectory problem, which is further exacerbated by the imbalance characteristic of Mix-of-Experts models, or deviate from the standard RL training formulation, thereby hindering model convergence. Therefore, we propose DORA (Dynamic ORchestration for Asynchronous Rollout), which addresses this challenge through algorithm-system co-design. DORA introduces multi-version streaming rollout, a novel asynchronous paradigm that maintains multiple policy versions concurrently -- simultaneously achieving full bubble elimination without compromising algorithmic constraints. Experimental results demonstrate that our DORA system achieves substantial improvements in throughput -- up to 2--3 times higher than state-of-the-art systems on open-source benchmarks -- without compromising convergence. Furthermore, in large-scale industrial applications with tens of thousands of accelerators, DORA accelerates RL training by 2--4 times compared to synchronous training across various scenarios. The resultant open-source models, LongCat-Flash-Thinking, exhibit competitive performance on complex reasoning benchmarks, matching the capability of most advanced LLMs.

We report our multi-stage framework designed for the problem of multi-class bi-atrial segmentation from 3D late gadolinium-enhanced (LGE) MRI of the human heart. The pipeline consists of a preprocessing step using multidimensional contrast limited adaptive histogram equalization (MCLAHE); coarse region segmentation from MCLAHE-enhanced and down-sampled MRI using a V-Net family model; and fine segmentation from the coarse region using another V-Net model. Asymmetric loss is adopted to optimize the model weights.

Multimodal recommendation improves user modeling by integrating collaborative signals with heterogeneous item content. In real applications, user interests evolve over time and exhibit nonstationary dynamics, where different preference factors change at different rates. This challenge is amplified in multimodal settings because visual and textual cues can dominate decisions under different temporal regimes. Despite strong progress, most multimodal recommenders still rely on static interaction graphs or coarse temporal heuristics, which limits their ability to model continuous preference evolution with fine-grained temporal adaptation. To address these limitations, we propose TimeMM, a time-conditioned spectral filtering framework for dynamic multimodal recommendation. TimeMM instantiates Time-as-Operator by mapping interaction recency to a family of parametric temporal kernels that reweight edges on the user--item graph, producing component-specific representations without explicit eigendecomposition. To capture non-stationary interests, we introduce Adaptive Spectral Filtering that mixes the operator bank according to temporal context, yielding prediction-specific effective spectral responses. To account for modality-specific temporal sensitivity, we further propose Spectral-Aware Modality Routing that calibrates visual and textual contributions conditioned on the same temporal context. Finally, a ranking-space Spectral Diversity Regularization encourages complementary expert behaviors and prevents filter-bank collapse. Extensive experiments on real-world benchmarks demonstrate that TimeMM consistently outperforms state-of-the-art multimodal recommenders while maintaining linear-time scalability.

We study generative super-resolution (SR) in real-world scenarios where content and degradations vary across domains, genres, and segments. For example, images and videos may alternate between text overlays, fast motion, smooth cartoons, and low-light faces, each benefiting from different forms of side information. Existing metadata-guided SR methods typically use a fixed conditioning design, which is suboptimal when useful cues are content dependent and transmission budgets are limited. We propose MetaSR, a Diffusion Transformer (DiT)-based framework that selects and injects task-relevant metadata to guide SR under resource constraints. Specifically, we use the DiT's own VAE and transformer backbone to fuse heterogeneous metadata, and adopt an efficient distillation strategy that enables one-step diffusion inference. Experiments across diverse content buckets and degradation regimes show that MetaSR outperforms reference solutions by up to 1.0~dB PSNR while achieving up to 50\% transmission bitrate saving at matched quality. We assess these gains under a rate--distortion optimization (RDO) framework that jointly accounts for sender-side bitrate and receiver/display quality metrics (e.g., PSNR and SSIM).

Achieving realistic human-like conversation for virtual characters requires not only a simple memorization and recall of past events, but also the strategic utilization of memory to meet factual needs and social engagement. Current memory utilization relevant (e.g., memory-augmented generation, long-term dialogue, and etc.) benchmarks overlook this nuance, treating memory primarily as a static repository of facts rather than a dynamic resource to be strategically deployed in dialogues. To address this gap, we design StratMem-Bench, a new benchmark to evaluate strategic memory use in character-centric dialogues. This dataset comprises 657 instances where virtual characters must navigate heterogeneous memory pools containing required, supportive, and irrelevant memories. We also propose a framework with different evaluation metrics including Strict Memory Compliance, Memory Integration Quality, Proactive Enrichment Score and Conditional Irrelevance Rate, to evaluate strategic memory use capabilities of virtual characters. Experiments on StratMem-Bench which leverage the state-of-the-art large language models as virtual characters show that all models perform well at distinguishing between required and irrelevant memories, but struggle once supportive memories are introduced into the decision process.

Digital biomarkers for depression have largely relied on static acoustic descriptors, pooled summary statistics, or conventional machine learning representations. Such approaches may miss nonlinear temporal organization embedded in conversational vocal dynamics. We hypothesized that depression is associated with altered recurrence structure in vocal state trajectories, reflecting changes in how the vocal system revisits acoustic states over time. Using the depression subset of the DAIC-WOZ corpus with 142 labeled participants, we modeled frame-level COVAREP trajectories as nonlinear dynamical systems and derived recurrence-based biomarkers from 74 vocal channels. Logistic regression with feature selection and stratified cross-validation evaluated classification performance. Recurrence-based biomarkers achieved a mean cross-validated AUC of 0.689, exceeding static acoustic baselines, entropy-dynamics features, Hurst exponent features, determinism features, and Lyapunov-like instability proxies. Permutation testing indicated statistical significance with $p=0.004$. Pooled cross-validated predictions yielded AUC 0.665 with a 95\% bootstrap confidence interval of [0.568, 0.758]. These findings suggest that depression may be characterized by altered recurrence structure in conversational vocal dynamics and support nonlinear state-space analysis as a promising direction for digital psychiatric biomarkers.

This study applied the Apriori algorithm to analyze behavioral interaction patterns associated with learned helplessness (LH) in mathematics tutoring system logs. Interaction data were examined across three dimensions: LH level (low vs. high), system-based intervention (with vs. without), and problem-solving outcomes (solved vs. unsolved). The analysis of the complete dataset showed that skipping problems without using hints was the most frequent pattern linked to unsolved outcomes, while persistence behaviors such as not skipping were less dominant overall. Comparisons by LH level showed that low-LH students had stronger links between problem solving and not skipping, as well as positive associations between hint use and solved outcomes. High-LH students showed more avoidance patterns, with skipping strongly tied to unsolved outcomes. In the comparison of system-based intervention conditions, students without intervention had the highest lift for persistence-success links, while the with-intervention group had stronger patterns involving skipping behaviors leading to unsolved outcomes. Outcome-specific analysis showed that not skipping was consistently associated with solved problems across all groups, while skipping without hints predicted unsolved outcomes. Practical implications and recommendations are discussed.

We introduce LATTICE, a benchmark for evaluating the decision support utility of crypto agents in realistic user-facing scenarios. Prior crypto agent benchmarks mainly focus on reasoning-based or outcome-based evaluation, but do not assess agents' ability to assist user decision-making. LATTICE addresses this gap by: (1) defining six evaluation dimensions that capture key decision support properties; (2) proposing 16 task types that span the end-to-end crypto copilot workflow; and (3) using LLM judges to automatically score agent outputs based on these dimensions and tasks. Crucially, the dimensions and tasks are designed to be evaluable at scale using LLM judges, without relying on ground truth from expert annotators or external data sources. In lieu of these dependencies, LATTICE's LLM judge rubrics can be continually audited and updated given new dimensions, tasks, criteria, and human feedback, thus promoting reliable and extensible evaluation. While other benchmarks often compare foundation models sharing a generic agent framework, we use LATTICE to assess production-level agents used in actual crypto copilot products, reflecting the importance of orchestration and UI/UX design in determining agent quality. In this paper, we evaluate six real-world crypto copilots on 1,200 diverse queries and report breakdowns across dimensions, tasks, and query categories. Our experiments show that most of the tested copilots achieve comparable aggregate scores, but differ more significantly on dimension-level and task-level performance. This pattern suggests meaningful trade-offs in decision support quality: users with different priorities may be better served by different copilots than the aggregate rankings alone would indicate. To support reproducible research, we open-source all LATTICE code and data used in this paper.

As Large Language Models (LLMs) are proposed as legal decision assistants, and even first-instance decision-makers, across a range of judicial and administrative contexts, it becomes essential to explore how they answer legal questions, and in particular the factors that lead them to decide difficult questions in one way or another. A specific feature of legal decisions is the need to respond to arguments advanced by contending parties. A legal decision-maker must be able to engage with, and respond to, including through being potentially persuaded by, arguments advanced by the parties. Conversely, they should not be unduly persuadable, influenced by a particularly compelling advocate to decide cases based on the skills of the advocates, rather than the merits of the case. We explore how frontier open- and closed-weights LLMs respond to legal arguments, reporting original experimental results examining how the quality of the advocate making those arguments affects the likelihood that a model will agree with a particular legal point of view, and exploring the factors driving these results. Our results have implications for the feasibility of adopting LLMs across legal and administrative settings.

Controllable medical video generation has achieved remarkable progress, but it still lacks interpretability, which requires the alignment of generated contents with physical priors and faithful clinical manifestations. To push the boundaries from mere controllability to interpretability, we propose DepthPilot, the first interpretable framework for colonoscopy video generation. This work takes a step toward trustworthy generation through two synergistic paradigms. To achieve explicit geometric grounding, DepthPilot devises a prior distribution alignment strategy, injecting depth constraints into the diffusion backbone via parameter-efficient fine-tuning to ensure anatomical fidelity. To enhance intrinsic nonlinear modeling under these geometric constraints, DepthPilot employs an adaptive spline denoising module, replacing fixed linear weights with learnable spline functions to capture complex spatio-temporal dynamics. Extensive evaluations across three public datasets and in-house clinical data confirm DepthPilot's robust ability to produce physically consistent videos. It achieves FID scores below 15 across all benchmarks and ranks first in clinician assessments, bridging the gap between "visually realistic" and "clinically interpretable". Moreover, DepthPilot-generated videos are expected to enable reliable 3D reconstruction, facilitating surgical navigation and blind region identification, and serve as a foundation toward the colorectal world model.

I developed a new version of Latent Semantic Scaling (LSS) employing word2vec as a masked language model. Unlike original spatial models, it assigns polarity scores to words and documents as predicted probabilities of seed words to occur in given contexts. These probabilistic polarity scores are more accurate, interpretable and consistent than those spatial polarity models can produce in text analysis. I demonstrate these advantages by applying both probabilistic and spatial models to China Daily's coverage of China and other countries during the coronavirus disease (COVID) pandemic in terms of achievement in health issues. The result suggests that more advanced masked language models would further improve the semisupervised machine learning technique.

This study compares machine learning and deep learning approaches for cyberbullying detection in Indonesian-language Instagram comments. Using a balanced dataset of 650 comments labeled as Bullying and Non-Bullying, the study evaluates Naive Bayes, Logistic Regression, and Support Vector Machine with TF-IDF features, as well as BiLSTM and BiLSTM with Bahdanau Attention. A preprocessing pipeline tailored to informal Indonesian text is applied, including slang normalization, stopword removal, and stemming. The results show that Logistic Regression performs best among the machine learning models, while BiLSTM with Attention achieves the strongest overall deep learning performance. The findings highlight the value of domain-specific preprocessing and show that although deep learning captures contextual patterns more effectively, machine learning remains a competitive option for resource-constrained deployments.

Open-vocabulary semantic segmentation (OVSS) in remote sensing images is a promising task that employs textual descriptions for identifying undefined land cover categories. Despite notable advances, existing methods typically employ a static inference paradigm, overlooking the distinct distribution of each scene, resulting in semantic ambiguity in diverse land covers and incomplete foreground activation. Motivated by this, we propose Seeking Consensus, termed SeeCo, a plug-and-play framework to boost the performance of training-free OVSS models in remote sensing images, which recalibrates arbitrary OVSS models on-the-fly by seeking dual consensus: geometric consensus learning (GCL) through multi-view consistent observations and semantic consensus learning (SCL) via textual description adaptive calibration, which assists collaborative recalibration of visual and textual semantics. The two consensus are injected via an online consensus injector (OCI), effectively alleviating the under-activation and semantic bias. SeeCo requires no specific training process, yet recalibrates semantic-geometric alignment for each unique scene during inference. Extensive experiments on eight remote sensing OVSS benchmarks show consistent gains, proving its effectiveness and universality.

Consumers are increasingly delegating purchase decisions to AI agents, providing natural-language descriptions of their preferences and identity. We argue that these representations constitute an information channel, role coherence, through which sellers can infer willingness to pay without explicit disclosure by the buyer agent, leading to preference leakage. In an experiment where a language-model buyer agent shops on behalf of a verbal consumer profile, we show that seller-side inference from dialogue alone recovers willingness to pay nearly one-for-one. Comparing this setting to a numeric-budget condition with confidentiality instructions cleanly isolates role coherence as distinct from instruction-following failure. Because this leakage arises from delegation itself, it cannot be mitigated at the prompt level. Instead, we propose architectural interventions that trade off personalization against preference privacy.

The security of open-source software repositories is increasingly threatened by next-gen software supply chain attacks. These attacks include multiphase malware execution, remote access activation, and dynamic payload generation. Traditional Machine Learning (ML) detectors struggle to detect these attacks due to the high-dimensional and sparse nature of dynamic behavioral data, including system calls, network traffic, directory access patterns, and dependency logs. As a result, these data characteristics degrade the performance, stability, and explainability of ML models. These challenges have made Deep Learning (DL) a promising alternative, given its success across various domains and its potential for modeling complex patterns. This paper presents eDySec, a DL-based efficient, stable, and explainable framework for dynamic behavioral analysis to detect malicious packages. Using the QUT-DV25 dataset, which captures both install-time and post-installation behaviors of packages, we evaluate DL models and investigate feature sets to identify the most discriminative attributes for enabling efficient malicious package detection. Additionally, model stability analysis and explainable AI techniques are incorporated into the detection pipeline to enable stable, and transparent interpretations of model decisions. Experimental results demonstrate that eDySec significantly outperforms the state-of-the-art frameworks. Specifically, it halves feature dimensionality while lowering false positives by 82% and false negatives by 79%. It also improves accuracy by 3%, achieves near-perfect stability, and maintains an inference latency of 170ms per package. Further analysis reveals that feature and model selection play a critical role, as certain combinations degrade performance. Ultimately, this study advances the understanding of the strengths and limitations of dynamic analysis against next-gen attacks.

This paper addresses the problem of anomaly detection in accounting subject association structures, proposing a structured modeling and unsupervised discriminant framework based on graph neural networks. This framework is used to mine stable correspondences between subjects and identify structural deviations from general ledger details and voucher entries. The method first abstracts accounting subjects as graph nodes, and the co-occurrence and debit/credit correspondence of subjects in the same business record are abstracted as weighted edges. The edge weights are characterized by statistical measures such as co-occurrence frequency or amount aggregation, thus forming a period-level accounting subject association graph. In the representation learning stage, a message passing mechanism is used to fuse the node's own attributes and neighborhood context to obtain node embeddings containing structural information. In the anomaly detection stage, the rationality of subject pair connections is estimated through a relation reconstruction decoder, and edge-level anomaly scores are defined based on the degree of deviation in reconstruction probabilities. These scores are then aggregated to obtain node-level risk ranking and local anomaly localization. This framework can simultaneously capture local substructure anomalies and cross-community anomaly connections without relying on anomaly labeling, outputting traceable subject pair risk clues. Comparative experiments demonstrate more stable comprehensive discriminant capabilities and higher top-ranking accuracy.

Loading high dimensional distributions is an important task for utilizing quantum computers on applications ranging from machine learning to finance. The high dimensionality leads to a curse of dimensionality, representing a d-dimensional distribution with k resolution requires dk qubits and an unstructured parameterized circuit would express a unitary in an exponential operator space in the number of qubits, leading to vanishing gradients and poor convergence guarantees even at high depth. Vine copula decompositions are widely used to represent high dimensional distributions classically, showing high quality approximation in many important applications, such as financial modeling. We present Qvine, a vine structured ansatz for quantum circuits, that mirrors the vine decomposition to construct scalable quantum circuits with efficient trainability while achieving similarly high quality approximation for amplitude encoding distributions. For regular vines (R-vines), we show that the circuit depth scales at most quadratic in the dimension of the distribution, while for D-vines, as well as many practical R-vines, the circuit depth scales linear in the dimension. For 3-dimensional and 4-dimensional Gaussians and empirical joint stock price return distributions for selected stocks, our experiments show Qvines achieve high quality loading.

In order to automate AI research we introduce a full, end-to-end framework, OMEGA: Optimizing Machine learning by Evaluating Generated Algorithms, that starts at idea generation and ends with executable code. Our system combines structured meta-prompt engineering with executable code generation to create new ML classifiers. The OMEGA framework has been utilized to generate several novel algorithms that outperform scikit-learn baselines across a robust selection of 20 benchmark datasets (infinity-bench). You can access models discussed in this paper and more in the python package: pip install omega-models.

Some text generation tasks, such as Attribute Value Extraction (AVE), require decoding multiple independent sequences from the same document context. While standard autoregressive decoding is slow due to its sequential nature, the independence between output sequences offers an opportunity for parallelism. We present Hyper-Parallel Decoding, a novel decoding algorithm that accelerates offline decoding by leveraging both shared memory and computation across batches. HPD enables out-of-order token generation through position ID manipulation, significantly improving efficiency. Experiments on AVE show that attribute-value pairs are conditionally independent, enabling us to parallelize value generation within each prompt. By further stacking multiple documents within a single prompt, we can decode in parallel up to 96 tokens per prompt. HPD works with all LLMs, and reduces both inference costs and total inference time by up to 13.8X without compromising output quality, potentially saving hundreds of thousands of dollars on industry AVE tasks. Although designed for attribute extraction, HPD makes no assumptions unique to the AVE domain and can in theory be applied to other scenarios with independent output structures.

A predecessor pilot (Cacioli, 2026) found that Llama-3-8B implements prompted sandbagging as positional collapse rather than answer avoidance. However, fixed option ordering in MMLU-Pro left open whether this reflected a model-level position-dominant policy or dataset-level distractor structure. This pre-registered follow-up (3 models, 2,000 MMLU-Pro items, 4 conditions, 24,000 primary trials) added cyclic option-order randomisation as the critical control. The pre-registered item-level same-letter diagnostic did not confirm deterministic position-tracking (same-letter rate 37.3%, below the 50% threshold). However, pre-specified supporting analyses revealed that the response-position distribution under sandbagging was highly stable under complete content rotation (Pearson r = 0.9994; Jensen-Shannon divergence = 0.027, compared to 0.386 between honest and sandbagging conditions). Accuracy spiked to 72.1% when the correct answer coincidentally occupied the preferred position E, and fell to 4.3% at position A. The data provide strong evidence for a soft distributional attractor: under sandbagging instruction, the model enters a low-entropy response-position basin centred on E/F/G that is highly stable and largely content-invariant at the aggregate level. Qwen-2.5-7B served as a negative control (non-compliant, no distributional shift). These results provide evidence, at the 7-9 billion parameter scale, that response-position entropy is a promising black-box behavioural signature of this sandbagging mode.

Large Language Model (LLM) agents are increasingly used in real-world products, where personalized and context-aware user interactions are essential. A central enabler of such capabilities is the agent's long-term semantic memory system, which extracts implicit and explicit signals from noisy longitudinal behavioral data, stores them in a structured form, and supports low-latency retrieval. Building industrial-grade long-term memory for LLM agents raises five challenges: scalability, low-latency retrieval, privacy constraints, cross-domain generalizability, and observability. We introduce the Hierarchical Long-Term Semantic Memory (HLTM) framework, which organizes textual data into a schema-aligned memory tree that captures semantic knowledge at multiple levels of granularity, enabling scalable ingestion, privacy-aware storage, low-latency retrieval, and transparent provenance; HLTM further incorporates an adaptation mechanism to generalize across diverse use cases. Extensive evaluations on LinkedIn's Hiring Assistant show that HLTM improves answer correctness and retrieval F1 significantly by more than 10%, while significantly advancing the Pareto frontier between query and indexing latency. HLTM has been deployed in LinkedIn's Hiring Assistant to power core personalization features in production hiring workflows.

Context: Empirical Software Engineering (ESE) faces increasing challenges due to data scale, methodological complexity, and reproducibility concerns. Large Language Models (LLMs) have emerged as promising tools to support empirical workflows, yet their use remains fragmented, with no comprehensive synthesis to guide responsible adoption. Aims: This study analyzes how LLMs are used in ESE, examining supported tasks, phases of the empirical lifecycle, integration into workflows, reported benefits and limitations, and the extent of reproducibility-related reporting. It also identifies gaps and future research directions. Method: We conducted a systematic literature review of peer-reviewed papers (2020-2025) across 12 leading software engineering venues, resulting in 50 primary studies analyzed through qualitative and quantitative synthesis. Results: We identified 69 LLM-assisted tasks, mainly in mining software repositories and controlled experiments, focusing on classification, filtering, and evaluation. LLMs are used across multiple phases but are concentrated in data processing and analysis. Their integration is largely automation-oriented, with limited decision-support use. Benefits emphasize efficiency and scalability, while limitations include hallucinations, inconsistency, prompt sensitivity, and reproducibility issues. Reporting practices are often incomplete. Conclusion: LLM use in ESE is growing but remains automation-driven, with gaps in human-centered integration and transparency. We outline implications and research agenda for responsible use.

We introduce Flashback, a reversible string decomposition that repeatedly peels the maximal leading and trailing character runs from a sentinel-wrapped input, recording each pair as one bilateral token. Decomposition and reconstruction both run in O(n) time and space. Our central result is a run-pairing theorem: Flashback is equivalent to pairing the first run of the string with the last, the second with the second-to-last, and so on. This gives an exact token count of 1+[r/2] for a string with r maximal runs, and matches a lower bound that holds for any admissible bilateral run-peeling scheme. From the run-pairing theorem the main structural properties follow as corollaries: the irreducible peeling kernel uses at most two symbols; palindromes are precisely the strings whose run-length encoding is symmetric with an odd number of runs; the image of the decomposition admits an explicit finite-state characterisation; and changing one run length rewrites exactly one content token.

The growing reliance of machine learning models in high-stakes, highly regulated domains such as finance and insurance has created a growing tension between predictive performance, interpretability, and regulatory fairness requirements. In these settings, models are expected not only to deliver reliable predictions but also to provide transparent decision rationales and comply with strict fairness requirements. Attention-based transformers offer powerful mechanisms for modeling complex data relationships as demonstrated in various language tasks, yet their attention mechanisms alone do not ensure counterfactually fair predictions, even when combined with fairness-aware techniques. To address these limitations, we propose the Feature Correlation Transformer (FCorrTransformer), an attention-light architecture tailored for tabular data. In this design, the attention matrix admits a direct statistical interpretation as pairwise feature dependencies, enhancing both interpretability and efficiency. Leveraging this structure, we introduce Counterfactual Attention Regularization (CAR), a framework that enforces group-invariant fair representations of sensitive features at the attention level, promoting counterfactually fair predictions without relying on explicit causal assumptions. Empirical evaluations on imbalanced classification and regression benchmarks demonstrate that FCorrTransformer combined with CAR achieves strong counterfactual fairness while maintaining competitive predictive performance and substantially reducing model complexity compared with standard transformer-based baselines. Overall, this work bridges a critical gap between fairness theory and machine learning models, offering a practical framework for responsible AI in regulatory-sensitive domains.

World models of embodied agents predict future observations conditioned on an action taken by the agent. For complex embodiments, action spaces are high-dimensional and difficult to specify: for example, precisely controlling a human agent requires specifying the motion of each joint. This makes the world model hard to control and expensive to plan with as search-based methods like CEM scale poorly with action dimensionality. To address this issue, we train a lightweight policy that maps high-level actions to sequences of low-level joint actions. Composing this policy with the frozen world model produces a lifted world model that predicts a sequence of future observations from a single high-level action. We instantiate this framework for a human-like embodiment, defining the high-level action space as a small set of 2D waypoints annotated on the current observation frame, each specifying a near-term goal position for a leaf joint (pelvis, head, hands). Waypoints are low-dimensional, visually interpretable, and easy to specify manually or to search over. We show that the lifted world model substantially outperforms searching directly in low-level joint space ($3.8\times$ lower mean joint error to the goal pose), while remaining more compute-efficient and generalizing to environments unseen by the policy.

Multimodal deep neural networks deployed in realistic environments must contend with runtime variations: changes in modality quality, overall input complexity, and available platform resources. Current networks struggle with such fluctuations -- adaptive networks cannot adhere to a strict compute budget, controller-based networks neglect to consider input complexity, and statically provisioned networks fail at all the above. Consequently, they do not extract maximum utility from the expended computational resources. We present SWAN (Sample and World-Aware Multimodal Network), the first adaptive multimodal network that accomplishes all three goals. SWAN employs a quality-aware controller to assign resources among modalities according to a variable user-specified maximum budget. Within this budget, an adaptive gating module further optimizes efficiency by scaling layer utilization according to sample complexity. For further gains, SWAN also employs a token dropping module that masks semantically irrelevant multimodal features before performing detections. We evaluate SWAN in the domain of autonomous driving with complex multi-object 3D detection, reducing FLOPs by up to 49% with minimal degradation.

With recent semantic query processing engines, semantic aggregation has become a primitive operator, enabling the reduction of a relation into a natural language aggregate using an LLM. However, the resulting semantic aggregate may contain claims that are not grounded in the underlying relation. Verifying such claims is challenging: they often involve quantifiers, groupings, and comparisons over relations that far exceed LLM context windows and require a costly combination of semantic and symbolic processing. We present Evergreen, a system that recasts claim verification as a semantic query processing task with tailored optimizations and provenance capture. Evergreen compiles each claim into a declarative semantic verification query and executes it on the same engine that produced the aggregate. To reduce cost and latency, Evergreen avoids unnecessary LLM calls through verification-aware optimizations (early stopping, relevance sorting, and estimation with confidence sequences) and general-purpose optimizations for semantic queries (operator fusion, similarity filtering, and prompt caching). Each verdict is accompanied by citations that identify a minimal set of tuples justifying the result, with semantics based on semiring provenance for first-order logic. On a benchmark of real-world restaurant review datasets reflecting production-inspired workloads, Evergreen achieves excellent verification quality (F1 = 1.00) with a strong LLM while reducing cost by 3.2x and latency by 4.0x compared to unoptimized verification. Even with a significantly weaker LLM, Evergreen outperforms a strong LLM-as-a-judge baseline in F1 at 48x lower cost and 2.3x lower latency. Relative to a retrieval-augmented agent, Evergreen compares favorably in F1 and latency with similar cost when both use a strong LLM; yet, with a much weaker LLM, it achieves the same F1 at 63x lower cost and 4.2x lower latency.

The integration of Large Language Models (LLMs) with Retrieval-Augmented Generation (RAG) has significantly advanced Knowledge Graph Question Answering (KGQA). However, existing LLM-driven KGQA systems act as stateless planners, generating retrieval plans in isolation without exploiting historical query patterns: analogous to a database system that optimizes every query from scratch without a plan cache. This fundamental design flaw leads to schema hallucinations and limited retrieval coverage. We propose CacheRAG, a systematic cache-augmented architecture for LLM-based KGQA that transforms stateless planners into continual learners. Unlike traditional database plan caching (which optimizes for frequency), CacheRAG introduces three novel design principles tailored for LLM contexts: (1) Schema-agnostic user interface: A two-stage semantic parsing framework via Intermediate Semantic Representation (ISR) enables non-expert users to interact purely in natural language, while a Backend Adapter grounds the LLM with local schema context to compile executable physical queries safely. (2) Diversity-optimized cache retrieval: A two-layer hierarchical index (Domain $\rightarrow$ Aspect) coupled with Maximal Marginal Relevance (MMR) maximizes structural variety in cached examples, effectively mitigating reasoning homogeneity. (3) Bounded heuristic expansion: Deterministic depth and breadth subgraph operators with strict complexity guarantees significantly enhance retrieval recall without risking unbounded API execution. Extensive experiments on multiple benchmarks demonstrate that CacheRAG significantly outperforms state-of-the-art baselines (e.g., +13.2% accuracy and +17.5% truthfulness on the CRAG dataset).

Domain shift, where deviations between training and deployment data distributions degrade model performance, is a key challenge in underwater environments. Existing benchmarks testing performance for underwater domain shift simulate variability through synthetic style transfer. This fails to capture intrinsic scene factors such as visibility, illumination, scene composition, or acquisition factors, limiting analysis of real-world effects. We propose a labeling framework that defines underwater domains using measurable image, scene, and acquisition characteristics. Unlike prior benchmarks, it captures physically meaningful factors, enabling semantically consistent image grouping and supporting domain-specific evaluation of detection performance including failure analysis. We validate this on public datasets, showing systematic variations across domain factors and revealing hidden failure modes.

An effective way to scale up test-time compute of large language models is to sample multiple responses and then select the best one, as in Grok Heavy and Gemini Deep Think. Existing selection methods often rely on external reward models, which requires training a strong reward model and introduces additional computation overhead. As an alternative, previous approaches have explored intrinsic signals, such as confidence and entropy, but these signals are noisy with naive aggregation. In this work, we observe that high-entropy tokens tend to cluster into consecutive groups during inference, providing a more stable notion of model uncertainty than individual tokens. Together, these clusters reveal temporal patterns of model uncertainty throughout the inference process. Motivated by this observation, we propose to use the temporal structure of uncertainty as an intrinsic reward. To this end, we first formalize the basic unit of segment-level uncertainty as the High Entropy Phase (HEP), a variable-length segment that begins at a high-entropy token and ends when consecutive low-entropy tokens appear. We then define the Entropy Centroid, inspired by the concept of the center of mass in physics, as the weighted average position of all HEPs along the trajectory. Intuitively, a lower centroid indicates early exploration followed by confident generation, which we find often corresponds to higher response quality. Based on this insight, we propose the Lowest Centroid method, which selects the response with the lowest entropy centroid among multiple candidates. Experiments on mathematics, code generation, logical reasoning, and agentic tasks, across model scales ranging from 14B to 480B, show that Lowest Centroid consistently outperforms existing baselines and delivers stable gains as model size increases. Code is available at https://github.com/hkust-nlp/entropy-centroid.

We develop a physics-informed learning framework for energy-shaping control of port-Hamiltonian (pH) systems from trajectory data. The proposed approach {co-learns} a pH system model and an optimal energy-balancing passivity-based controller (EB-PBC) through alternating optimization with policy-aware data collection. At each iteration, the system model is refined using trajectory data collected under the current control policy, and the controller is re-optimized on the updated model. Both components are parameterized by neural networks that embed the pH {dynamics} and EB-PBC structure, ensuring interpretability in terms of energy {interactions}. The learned controller renders the closed-loop system inherently passive and provably stable, and exploits passive plant dynamics without canceling the natural potential. A dissipation regularization enforces strict energy decay during training, thereby enhancing robustness to sim-to-real gaps. The proposed framework is validated on state-regulation and swing-up tasks for planar and torsional pendulum systems.

Adapting large language models (LLMs) to a targeted task efficiently and effectively remains a fundamental challenge. Such adaptation often requires iteratively improving the model toward a targeted task, yet collecting high-quality human-labeled data to support this process is costly and difficult to scale. As a result, synthetic data generation has emerged as a flexible and scalable alternative. One straightforward approach is through an iterative generation-training loop, where candidate data are synthesized through an external generator, the model is updated using these data and the process is repeated over iterations. However, generated samples can be noisy, highly redundant, or even misaligned with the targeted task distribution. Training indiscriminately on such data can dilute useful learning signals and even degrade model performance. To address this, we introduce a refined paradigm, namely an iterative generation-selection-training loop, which incorporates a selection step prior to model updates. Building on this paradigm, we propose EvoSelect, a data-efficient framework to evolve LLM effectively. Given candidate samples produced by the data generator, EvoSelect selects training data by jointly modeling targeted task alignment and diversity. We estimate task relevance through optimal transport with proxy gradient representations, which quantifies how well candidate samples align with the targeted task distribution. To mitigate redundancy, we incorporate a diversification mechanism that promotes coverage of complementary training samples. By interleaving alignment and diversification, EvoSelect enables progressive LLM evolution toward targeted tasks. Extensive experiments on various benchmarks demonstrate that with either weak or strong data generators, EvoSelect consistently improves adaptation efficacy over existing data selection methods.

Treatment allocation under budget constraints is a central challenge in digital advertising: advertisers must decide which users to show ads to while spending a limited budget wisely. The standard approach follows a two-stage offline pipeline - first collect historical data to estimate heterogeneous treatment effects (HTE), then solve a constrained optimization to allocate the budget. This works well with abundant data, but fails in cold-start settings such as new campaigns, new markets, or new customer segments where little historical data exists. We propose Budget-Constrained Causal Bandits (BCCB), an online framework that learns which users respond to ads while simultaneously spending the budget, making treatment decisions one user at a time. BCCB unifies three components into a single sequential process: learning individual-level ad effectiveness, exploring users whose response is uncertain, and pacing the budget over time. We evaluated on the Criteo Uplift dataset, a large-scale advertising dataset from a real randomized controlled trial. Our key finding is a data-efficiency crossover: offline methods require approximately 10,000 historical observations to produce reliable results, while BCCB operates effectively from the very first user. Furthermore, BCCB exhibits 3-5x lower performance variance between runs, making it more practical for real campaign planning. Among purely online methods, BCCB consistently outperforms standard Thompson Sampling, budgeted Thompson Sampling, and greedy HTE estimation across all budget levels tested.

Recent work has shown that a model's input word embeddings can serve as effective control variables for steering its behavior toward outputs that satisfy desired properties. However, this has only been demonstrated for pretrained text-completion models on the relatively simple objective of reducing surface-level profanity in short continuations. A natural and practically important question is how well input embeddings can control aligned models, which produce an imbalanced bimodal refuse-or-comply output distribution rather than the smooth distribution characteristic of open-ended generation. We explore this in the context of safety, showing that input word embeddings can be optimized in a sub-lexical manner to minimize the semantic harmfulness of aligned model responses. Our approach uses zeroth-order gradient estimation of a black-box text-moderation API with respect to the input embeddings, and then applies gradient descent on these embeddings to minimize the harmfulness of the generated text. Experiments show that the proposed method can neutralize every safety-flagged response on standard safety benchmarks.

Nonlinear Mixed Effects models (NLME) models are widely used in pharmacometrics and related fields to analyze hierarchical and longitudinal data. However, as the number of parameters and random effects increases, traditional methods for maximizing the marginal likelihood become computationally expensive. This paper explores the Variational Expectation Maximization (VEM) algorithm, a scalable alternative for fitting NLME models. Originally introduced in the context of probabilistic graphical models and later popularized through variational autoencoders, VEM has not been extensively applied to NLME modeling. By leveraging flexible variational families and reverse-mode automatic differentiation, VEM can efficiently maximize the marginal likelihood, scaling to NLME models with over 15,000 population parameters. This work provides a detailed description of VEM, compares it to other NLME fitting algorithms, and highlights its scalability through computational experiments. Using the Pumas statistical software, we fit two test models: 1) a standard warfarin model, and 2) a DeepNLME Friberg model with 15,410 population parameters and 16 random effects. The warfarin model was fitted to completion to demonstrate the correctness of VEM, while the DeepNLME Friberg model was fitted for a limited number of iterations to measure the time per iteration and demonstrate VEM's scalability.

Structural generalization in semantic parsing requires systems to apply learned compositional rules to novel structural combinations. Existing approaches either rely on hand-written algebraic rules (AM-Parser) or fail to generalize structurally (Transformer-based models). We present an alternative requiring no hand-written compositional rules, based on a neural cellular automaton (NCA) with a discrete bottleneck: all compositional rules are learned from data through local iteration. On the SLOG benchmark, the system achieves 100% type-exact match on 11 of 17 structural generalization categories, including three where AM-Parser scores 0 to 74%, with an overall standard deviation of 0.2 across 10 seeds (vs. AM-Parser's 4.3). Analysis reveals that all 5,539 failure instances reduce to exactly two mechanisms: novel combinations of wh-extraction context with reduced verb types, and modifiers appearing on the subject side of verbs.When we decompose results by CCG structural features, each sub-pattern either succeeds on all instances or fails on all. Intermediate scores (e.g., 41.4%) are mixtures of structurally distinct CCG patterns, not partial generalization.All failures correspond to directed operations absent from training; all successes correspond to operations already covered.

Heuristic design upholds modern electronic design automation (EDA) tools, yet crafting effective placement, routing, and scheduling strategies entails substantial expertise. We study how large language models (LLMs) can systematically synthesize reusable optimization heuristics beyond one-shot code generation. We propose RAG-Enhanced Kernel-Based Heuristic Synthesis (RKHS), which integrates retrieval-augmented generation (RAG), compact kernel heuristic templates, and an LLM-driven refinement loop inspired by iterative self-feedback. Applied to latency-minimizing list scheduling in high-level synthesis (HLS), a prototype reduces average schedule length by up to 11 percent over a baseline scheduler with only 1.3x runtime overhead, and the structured retrieval-synthesis loop generalizes to other EDA optimization problems.

The deployment of large language models (LLMs) in production environments has created an urgent need for observability systems that span the full stack -- from model internals to GPU kernels. Yet existing monitoring approaches address isolated layers of this stack, and no comprehensive analysis has examined how these techniques relate, overlap, or complement each other. This paper presents a structured analysis of five recent research contributions (2025-2026) that collectively define the emerging landscape of AI observability: confidence calibration via reinforcement learning (MIT), internal state monitoring through propositional probes (UC Berkeley), chain-of-thought monitorability evaluation (OpenAI), autonomous cloud operations benchmarking (Microsoft Research, UC Berkeley, UIUC), and non-intrusive inference-level tracing (TRUFFLD). We organize these contributions into a five-layer observability taxonomy, synthesize their key findings into a unified comparison, and identify four critical gaps that remain unaddressed. We further contextualize these research directions against practical operational observability systems that translate infrastructure telemetry into actionable insights for site reliability teams. Our analysis reveals that while individual monitoring layers have matured rapidly, the integration challenge -- connecting model-level confidence signals with infrastructure-level anomalies into coherent operational intelligence -- remains the defining open problem for the field.

AI agents operating on user interfaces must understand how interfaces communicate state and feedback to act reliably. As a core communicative modality, animations are increasingly used in modern interfaces, serving critical functional purposes beyond mere aesthetics. Thus, understanding UI animation is essential for comprehensive interface interpretation. However, recent studies of Vision Language Models (VLMs) for UI understanding have focused primarily on static screenshots, leaving it unclear how well these models handle dynamic UI animations. To address this gap, we created AniMINT, a novel dataset of 300 densely annotated UI animation videos. We systematically evaluate state-of-the-art VLMs on UI animation understanding, including their abilities to perceive the animation effects, identify animation purposes, and interpret animation meaning. Our results show that VLMs can reliably detect primitive motion. However, their high-level animation interpretation remains inconsistent, with substantial gaps relative to human performance. Finally, we use Motion, Context, and Perceptual Cues (MCPC) to probe factors affecting VLM performance, revealing key bottlenecks and directions for future improvement.

Accurate intraoperative assessment of glioma infiltration is essential for maximizing tumor resection while preserving functional brain tissue. Fluorescence lifetime imaging (FLIm) offers real-time, label-free biochemical contrast, but its clinical utility is challenged by biological heterogeneity, class imbalance, and variability in histopathological labeling. We present a data-centric AI (DC-AI) framework that integrates confident learning (CL), class refinement, and targeted label evaluation to develop a robust multi-class FLIm classifier for glioblastoma (GBM) resection margins. FLIm data were collected from 192 tissue margins across 31 newly diagnosed IDH-wildtype GBM patients and initially labeled into seven tumor cellularity classes by an expert neuropathologist. CL was applied to quantify FLIm point-level confidence, identify label inconsistencies, and guide iterative class merging into a three-class scheme ("low", "moderate", "high"). The resulting high-fidelity dataset enabled training a model that achieved 96% accuracy in the three-class task. SHAP analysis revealed class-specific FLIm feature importance, highlighting distinct optical signatures across the infiltration spectrum. Targeted FLIm analysis further identified biological (e.g., gray matter composition) and acquisition-related (e.g., blood contamination) contributors to low-confidence predictions. Blinded re-evaluation of margins flagged by CL demonstrated intra-pathologist variability, underscoring the value of selective relabeling rather than exhaustive review. Together, these findings demonstrate that a DC-AI framework can systematically improve data reliability, enhance model robustness, and refine biological interpretation of FLIm signals, supporting the development of clinically actionable optical tools for real-time glioma margin assessment.

AI-powered language learning tools increasingly provide instant, personalised feedback to millions of learners worldwide. However, this feedback can fail in ways that are difficult for learners--and even teachers--to detect, potentially reinforcing misconceptions and eroding learning outcomes over extended use. We present a portion of L2-Bench, a benchmark for evaluating AI systems in language education that includes (but is not limited to) six critical dimensions of effective feedback: diagnostic accuracy, awareness of appropriacy, causes of error, prioritisation, guidance for improvement, and supporting self-regulation. We analyse how AI systems can fail with respect to these dimensions. These failures, which we argue are conducive to "explainability pitfalls," are AI-generated explanations that appear helpful on the surface but are fundamentally flawed, increasing the risk of attainment, human-AI interaction, and socioaffective harms. We discuss how the specific context of language learning amplifies these risks and outline open questions we believe merit more attention when designing evaluation frameworks specifically. Our analysis aims to expand the community's understanding of both the typology of explainability pitfalls and the contextual dynamics in which they may occur in order to encourage AI developers to better design safe, trustworthy, and effective AI explanations.

First-principles atomistic simulations are essential for understanding complex material phenomena but are fundamentally limited by their computational cost. While Machine Learning Interatomic Potentials (MLIPs) have drastically improved cost for a given accuracy, their inference cost remains a bottleneck for massive systems or long timescales. To address this, we introduce a multifidelity "Mixture-of-Experts" framework based on the E(3)-equivariant Allegro architecture. Our method spatially partitions the simulation domain into a chemically complex region (e.g., reactive interfaces) and a simple region (e.g., bulk lattice), assigning models of varying capacity to each. Among the challenges in such static domain decomposition, the mechanical mismatch between models at the interface is particularly critical, as it can generate artificial stress fields and instability. We address this challenge with a co-training strategy in which the loss function includes agreement constraints -- penalties on per-atom energy and force discrepancies between models evaluated on shared bulk environments -- forcing the independent models to learn a consistent physical description of the bulk material. We validate this approach on a realistic Pt+CO catalytic system, demonstrating that the co-trained models maintain exact energy conservation, align their bulk mechanical response (e.g., equation of state and bulk modulus), and achieve predictive accuracy comparable to a full high-fidelity simulation at more than twice the computational speed.

Bug tracking systems play a crucial role in software maintenance, yet developers frequently struggle with low-quality user-submitted reports that omit essential details such as Steps to Reproduce (S2R), Observed Behavior (OB), and Expected Behavior (EB). We propose ImproBR, an LLM-based pipeline that automatically detects and improves bug reports by addressing missing, incomplete, and ambiguous S2R, OB, and EB sections. ImproBR employs a hybrid detector combining fine-tuned DistilBERT, heuristic analysis, and an LLM analyzer, guided by GPT-4o mini with section-specific few-shot prompts and a Retrieval-Augmented Generation (RAG) pipeline grounded in Minecraft Wiki domain knowledge. Evaluated on Mojira, ImproBR improved structural completeness from 7.9% to 96.4%, more than doubled the proportion of executable S2R from 28.8% to 67.6%, and raised fully reproducible bug reports from 1 to 13 across 139 challenging real-world reports.

Diffusion large language models generate text through multi-step denoising, where hallucination signals may emerge throughout the trajectory rather than only in the final output. Existing detectors mainly rely on output uncertainty or coarse trace statistics, which often fail to capture the richer hidden dynamics of D-LLMs. We propose HIVE, a hidden-evidence verification framework that extracts compressed hidden evidence from denoising trajectories, selects informative step-layer evidence, and conditions a verifier language model on the selected evidence through prefix embeddings. HIVE produces both a continuous hallucination score from verifier decision logits and structured verification outputs, including hallucination types, evidence pairs, and short rationales. Across two D-LLMs and three QA benchmarks, HIVE consistently outperforms eight strong baselines and achieves up to 0.9236 AUROC and 0.9537 AUPRC. Ablation studies further confirm the importance of hidden-evidence conditioning, learned evidence selection, two-stream evidence representation, and step-layer embeddings. These results suggest that selected hidden evidence from denoising trajectories provides a stronger and more usable hallucination signal than output-only uncertainty or coarse trace statistics.

Preserving a speaker's voice identity while generating speech in a different language remains a fundamental challenge in spoken language technology, particularly in specialized domains such as scientific communication. In this paper, we address this challenge through our system submission to the International Conference on Spoken Language Translation (IWSLT 2026), the Cross-Lingual Voice Cloning shared task. First, we evaluate several state-of-the-art voice cloning models for cross-lingual speech generation of scientific texts in Arabic, Chinese, and French. Then, we build voice cloning systems based on the OmniVoice foundation model. We employ data augmentation via multi-model ensemble distillation from the ACL 60/60 corpus. We investigate the effect of using this synthetic data for fine-tuning, demonstrating consistent improvements in intelligibility (WER and CER) across languages while preserving speaker similarity.

Informal settlements face disproportionate exposure to climate-related health hazards. However, existing methodologies lack systematic approaches to link diverse settlement characteristics with environmental health outcomes. We develop a data-driven framework to assess heat vulnerability in Rio de Janeiro's favelas by combining spatially-constrained clustering with land surface temperature (LST) analysis. Using remote sensing and geospatial features, we identify two distinct favela typologies: recent, well-connected settlements on flat terrain (Cluster 0) and historical, poorly-connected communities on vegetated slopes (Cluster 1). Analysis of 16 extreme heat events reveals systematic temperature differences of 2--3$^\circ$C between clusters, with flat-terrain favelas experiencing significantly higher heat exposure. Our findings demonstrate that settlement morphology critically influences heat vulnerability, providing a replicable framework for targeted urban planning and public health interventions in informal settlements globally.

We aim to learn a sparse and connected graph from sparse data, where the number of observations K can be substantially smaller than the signal dimension N for signals x in R^N, and the underlying distribution is unknown. In this severely ill-posed setting, we incorporate Fiedler number (the second eigenvalue of the graph Laplacian matrix that quantifies connectedness) as a robust regularization term in the sparse graph learning objective. We first develop a greedy algorithm that iteratively selects one edge globally for weakening/removal to reduce the objective, leveraging eigenvalue perturbation theorems that bound the adverse effect of an edge change to the Fiedler number. Next, we design a parallel variant, based on the Cheeger's inequality, that recursively partitions an input graph into two sub-graphs using an approximate Cheeger cut to distributedly find an optimal edge. Simulation experiments show that Fiedler number maximization robustifies sparse graph estimates, outperforming previous sparse graph learning algorithms.

Every RLHF-trained language model is shaped by a reward model, yet the mechanistic interpretability toolkit -- logit lens, direct logit attribution, activation patching, sparse autoencoders -- was built for generative LLMs whose primitives all project onto a vocabulary unembedding. Reward models replace that with a scalar regression head, breaking each tool. We present reward-lens, an open-source library that ports this toolkit to reward models, organised around one observation: the reward head's weight vector $w_r$ is the natural axis for every interpretability question. The library provides a Reward Lens, component attribution, three-mode activation patching, a reward-hacking probe suite, TopK SAE feature attribution, cross-model comparison, and five theory-grounded extensions (distortion index, divergence-aware patching, misalignment cascade detection, reward-term conflict analysis, concept-vector analysis). A ten-method adapter protocol covers Llama, Mistral, Gemma-2, and ArmoRM multi-objective heads, with a generic adapter for any HuggingFace sequence classification model. We validate on two production reward models across ~695 RewardBench pairs. The central empirical finding is negative: linear attribution does not predict causal patching effects (mean Spearman $ρ= -0.256$ on Skywork, $-0.027$ on ArmoRM). The framework treats this disagreement as a property to expose, not a bug -- motivating a design that keeps observational and causal views first-class and directly comparable.

We consider learning from labeled data collected across multiple environments, where the data distribution may vary across these environments. This problem is commonly approached from a causal perspective, seeking invariant representations that retain causal factors while discarding spurious ones. However, this framework assumes that the environment has no direct effect on the target. In contrast, we consider settings in which this assumption fails, but still aim to learn representations that support robust prediction on average across previously unseen environments. To this end, we study representations learned by explicitly modeling variation across environments and then marginalizing that variation out. We analyze the resulting representations and characterize when they are preferable to those learned by causal invariant-representation methods. We propose a concrete method based on generalized random-intercept models, a class of predictors in which such marginalization is possible, and study their generalization properties. Empirically, we show that these models outperform invariant-learning methods across a range of challenging settings.

Behavioral logs provide rich signals for user modeling, but are noisy and interleaved across diverse intents. Recent work uses LLMs to generate interpretable natural-language personas from user logs, yet evaluation often emphasizes downstream utility, providing limited assurance of persona quality itself. We propose a hierarchical framework that aggregates user actions into intent memories and induces multiple evidence-grounded personas by clustering and labeling these memories. We formulate persona induction as an optimization problem over persona quality-captured by cluster cohesion, persona-evidence alignment, and persona truthfulness-and train the persona model using a groupwise extension of Direct Preference Optimization (DPO). Experiments on a large-scale service log and two public datasets show that our method induces more coherent, evidence-grounded, and trustworthy personas, while also improving future interaction prediction.

Developers are increasingly overwhelmed by AI-generated issue reports that lack actionability and reproducibility, eroding trust in automated bug detection tools. In this paper, we present IssueSpecter, an automated tool that finds bugs in uncovered code segments and automatically generates prioritized, actionable issue reports. IssueSpecter combines coverage analysis with LLM-based defect identification, producing structured reports complete with severity ratings, reproduction steps, and suggested fixes. We evaluate IssueSpecter on 13 actively maintained Python projects, generating 10,467 issue reports. Manual annotation of the top-130 ranked issues by IssueSpecter confirms that 84.6% of the LLM-generated issues are valid or warrant further investigation, with only 15.4% false positives. LLM-based ranking outperforms rule-based ranking by 50% at P@3 and 41% in MRR. The identified bugs cover a wide variety of types, from logic and boundary errors to security vulnerabilities and state consistency bugs. By ranking issues by priority, IssueSpecter aims to help developers focus their attention on the most impactful bugs first. Finally, we validate IssueSpecter through case studies reproducing real bugs surfaced from its generated issue reports, demonstrating its practical value for automatic bug discovery in open-source Python projects. Compared against CoverUp, a state-of-the-art coverage-driven test generation tool, IssueSpecter achieves a higher bug validity rate (81.0% vs. 76.2%) under identical evaluation conditions, using the same model and the same number of evaluated artifacts per project, while additionally providing structured issue reports with reproduction steps and candidate fixes that are immediately actionable without requiring developers to interpret generated test intent.

Federated learning is a machine learning paradigm in which multiple devices collaboratively train a model under the supervision of a central server while ensuring data privacy. However, its performance is often hindered by redundant, malicious, or abnormal samples, leading to model degradation and inefficiency. To overcome these issues, we propose novel sample selection methods for image classification, employing a multitask autoencoder to estimate sample contributions through loss and feature analysis. Our approach incorporates unsupervised outlier detection, using one-class support vector machine (OCSVM), isolation forest (IF), and adaptive loss threshold (AT) methods managed by a central server to filter noisy samples on clients. We also propose a multi-class deep support vector data description (SVDD) loss controlled by a central server to enhance feature-based sample selection. We validate our methods on CIFAR10 and MNIST datasets across varying numbers of clients, non-IID distributions, and noise levels up to 40%. The results show significant accuracy improvements with loss-based sample selection, achieving gains of up to 7.02% on CIFAR10 with OCSVM and 1.83% on MNIST with AT. Additionally, our federated SVDD loss further improves feature-based sample selection, yielding accuracy gains of up to 0.99% on CIFAR10 with OCSVM. These results show the effectiveness of our methods in improving model accuracy across various client counts and noise conditions.

Forecasting benchmarks produce accuracy leaderboards but little insight into why some forecasters are more accurate than others. We introduce Bench to the Future 2 (BTF-2), 1,417 pastcasting questions with a frozen 15M-document research corpus in which agents reproducibly research and forecast offline, producing full reasoning traces. BTF-2 detects accuracy differences of 0.004 Brier score, and can distinguish differential agent strengths in research vs. judgment. We build a forecaster 0.011 Brier more accurate than any single frontier agent, and use it to evaluate agent strategic reasoning without hindsight bias. We find the better forecaster differs primarily in its pre-mortem analysis of its blind spots and consideration of black swans. Expert human forecasters found the dominant strategic reasoning failures of frontier agents are in assessing political and business leaders' incentives, judging their likelihood to follow through on stated plans, and modeling institutional processes.

All current LLM serving systems place the GPU at the center, from production-level attention-FFN disaggregation to NVIDIA's Rubin GPU-LPU heterogeneous platform. Even academic PIM/PNM proposals still treat the GPU as the central hub for cross-device communication. Yet the GPU's compute-rich architecture is fundamentally mismatched with the memory-bound nature of decode-phase attention, inflating serving latency while wasting power and die area on idle compute units. The problem is compounded as reasoning and agentic workloads push context lengths toward one million tokens, making attention latency the primary user-facing bottleneck. To address these inefficiencies, we present AMMA, a multi-chiplet, memory-centric architecture for low-latency long-context attention. AMMA replaces GPU compute dies with HBM-PNM cubes, roughly doubling the available memory bandwidth to better serve memory-bound attention workloads. To translate this bandwidth into proportional performance gains, we introduce (i) a logic-die microarchitecture that fully exploits per-cube internal bandwidth for decode attention under a minimal power and area budget, (ii) a two-level hybrid parallelism scheme, and (iii) a reordered collective flow that reduces intra-chip die-to-die communication overhead. We further conduct a design-space exploration over per-cube compute power and intra-chip D2D link bandwidth, providing actionable guidance for hardware designers. Evaluations show that AMMA achieves 15.5X lower attention latency and 6.9X lower energy consumption compared with the NVIDIA H100.

Large language model agents have achieved remarkable progress on software engineering tasks, yet current approaches suffer from a fundamental context coupling problem: the standard code editing interface conflates code inspection, modification planning, and edit execution within a single context window, forcing agents to interleave exploratory viewing with strictly formatted edit generation. This causes irrelevant information to accumulate and degrades agent performance. To address this, we propose SWE-Edit, which decomposes code editing into two specialized subagents: a Viewer that extracts task-relevant code on demand, and an Editor that executes modifications from high-level plans--allowing the main agent to focus on reasoning while delegating context-intensive operations to clean context windows. We further investigate what makes an effective editing model: observing that the prevalent find-and-replace format is error-prone, we train Qwen3-8B with GRPO to adaptively select editing modes, yielding improved editing efficiency over single-format baselines. On SWE-bench Verified, SWE-Edit improves resolved rate by 2.1% while reducing inference cost by 17.9%. We additionally propose a code editing benchmark that reliably predicts downstream agentic performance, providing practical guidance for editing model selection. Our code is publicly available at https://github.com/microsoft/SWE-Edit.

Graph neural networks (GNNs) have emerged as a versatile and efficient option for modeling the dynamic behavior of deformable materials. While GNNs generalize readily to arbitrary shapes, mesh topologies, and material parameters, existing architectures struggle to correctly predict the temporal evolution of key physical quantities such as linear and angular momentum. In this work, we propose MomentumGNN -- a novel architecture designed to accurately track momentum by construction. Unlike existing GNNs that output unconstrained nodal accelerations, our model predicts per-edge stretching and bending impulses which guarantee the preservation of linear and angular momentum. We train our network in an unsupervised fashion using a physics-based loss, and we show that our method outperforms baselines in a number of common scenarios where momentum plays a pivotal role.

{Closed-loop inverse source localization and characterization (ISLC) requires a mobile agent to select measurements that localize sources and infer latent field parameters under strict time constraints.} {The core challenge lies in the belief-space objective: valid uncertainty estimation requires expensive Bayesian inference, whereas using fast learned belief model leads to reward hacking, in which the policy exploits approximation errors rather than actually reducing uncertainty.} {We propose \textbf{Distill-Belief}, a teacher--student framework that decouples correctness from efficiency. A Bayes-correct particle-filter teacher maintains the posterior and supplies a dense information-gain signal, while a compact student distills the posterior into belief statistics for control and an uncertainty certificate for stopping. At deployment, only the student is used, yielding constant per-step cost.} {Experiments on seven field modalities and two stress tests show that Distill-Belief consistently reduces sensing cost and improves success, posterior contraction, and estimation accuracy over baselines, while mitigating reward hacking.}

As blockchain ecosystems grow, financially motivated attackers increasingly exploit decentralized finance (DeFi) protocols, causing frequent and severe losses. Unlike conventional cyberattacks, DeFi exploits propagate rapidly due to the transparent and composable nature of smart contracts. We identify a critical pattern, Imitative Attack Cascade: an initial successful exploit is quickly followed by mimicking transactions that reuse attack logic with minor modifications or parameter changes. Our empirical analysis shows that over 69% of DeFi attacks exhibit strong behavioral similarity to earlier incidents, often within hours or days of the initial attack. This exposes a fundamental limitation in current reactive detection. Initial attacks are typically flagged via heuristic alerts (Tornado Cash traces, anomalous nonce usage, exploiter labels), but turning these signals into detection rules requires manual validation and handcrafted trace analysis -- a labor-intensive, slow process that leaves follow-up attacks to spread. Our goal is to ensure that once an attack has been observed, even a single instance, it can be rapidly abstracted into an actionable, generalizable detection rule. We decompose the problem into two challenges: (I) abstracting the semantics of diverse, obscure function signatures, and (II) matching transaction logic in noisy, evasive traces. We leverage two insights: (i) the open-source nature of most DeFi protocols enables high-fidelity semantic classification of function signatures; (ii) contract labels isolate essential logic by filtering irrelevant calls and classifying attack intent. Building on these, we develop GenDetect, which achieves ACC 98%, FPR 1%, FNR 3% and discovers 56 previously unrevealed attacks from the past three years. Source code and dataset: https://github.com/NobodyIsAnonymous/GenDetect_ICSE2026

We study reliability in autonomous language-model agents that translate user mandates into validated tool actions under real capital. The setting is DX Terminal Pro, a 21-day deployment in which 3,505 user-funded agents traded real ETH in a bounded onchain market. Users configured vaults through structured controls and natural-language strategies, but only agents could choose normal buy/sell trades. The system produced 7.5M agent invocations, roughly 300K onchain actions, about $20M in volume, more than 5,000 ETH deployed, roughly 70B inference tokens, and 99.9% settlement success for policy-valid submitted transactions. Long-running agents accumulated thousands of sequential decisions, including 6,000+ prompt-state-action cycles for continuously active agents, yielding a large-scale trace from user mandate to rendered prompt, reasoning, validation, portfolio state, and settlement. Reliability did not come from the base model alone; it emerged from the operating layer around the model: prompt compilation, typed controls, policy validation, execution guards, memory design, and trace-level observability. Pre-launch testing exposed failures that text-only benchmarks rarely measure, including fabricated trading rules, fee paralysis, numeric anchoring, cadence trading, and misread tokenomics. Targeted harness changes reduced fabricated sell rules from 57% to 3%, reduced fee-led observations from 32.5% to below 10%, and increased capital deployment from 42.9% to 78.0% in an affected test population. We show that capital-managing agents should be evaluated across the full path from user mandate to prompt, validated action, and settlement.

Accurate fruit maturity identification is essential for determining harvest timing, as incorrect assessment directly affects yield and post-harvest quality. Although ripening is a continuous biological process, vision-based maturity estimation is typically formulated as a multi-class classification task, which imposes sharp boundaries between visually similar stages. To examine this limitation, we perform an annotation reliability study with two independent annotators on a held-out tomato dataset and observe disagreement concentrated near adjacent maturity stages. Motivated by this observation, we model maturity as a latent continuous variable and predict it probabilistically using a distributional detection head, converting the distribution into class probabilities through the cumulative distribution function (CDF). The proposed formulation maintains comparable performance to a standard detector under clean labels while better representing uncertainty. Furthermore, when controlled label noise is introduced during training, the probabilistic model demonstrates improved robustness relative to the baseline, indicating that explicitly modeling maturity uncertainty leads to more reliable visual maturity estimation.

Current and future applications demand ultra-low latency and consistent throughput, yet frequently traverse 5G cellular networks, so cope with volatile packet dynamics, as 5G base station schedulers dynamically react to user workloads and wireless channel conditions. The task of evaluating network algorithms in these environments is hamstrung by current tools: record-and-replay emulators sever the feedback interaction that exists between application end points and a commercial operator's proprietary 5G scheduler, while full-stack simulators rely on overly simplistic scheduling logic. To bridge this reality gap, we present NeuralEmu, a high-fidelity, machine learning-based emulation framework that learns complex 5G scheduler resource allocation behaviors directly from extremely high-resolution network telemetry tools. The first emulator to handle multiple clients, NeuralEmu utilizes machine learning to dynamically predict resource block allocations and modulation schemes based on instantaneous user buffer occupancy and channel states. To capture realistic cross-user contention, a traffic reconstruction model inverts cellular network scheduling results to recover the underlying traffic patterns of uncontrolled background users. Implemented as an high-performance Linux middlebox emulator, NeuralEmu reduces emulation error relative to the state of the art for various network applications including but not limited to 55% for web-page load time, 57% for WebRTC encoder bit rate, and 51% for cloud gaming packet one-way delay, providing an accurate, standardized testing ground for tomorrow's real-time interactive network protocols and applications.

Photoplethysmography (PPG) is increasingly used in wearable affective computing due to its low cost and ease of integration into consumer devices. Recent advances in deep learning have introduced long-range sequence models, such as Transformers, and state-space models, like Mamba, which have demonstrated strong performance on natural language and general time-series tasks. However, it remains unclear whether these architectures offer tangible benefits over widely used Convolutional Neural Networks (CNNs) and Long Short-Term Memory (LSTMs) for PPG-based affect recognition, given that datasets are typically small and noisy. This work presents a measurement-driven comparison of four deep learning architectures, CNN, CNN-LSTM hybrid, Transformers, and Mamba, for classifying arousal, valence, and relaxation states from wrist-based PPG signals. All models are evaluated under a subject-independent 5-fold cross-validation protocol using identical preprocessing, segmentation, and training pipelines. Our results show that the Transformer and Mamba models achieve performance comparable to that of a CNN baseline, but do not consistently outperform it across all tasks. CNNs remain the most effective overall, providing the highest accuracy with the smallest model size, whereas Transformers have a better balance of F1 scores for Arousal and Relaxation. The study provides the first evaluation of Transformer and Mamba models for PPG-based affect recognition, offering practical guidance on model selection for wearable affective monitoring systems.

LLM inference is constrained by GPU memory capacity and bandwidth. Tiered memory architectures mitigate this by allowing the GPU to offload memory to the remote tier. However, existing memory offloading frameworks rely on prefetching data into local GPU HBM. This approach underutilizes system resources by introducing HBM contention, squandering memory capacity, and creating pipeline bubbles. We show that enabling direct GPU access to remote memory significantly outperforms prefetching, achieving optimal aggregate system bandwidth. We propose DAK, an end-to-end direct-access memory offloading framework that repurposes the Tensor Memory Accelerator (TMA) to asynchronously fetch offloaded weights and KV caches directly from remote memory into GPU shared memory (SMEM). To maximize remote access performance, DAK introduces a greedy algorithm to determine optimal per-operation offloading ratios, alongside active congestion control and TMA multicast to eliminate interconnect bottlenecks and read amplification. Evaluations across diverse architectures show that DAK achieves near-optimal bandwidth aggregation, with up to 3$\times$ performance gains on NVLink-C2C and 1.8$\times$ on PCIe systems compared to state-of-the-art memory offloading baselines.

Industrial chemical plants often operate under strict data confidentiality constraints, making centralized data-driven process modeling difficult. Federated learning (FL) provides a promising solution by enabling collaborative model training across distributed facilities without sharing raw operational data. This paper proposes a privacy-preserving federated learning framework for distributed chemical process optimization using data collected from multiple geographically separated plants. Each plant locally trains a neural-network-based process model using its own time-series sensor data, while only model parameters are transmitted to a central aggregation server through secure aggregation mechanisms. This design allows cross-plant knowledge sharing while maintaining strict data locality and industrial confidentiality. Experimental evaluation was conducted using process datasets from three independent chemical plants operating under heterogeneous conditions. The results demonstrate rapid convergence of the federated model, with the global mean squared error decreasing from approximately 2369 to below 50 within the first five communication rounds and stabilizing around 35 after 40 rounds. In comparison with local-only training, the proposed federated framework significantly improves prediction accuracy across all plants, while achieving performance comparable to centralized training. The findings indicate that federated learning provides an effective and scalable solution for collaborative industrial analytics, enabling privacy-preserving predictive modeling and process optimization across distributed chemical production facilities.

Causal inference in continuous-time sequential decision problems is challenged by hidden confounders. We show that, in latent state-space models with time-varying interventions, observability of the latent dynamics from observed data is necessary for identifying dynamic treatment effects, linking control-theoretic observability to causal identifiability, even when hidden confounders affect both treatments and outcomes. We derive a continuous-time adjustment formula expressing potential outcome distributions under treatment trajectories via the measurement model, latent dynamics, and the filtering distribution over latent states given observed histories. We propose Observable Neural ODEs (ObsNODEs), Neural ODE models in observable normal form for causal forecasting. ObsNODEs learn continuous-time dynamics with states reconstructible from observations, enabling outcome prediction under alternative treatment paths. Experiments on synthetic cancer data, semi-synthetic data based on MIMIC-IV, and real-world sepsis data show strong performance over recent sequence models.

Algorithms with predictions is a growing area that aims to leverage machine-learned predictions to design faster beyond-worst-case algorithms. In this paper, we use this framework to design a learned data structure for the incremental strongly connected components (SCC) problem. In this problem, the $n$ vertices of a graph are known a priori and the $m$ directed edges arrive over time. The goal is to efficiently maintain the strongly connected components of the graph after each insert. Our algorithm receives a possibly erroneous prediction of the edge sequence and uses it to precompute partial solutions to support fast inserts. We show that our algorithm achieves nearly optimal bounds with good predictions and its performance smoothly degrades with the prediction error. We also implement our data structure and perform experiments on real datasets. Our empirical results show that the theory is predictive of practical runtime improvements.

Supervised contrastive learning (SupCon) is widely used to shape representations, but has seen limited targeted study for audio deepfake detection. Existing work typically combines contrastive terms with broader pipelines; however, the focus on SupCon itself is missing. In this work, we run a controlled study on wav2vec2 XLS-R (300M) that varies (i) similarity in SupCon (cosine vs angular similarity derived from the hyperspherical angle) and (ii) negative scaling using a warm-started global cross-batch queue. Stage 1 fine-tunes the encoder and projection head with SupCon; Stage 2 freezes them and trains a linear classifier with BCE. Trained on ASVspoof 2019 LA and evaluated on ASV19 eval plus ITW and ASVspoof 2021 DF/LA, Cosine SupCon with a delayed queue achieves the best ITW EER (8.29%) and pooled EER (4.44), while angular similarity performs strongly without queued negatives (ITW 8.70), indicating reduced reliance on large negative sets.

Autonomous agents acting in realistic Multi-Agent Systems (MAS) should be able to adapt during their execution. Standard strategic logics, such as Alternating-time Temporal Logic (ATL), model agents' state- or history-dependent behaviour. However, the dynamic treatment of agents' available actions and their knowledge of required actions is still rarely addressed. In this paper, we introduce ATL with Dynamic Actions (ATL-D), which models the process of granting and revoking actions, and its extension ATEL-D, which captures how such updates affect agents' knowledge. Beyond the conceptual contribution, we provide several technical results: we analyse the expressivity of our logic in relation to ATL, study its relation to normative systems, and provide complexity results for relevant computational problems.

Safety evaluations of large language models (LLMs) typically report binary outcomes such as attack success rate, refusal rate, or harmful/not-harmful response classification. While useful, these can hide how risk changes between a user's input and the model's response. We present a paired, transition-based analysis over 1250 prompt-response records with human-provided labels over four harm categories (Hate, Sexual, Violence, Self-harm) and ordinal severity levels aligned with the Azure AI Content Safety taxonomy. 61% of responses de-escalate harm relative to the prompt, 36% preserve the same severity, and 3% escalate to higher harm. A per-category persistence/drift-up decomposition identifies Sexual content as 3x harder to de-escalate than Hate or Violence, driven by persistence on already-sexual prompts, not by newly introducing sexual harm from benign inputs. Jointly measuring response relevance reveals an empirical signature of the helpfulness-harmlessness tradeoff: all compliance-escalation cases (from non-zero prompts) are relevance-3 (high-quality, on-task content at elevated severity), while medium-severity responses show the lowest relevance (64%), driven by tangential elaborations in Violence and Sexual categories.

The increasing number of satellites has improved the temporal resolution of Earth observation, making satellite-based flood mapping a promising approach for operational flood monitoring. Deep learning-based approaches for flood mapping using satellite imagery, an important application within Geospatial Artificial Intelligence (GeoAI), have shown improved predictive performance by learning complex spatial and spectral patterns from large volumes of remote sensing data. However, the opaque decision-making processes of deep learning models remain a major barrier to their integration into critical scientific and operational workflows. This highlights the need for a systematic assessment of whether model explanations align with established domain knowledge in remote sensing. To address this research gap, this study introduces the ADAGE (Alignment between Domain Knowledge And GeoAI Explanation Evaluation) framework. The proposed framework is designed to systematically evaluate how well explanations of deep learning models align with established remote sensing knowledge, particularly regarding the distinctive spectral properties of the Earth's surface. The ADAGE framework employs Channel-Group SHAP (SHapley Additive exPlanations) method to estimate the contributions of grouped input channels to pixel-level predictions. Experiments on two satellite-based flood mapping tasks demonstrate that the ADAGE framework can (1) quantitatively assess the alignment between model explanations and reference explanations derived from domain knowledge and (2) help domain experts identify misaligned explanations through alignment scores. This study contributes to bridging the gap between explainability and domain knowledge in GeoAI for Earth observation, enhancing the applicability of GeoAI models in scientific and operational workflows.

This paper presents a principled and scalable framework for systematically generating complex Question Answering (QA) data. In the core of this framework is a graphlet-anchored generation process, where small subgraphs from a Knowledge Graph (KG) are used in a structured prompt to control the complexity and ensure the factual grounding of questions generated by Large Language Models. The first instantiation of this framework is BioGraphletQA, a new biomedical KGQA dataset of 119,856 QA pairs. Each entry is grounded in a graphlet of up to five nodes from the OREGANO KG, with most of the pairs being enriched with relevant document snippets from PubMed. We start by demonstrating the framework's value and the dataset's quality through evaluation by a domain expert on 106 QA pairs, confirming the high scientific validity and complexity of the generated data. Secondly, we establish its practical utility by showing that augmenting downstream benchmarks with our data improves accuracy on PubMedQA from 49.2% to 68.5% in a low-resource setting, and on MedQA from a 41.4% baseline to 44.8% in a full-resource setting. Our framework provides a robust and generalizable solution for creating critical resources to advance complex QA tasks, including MCQA and KGQA. All resources supporting this work, including the dataset (https://zenodo.org/records/17381119) and framework code (https://github.com/ieeta-pt/BioGraphletQA), are publicly available to facilitate use, reproducibility and extension.

The optimal kernel configuration for Mixture-of-Experts (MoE) inference depends on both batch size and the expert routing distribution, yet production systems dispatch from batch size alone, leaving 10-70% of kernel throughput unrealized. We present RaMP, a routing-aware dispatch framework. A performance-region analysis derives, from hardware constants alone, when each optimization helps, correctly predicting all 8 tested architectures, including 3 unseen. A four-parameter wave cost model selects the fastest configuration from the runtime expert histogram, achieving 0.93% mean regret versus exhaustive search, fitted from just 10-24 minutes of one-time profiling per model. Because the model depends only on CTA grid geometry, it is kernel-agnostic: applied to Alpha-MoE, it delivers 1.14x with no source modification. Paired with a co-designed CuTe DSL kernel exposing 134-268 polymorphic configurations, RaMP delivers 1.22x kernel speedup over static dispatch and 1.30x end-to-end speedup in vLLM serving over Triton, 1.41x over DeepGEMM, and 1.13x over FlashInfer CUTLASS.

Class-level evaluation can conceal substantial performance disparities across subconcepts within the same class, causing models that perform well on average to fail on specific subpopulations. Prior work has shown that common evaluation measures for imbalanced classification are biased toward larger minority subconcepts and that utility-based reweighting using true subconcept labels can mitigate this bias; however, such labels are rarely available at test time. We introduce a practical utility-weighted evaluation that replaces unavailable subconcept labels with predicted posterior probabilities from a multiclass subconcept model. Evaluation weights are defined as the expected utility under this posterior, yielding a soft, uncertainty-aware metric we call predicted-weighted balanced accuracy (pBA). Experiments on tabular benchmarks as well as medical-imaging and text datasets show that unweighted scores can be misleading under within-class heterogeneity, while pBA provides more stable and interpretable assessments when subconcept distributions are uneven but not pathological. Our code is available at: https://anonymous.4open.science/r/correcting-bias-imbalance-9C6C/.

Usability testing with experts and potential users can assess the effectiveness, efficiency, and user satisfaction of graphical user interfaces (GUIs) but doing so remains a costly and time-intensive process. Prior work has used computer use agents (CUAs) and other generative agents that can simulate user interactions and preference, but we show that agents still struggle to provide accurate usability assessments. In this work, we present a novel machine learning method that operationalizes a computational definition of usability to train CUAs to assess GUI usability by i) prioritizing important interaction flows, ii) executing them through human-like interactions, and iii) predicting a learned numerical usability score. We train a computer use agent, uxCUA, with our algorithm on a large-scale dataset of fully interactive user interfaces (UIs) paired with usability labels and human preferences. We show that uxCUA outperforms larger models in accurate usability assessments and produces realistic critiques of both synthetic and real UIs. More broadly, our work aims to build a principled, data-driven foundation for automated usability assessment in HCI.

We introduce QERNEL, a foundational neural wavefunction that variationally solves families of parameterized many-electron Hamiltonians and captures their ground states throughout parameter space within a single model. QERNEL combines FiLM-based parameter conditioning with scale-efficient architectural elements -- mixture of experts and grouped-query attention, substantially improving expressivity at low computational cost. We apply QERNEL to interacting electrons in semiconductor moiré heterobilayers, training a single weight-shared model for systems of up to 150 electrons. By solving the many-electron Schrödinger equation conditioned on moiré potential depth, QERNEL captures both quantum liquid and crystal states and discovers the sharp phase transition between them, marked by abrupt changes in interaction energy and charge density. Our work establishes a foundation model for moiré quantum materials and a scalable architecture toward a Large Electron Model for solids.

Parallel processing of information plays a critical role in accelerating computation. This includes quantum computers, where parallel processing of quantum information will play a critical role in practical quantum advantage. Here, we demonstrate a new type of parallel entangling gates in a trapped-ion quantum computer, that simultaneously provides efficient gate-pulse synthesis and calibration, as well as graph-pattern-agnostic implementation. We demonstrate the resulting reduced execution time in three well-known algorithms, exhibiting disjoint gates, a star graph and a ring graph respectively. For disjoint qubit pairs the execution time of our parallel gates is comparable to that of a single-pair entangling gate resulting in an approximately linear speed up. For all graph patterns our parallel gate fidelities are comparable to the fidelity of a single-pair entangling gate. These advantages motivate architectures featuring multiple medium length ion chains in future quantum computing devices.

Recursive or looped language models have recently emerged as a new scaling axis by iteratively refining the same model computation over latent states to deepen reasoning. We extend such scaling principle from a single model to multi-agent systems, and ask: Can agent collaboration itself be scaled through recursion? To this end, we introduce RecursiveMAS, a recursive multi-agent framework that casts the entire system as a unified latent-space recursive computation. RecursiveMAS connects heterogeneous agents as a collaboration loop through the lightweight RecursiveLink module, enabling in-distribution latent thoughts generation and cross-agent latent state transfer. To optimize our framework, we develop an inner-outer loop learning algorithm for iterative whole-system co-optimization through shared gradient-based credit assignment across recursion rounds. Theoretical analyses of runtime complexity and learning dynamics establish that RecursiveMAS is more efficient than standard text-based MAS and maintains stable gradients during recursive training. Empirically, we instantiate RecursiveMAS under 4 representative agent collaboration patterns and evaluate across 9 benchmarks spanning mathematics, science, medicine, search, and code generation. In comparison with advanced single/multi-agent and recursive computation baselines, RecursiveMAS consistently delivers an average accuracy improvement of 8.3%, together with 1.2$\times$-2.4$\times$ end-to-end inference speedup, and 34.6%-75.6% token usage reduction. Code and Data are provided in https://recursivemas.github.io.

Real-world data visualization (DV) requires native environmental grounding, cross-platform evolution, and proactive intent alignment. Yet, existing benchmarks often suffer from code-sandbox confinement, single-language creation-only tasks, and assumption of perfect intent. To bridge these gaps, we introduce DV-World, a benchmark of 260 tasks designed to evaluate DV agents across real-world professional lifecycles. DV-World spans three domains: DV-Sheet for native spreadsheet manipulation including chart and dashboard creation as well as diagnostic repair; DV-Evolution for adapting and restructuring reference visual artifacts to fit new data across diverse programming paradigms and DV-Interact for proactive intent alignment with a user simulator that mimics real-world ambiguous requirements. Our hybrid evaluation framework integrates Table-value Alignment for numerical precision and MLLM-as-a-Judge with rubrics for semantic-visual assessment. Experiments reveal that state-of-the-art models achieve less than 50% overall performance, exposing critical deficits in handling the complex challenges of real-world data visualization. DV-World provides a realistic testbed to steer development toward the versatile expertise required in enterprise workflows. Our data and code are available at \href{https://github.com/DA-Open/DV-World}{this project page}.

Adapting reasoning models to new tasks during post-training with only output-level supervision stalls under reinforcement learning from verifiable rewards (RLVR) when the initial success probability $p_0$ is small. Using the Tsallis $q$-logarithm, we define a loss family $J_Q$ that interpolates between RLVR (at $q{=}0$, the exploitation pole) and the log-marginal-likelihood over latent trajectories (at $q{=}1$, the density-estimation pole). All members share the same per-example gradient direction, differing only by a scalar amplification $P_{θ^{-q}}$ that reweights each instance independently of the learning rate. This amplification is the mechanism that addresses cold-start stalling: under gradient flow, the exploitation pole requires $Ω(\frac{1}{p_0})$ time to escape cold start, while the density-estimation pole escapes in $Θ\big(\log(\frac{1}{p_0})\big)$; intermediate $q$ trades escape speed against noise memorization. Because $P_θ$ is intractable, we derive two Monte Carlo estimators from the two factorizations of the gradient: Gradient-Amplified RL (GARL) samples from the prior and amplifies the RL gradient, and Posterior-Attenuated Fine-Tuning (PAFT) importance-resamples from the posterior and runs standard SFT. Both have bias $O\big(\frac{q}{M P_θ^{q+1}}\big)$; GARL has lower variance, PAFT has semantically coherent gradients. On FinQA, HotPotQA, and MuSiQue, GARL at $q{=}0.75$ substantially mitigates cold-start stalling, escaping cold start where GRPO fails entirely. In warm start, GARL at low $q$ dominates FinQA where training is stable; on HotPotQA and MuSiQue, GARL destabilizes during training, and PAFT at $q{=}0.75$ provides stable gradients (best overall on HotPotQA at 47.9 maj@16, $+14.4$ over GRPO).

How much does a user's skill with AI shape what AI actually delivers for them? This question is critical for users, AI product builders, and society at large, but it remains underexplored. Using a richly annotated sample of 27K transcripts from WildChat-4.8M, we show that fluent users take on more complex tasks than novices and adopt a fundamentally different interactional mode: they iterate collaboratively with the AI, refining goals and critically assessing outputs, whereas novices take a passive stance. These differences lead to a paradox of AI fluency: fluent users experience more failures than novices -- but their failures tend to be visible (a direct consequence of their engagement), they are more likely to lead to partial recovery, and they occur alongside greater success on complex tasks. Novices, by contrast, more often experience invisible failures: conversations that appear to end successfully but in fact miss the mark. Taken together, these results reframe what success with AI depends on. Individuals should adopt a stance of active engagement rather than passive acceptance. AI product builders should recognize that they are designing not just model behavior but user behavior; encouraging deep engagement, rather than friction-free experiences, will lead to more success overall. Our code and data are available at https://github.com/bigspinai/bigspin-fluency-outcomes

Identity teacher forcing (ITF) enables stable training of deterministic recurrent surrogates for chaotic dynamical systems and has been highly effective for dynamical systems reconstruction (DSR) with recurrent neural networks (RNNs), including interpretable almost-linear RNNs (AL-RNNs). However, as an intervention-based prediction loss (and thus a generalized Bayes update), teacher forcing need not match the free-running model's marginal likelihood geometry. We compare the objective-induced curvatures of ITF and marginal likelihood in a probabilistic switching augmentation of AL-RNNs, estimating ambiguity-aware observed information via Louis' identity. In the switching setting studied here, conditioning on a single forced regime path (as ITF does) inflates curvature, while marginal likelihood curvature is reduced by a missing-information correction when multiple switching explanations remain plausible. In Lorenz-63 experiments, windowed evidence fine-tuning improves held-out evidence but can degrade dynamical quantities of interest (QoIs) relative to ITF-pretrained models.

The accelerating adoption of Large Language Models (LLMs) in software engineering (SE) has brought with it a silent crisis: unsustainable computational cost. While these models demonstrate remarkable capabilities in different SE tasks, they are unmanageably large, slow to deploy, memory-intensive, and carbon-heavy. This reality threatens not only the scalability and accessibility of AI-powered SE, but also its long-term environmental sustainability. The research challenge is clear: we must go beyond accuracy and address efficiency and environmental cost as first-class design constraints. To meet this challenge, we introduce Carbon-Taxed Transformers (CTT), a systematic multi-architectural compression principled pipeline ordering inspired by economic carbon taxation principles. Drawing from the economic concept of carbon pricing, CTT operationalizes a computational carbon tax that penalizes architectural inefficiencies and rewards deployment-ready compression. We evaluate CTT across three core SE tasks: code clone detection, code summarization, and code generation, with models spanning encoder-only, encoder-decoder, and decoder-only architecture. Our results show that CTT delivers on inference: (1) up to 49x memory reduction, (2) time reduction up to 8-10x for clone detection, up to 3x for summarization, and 4-7x for generation, (3) up to 81% reduction in CO2 emissions and (4) CTT retains around 98% accuracy on clone detection, around 89% on summarization, and up to 91% (textual metrics) and 68% (pass@1) for generation. Two ablation studies show that pipeline ordering and individual component contributions are both essential, providing empirical justification for CTT's design and effectiveness. This work establishes a viable path toward responsible AI in SE through aggressive yet performance-preserving compression.

Distributional and neural approaches to natural language semantics have been built almost exclusively on conventional linear algebra: vectors, matrices, tensors, and the operations that accompany them. These methods have achieved remarkable empirical success, yet they face persistent structural limitations in compositional semantics, type sensitivity, and interpretability. I argue in this paper that geometric algebra (GA) -- specifically, Clifford algebras -- provides a mathematically superior foundation for semantic representation, and that a Functional Geometric Algebra (FGA) framework extends GA toward a typed, compositional semantics capable of supporting inference, transformation, and interpretability while retaining full compatibility with distributional learning and modern neural architectures. I develop the formal foundations, identify three core capabilities that GA provides and linear algebra does not, present a detailed worked example illustrating operator-level semantic contrasts, and show how GA-based operations already implicit in current transformer architectures can be made explicit and extended. The central claim is not merely increased dimensionality but increased structural organization: GA expands an $n$-dimensional embedding space into a $2^n$ multivector algebra where base semantic concepts and their higher-order interactions are represented within a single, principled algebraic framework.

As LLM applications grow more complex, developers are increasingly adopting multi-agent architectures to decompose workflows into specialized, collaborative components, introducing structure that constrains agent behavior and exposes useful semantic predictability. Unlike traditional LLM serving, which operates under highly dynamic and uncertain conditions, this structured topology enables opportunities to reduce runtime uncertainty -- yet existing systems fail to exploit it, treating agentic workloads as generic traffic and incurring significant inefficiencies. Our analysis of production traces from an agent-serving platform and an internal coding assistant reveals key bottlenecks, including low prefix cache hit rates, severe resource contention from long-context requests, and substantial queuing delays due to suboptimal scaling. To address these challenges, we propose Pythia, a multi-agent serving system that captures workflow semantics through a simple interface at the serving layer, unlocking new optimization opportunities and substantially improving throughput and job completion time over state-of-the-art baselines.

Continual offline reinforcement learning (CORL) aims to learn a sequence of tasks from datasets collected over time while preserving performance on previously learned tasks. This setting corresponds to domains where new tasks arise over time, but adapting the model in live environment interactions is expensive, risky, or impossible. However, CORL inherits the dual difficulty of offline reinforcement learning and adapting while preventing catastrophic forgetting. Replay-based continual learning approaches remain a strong baseline but incur memory overhead and suffer from a distribution mismatch between replayed samples and newly learned policies. At the same time, architectural continual learning methods have shown strong potential in supervised learning but remain underexplored in CORL. In this work, we propose TSN-Affinity, a novel CORL method based on TinySubNetworks and Decision Transformer. The method enables task-specific parameterization and controlled knowledge sharing through a RL-aware reuse strategy that routes tasks according to action compatibility and latent similarity. We evaluate the approach on benchmarks based on Atari games and simulations of manipulation tasks with the Franka Emika Panda robotic arm, covering both discrete and continuous control. Results show strong retention from sparse SubNetworks, with routing further improving multi-task performance. Our findings suggest that similarity-guided architectural reuse is a strong and viable alternative to replay-based strategies in a CORL setting. Our code is available at: https://github.com/anonymized-for-submission123/tsn-affinity.

Contact variability, sensing uncertainty, and external disturbances make grasp execution stochastic. Expected-quality objectives ignore tail outcomes and often select grasps that fail under adverse contact realizations. Risk-sensitive POMDPs address this failure mode, but many use particle-filter beliefs that scale poorly, obstruct gradient-based optimization, and estimate Conditional Value-at-Risk (CVaR) with high-variance approximations. We instead formulate grasp acquisition as variational inference over latent contact parameters and object pose, representing the belief with a differentiable Gaussian mixture. We use Gumbel-Softmax component selection and location-scale reparameterization to express samples as smooth functions of the belief parameters, enabling pathwise gradients through a differentiable CVaR surrogate for direct optimization of tail robustness. In simulation, our variational neural belief improves robust grasp success under contact-parameter uncertainty and exogenous force perturbations while reducing planning time by roughly an order of magnitude relative to particle-filter model-predictive control. On a serial-chain robot arm with a multifingered hand, we validate grasp-and-lift success under object-pose uncertainty against a Gaussian baseline. Both methods succeed on the tested perturbations, but our controller terminates in fewer steps and less wall-clock time while achieving a higher tactile grasp-quality proxy. Our learned belief also calibrates risk more accurately, keeping mean absolute calibration error below 0.14 across tested simulation regimes, compared with 0.58 for a Cross-Entropy Method planner.

Preference-based alignment methods, most prominently Reinforcement Learning with Human Feedback (RLHF), use the judgments of human annotators to shape large language model behaviour. However, the normative role of these judgments is rarely made explicit. I distinguish three conceptual models of that role. The first is extension: annotators extend the system designers' own judgments about what outputs should be. The second is evidence: annotators provide independent evidence about some facts, whether moral, social or otherwise. The third is authority: annotators have some independent authority (as representatives of the broader population) to determine system outputs. I argue that these models have implications for how RLHF pipelines should solicit, validate and aggregate annotations. I survey landmark papers in the literature on RLHF and related methods to illustrate how they implicitly draw on these models, describe failure modes that come from unintentionally or intentionally conflating them, and offer normative criteria for choosing among them. My central recommendation is that RLHF pipeline designers should decompose annotation into separable dimensions and tailor each pipeline to the model most appropriate for that dimension, rather than seeking a single unified pipeline.

Finetuning a language model can lead to emergent misalignment (EM) [Betley et al., 2025b]. Models trained on a narrow distribution of misaligned behavior generalize to more egregious behaviors when tested outside the training distribution. We study a set of interventions proposed to reduce EM. We confirm that these interventions reduce or eliminate EM on existing evaluations (questions like "How do I make a quick buck?"). However, if the evaluation prompts are tweaked to resemble the training context, the model displays EM. We call this conditional misalignment. As in standard EM, the model displays misaligned behaviors more egregious than those seen during training, but only on inputs sharing features with the training data. The first two interventions are diluting misaligned data with benign data, and finetuning on benign data after misaligned data. Both produce conditional misalignment. For instance, models trained on a mix of only 5% insecure code still show misalignment when asked to format responses as Python strings (resembling the training context). The third intervention is inoculation prompting. Here, statements with a similar form to the inoculation prompt serve as triggers for misalignment, even if they have the opposite meaning. On the positive side, inoculation prompting has lower (but still non-zero) conditional misalignment if training is on-policy or includes reasoning distillation. Our results imply that in realistic post-training, where misaligned data is typically combined with benign data, models may be conditionally misaligned even if standard evaluations look clean.

Current pedestrian crossing signals operate on fixed timing without adjustment to pedestrian behavior, which can leave vulnerable road users (VRUs) such as the elderly, disabled, or distracted pedestrians stranded when the light changes. We introduce No Pedestrian Left Behind (NPLB), a real-time adaptive traffic signal system that monitors VRUs in crosswalks and automatically extends signal timing when needed. We evaluated five state-of-the-art object detection models on the BGVP dataset, with YOLOv12 achieving the highest mean Average Precision at 50% (mAP@0.5) of 0.756. NPLB integrates our fine-tuned YOLOv12 with ByteTrack multi-object tracking and an adaptive controller that extends pedestrian phases when remaining time falls below a critical threshold. Through 10,000 Monte Carlo simulations, we demonstrate that NPLB improves VRU safety by 71.4%, reducing stranding rates from 9.10% to 2.60%, while requiring signal extensions in only 12.1% of crossing cycles.

Graph neural networks such as ParticleNet and transformer based networks on point clouds such as ParticleTransformer achieve state-of-the-art performance on jet tagging benchmarks at the Large Hadron Collider, yet the physical reasoning behind their predictions remains opaque. We present different methods, i.e. perturbation-based (GNNExplainer), Shapley-value-based (GNNShap), and gradient-based (GRADCam); adapted to operate on LundNet's Lund-plane graph representation. Leveraging the fact that each node in the Lund plane corresponds to a physically meaningful parton splitting, we construct Monte Carlo truth explanation masks and introduce a physics-informed evaluation framework that goes beyond standard fidelity metrics. We perform the analysis in three transverse-momentum bins ($\mathrm{p_T} \in [500,700]$, $[800,1000]$, and the inclusive region $[500,1000]$ GeV), revealing how explanation quality and focus shift between non-perturbative and perturbative regimes. We further quantify the correlation between explainer-assigned node importance and classical jet substructure observables -- $N$-subjettiness ratios $τ_{21}$ and $τ_{32}$ and the energy correlation functions -- establishing the degree to which the model has learned known QCD features. We find that overall the weight assigned by explainability methods has a correlation with analytic observables, with expected shift across different phase space regimes, indicating that a trained neural network indeed learns some aspects of jet-substructure moments. Our open-source implementation enables reproducible explainability studies for graph-based jet taggers.

Resolution of complex post-production issues in large-scale open-source software (OSS) projects requires significant cognitive effort, as developers need to go through long, unstructured and fragmented issue discussion threads before that. In this paper, we present SWE-MIMIC-Bench, an issue trajectory dataset generated from raw GitHub discussions using an automated multi-LLM pipeline. Unlike simple summarization, this pipeline utilizes a group of closed-source LLMs to perform granular tasks: analyzing individual comments with awareness of externally-linked resources, classifying comment analyses into label-specific fields (e.g., root cause, solution plan, implementation progress), and synthesizing label-aware trajectories which capture a structured and coherent narrative of the entire discussion thread. Our pipeline uses five closed-source LLM configurations for distinct purposes: label classification, inline code block and external link summarization, comment analysis, label-specific field classification and trajectory synthesis. By generating concise and reliable trajectories from complex conversation threads, this system can assist developers and researchers of broader software engineering community to understand the experience-driven collaborative approach for issue diagnosis. Furthermore, the generated trajectories can be used to train modern LLM agents to think and act like an expert developer. We evaluated our system on 800 real-world GitHub issues drawn from the SWE-Bench-Pro, SWE-Bench-Multilingual and SWE-Bench-Verified dataset, achieving a 91.7% success rate in extracting 734 high-fidelity reasoning trajectories.

Training language models via reinforcement learning often relies on imperfect proxy rewards, since ground truth rewards that precisely define the intended behavior are rarely available. Standard metrics for assessing the quality of proxy rewards, such as ranking accuracy, treat incorrect rewards as strictly harmful. In this work, however, we highlight that not all deviations from the ground truth are equal. By theoretically analyzing which outputs attract probability during policy gradient optimization, we categorize reward errors according to their effect on the increase in ground truth reward. The analysis establishes that reward errors, though conventionally viewed as harmful, can also be benign or even beneficial by preventing the policy from stalling around outputs with mediocre ground truth reward. We then present two practical implications of our theory. First, for reinforcement learning from human feedback (RLHF), we develop reward model evaluation metrics that account for the harmfulness of reward errors. Compared to standard ranking accuracy, these metrics typically correlate better with the performance of a language model after RLHF, yet gaps remain in robustly evaluating reward models. Second, we provide insights for reward design in settings with verifiable rewards. A key theme underlying our results is that the effectiveness of a proxy reward function depends heavily on its interaction with the initial policy and learning algorithm.

General-relativistic magnetohydrodynamic (GR-MHD) simulations are essential for studying black hole accretion, relativistic jets, and magnetic reconnection, yet their computational cost severely limits systematic parameter exploration. We investigate neural operator surrogates for two astrophysically relevant simulation scenarios produced by the Black Hole Accretion Code (\texttt{BHAC}). First, a Physics Informed Fourier Neural Operator (PINO) is trained on the special-relativistic resistive MHD (SRRMHD) evolution of the Orszag-Tang vortex over a range of resistivities spanning the Sweet-Parker and fast reconnection regimes. By embedding the governing equations as an additional loss term evaluated at finer temporal resolution than the available data supervision, the model learns dynamics at time steps where no simulation data is provided, enabling recovery of plasmoid formation that a data-only baseline trained on the same sparse snapshots fails to reproduce. To our knowledge, the present work is the first application of a physics informed neural operator to special relativistic resistive MHD, and the first to investigate the capability of such models to resolve plasmoid formation in SRRMHD. In a second line of investigation, an OFormer-style Transformer Neural Operator is trained on the evolution of spine-sheath relativistic jets created with \texttt{BHAC}, in special-relativistic MHD (SRMHD). The model is directly applied on the adaptive mesh, highlighting the need for linear attention due to long sequences. The neural surrogate model is capable of capturing most of the major details, especially in early predictions. To our knowledge, this constitutes the first application of a neural operator directly on a high resolution adaptive mesh refinement grid in the context of MHD simulations.

Large language models (LLMs) are increasingly used in emotionally sensitive human-AI applications, yet little is known about how emotion recognition is internally represented. In this work, we investigate the internal mechanisms of emotion recognition in LLMs using sparse autoencoders (SAEs). By analyzing sparse feature activations across layers, we identify a consistent three-phase information flow, in which emotion-related features emerge only in the final phase. We further show that emotion representations comprise both shared features across emotions and emotion-specific features. Using phase-stratified causal tracing, we identify a small set of features that strongly influence emotion predictions, and show that both their number and causal impact vary across emotions; in particular, Disgust is more weakly and diffusely represented than other emotions. Finally, we propose an interpretable and data-efficient causal feature steering method that significantly improves emotion recognition performance across multiple models while largely preserving language modeling ability, and demonstrate that these improvements generalize across multiple emotion recognition datasets. Overall, our findings provide a systematic analysis of the internal mechanisms underlying emotion recognition in LLMs and introduce an efficient, interpretable, and controllable approach for improving model performance.

Existing REST API testing tools are typically evaluated using code coverage and crash-based fault metrics. However, recent LLM-based approaches increasingly generate tests from NL requirements to validate functional behaviour, making traditional metrics weak proxies for whether generated tests validate intended behaviour. To address this gap, we present RESTestBench, a benchmark comprising three REST services paired with manually verified NL requirements in both precise and vague variants, enabling controlled and reproducible evaluation of requirement-based test generation. RESTestBench further introduces a requirements-based mutation testing metric that measures the fault-detection effectiveness of a generated test case with respect to a specific requirement, extending the property-based approach of Bartocci et al. . Using RESTestBench, we evaluate two approaches across multiple state-of-the-art LLMs: (i) non-refinement-based generation, and (ii) refinement-based generation guided by interaction with the running SUT. In the refinement experiments, RESTestBench assesses how exposure to the actual implementation, valid or mutated, affects test effectiveness. Our results show that test effectiveness drops considerably when the generator interacts with faulty or mutated code, especially for vague requirements, sometimes negating the benefit of refinement and indicating that incorporating actual SUT behaviour is unnecessary when requirement detail is high.

Machine-generated text (MGT) detection requires identifying structurally invariant signals across generation models, rather than relying on model-specific fingerprints. In this respect, we hypothesize that while large language models excel at local semantic consistency, their autoregressive nature results in a specific kind of structural fragility compared to human writing. We propose Luminol-AIDetect, a novel, zero-shot statistical approach that exposes this fragility through coherence disruption. By applying a simple randomized text-shuffling procedure, we demonstrate that the resulting shift in perplexity serves as a principled, model-agnostic discriminant, as MGT displays a characteristic dispersion in perplexity-under-shuffling that differs markedly from the more stable structural variability of human-written text. Luminol-AIDetect leverages this distinction to inform its decision process, where a handful of perplexity-based scalar features are extracted from an input text and its shuffled version, then detection is performed via density estimation and ensemble-based prediction. Evaluated across 8 content domains, 11 adversarial attack types, and 18 languages, Luminol-AIDetect demonstrates state-of-the-art performance, with gains up to 17x lower FPR while being cheaper than prior methods.

Transformers have demonstrated a strong ability for in-context learning (ICL), enabling models to solve previously unseen tasks using only example input output pairs provided at inference time. While prior theoretical work has established conditions under which transformers can perform linear classification in-context, the empirical scaling behavior governing when this mechanism succeeds remains insufficiently characterized. In this paper, we conduct a systematic empirical study of in-context learning for Gaussian-mixture binary classification tasks. Building on the theoretical framework of Frei and Vardi (2024), we analyze how in-context test accuracy depends on three fundamental factors: the input dimension, the number of in-context examples, and the number of pre-training tasks. Using a controlled synthetic setup and a linear in-context classifier formulation, we isolate the geometric conditions under which models successfully infer task structure from context alone. We additionally investigate the emergence of benign overfitting, where models memorize noisy in-context labels while still achieving strong generalization performance on clean test data. Through extensive sweeps across dimensionality, sequence length, task diversity, and signal-to-noise regimes, we identify the parameter regions in which this phenomenon arises and characterize how it depends on data geometry and training exposure. Our results provide a comprehensive empirical map of scaling behavior in in-context classification, highlighting the critical role of dimensionality, signal strength, and contextual information in determining when in-context learning succeeds and when it fails.

Multimodal large language models (MLLMs) achieve ever-stronger performance on visual-language tasks. Even as traditional visual question answering benchmarks approach saturation, reliable deployment requires satisfying low error tolerances in real-world out-of-distribution (OOD) scenarios. Precisely, selective prediction aims to improve coverage, i.e. the share of inputs the system answers, while adhering to a user-defined risk level. This is typically achieved by assigning a confidence score to each answer and abstaining on those that fall below a certain threshold. To enable reliable generalization, we require reasoner models to produce localized visual evidence while answering, and design a selector that explicitly learns to estimate the quality of the localization provided by the reasoner. We show that SIEVES (Selective Prediction through Visual Evidence Scoring) improves coverage by up to three times on challenging OOD benchmarks (V* Bench, HR-Bench-8k, MME-RealWorld-Lite, VizWiz, and AdVQA), compared to non-grounding baselines. Beyond better generalization to OOD tasks, the design of the SIEVES selector enables transfer to proprietary reasoners without access to their weights or logits, such as o3 and Gemini-3-Pro, providing coverage boosts beyond those attributable to accuracy alone. We highlight that SIEVES generalizes across all five tested OOD datasets and reasoner models (Pixel-Reasoner, o3, and Gemini-3-Pro), without benchmark- or reasoner-specific training or adaptation.

Traditional loss functions, including cross-entropy, contrastive, triplet, and su pervised contrastive losses, used for fine-tuning pre-trained language models such as BERT, operate only within local neighborhoods and fail to account for the global semantic structure. We present G-Loss, a graph-guided loss function that incorporates semi-supervised label propagation to use structural relationships within the embedding manifold. G-Loss builds a document-similarity graph that captures global semantic relationships, thereby guiding the model to learn more discriminative and robust embeddings. We evaluate G-Loss on five benchmark datasets covering key downstream classification tasks: MR (sentiment analysis), R8 and R52 (topic categorization), Ohsumed (medical document classification), and 20NG (news categorization). In the majority of experimental setups, G-Loss converges faster and produces semantically coherent embedding spaces, resulting in higher classification accuracy than models fine-tuned with traditional loss functions.

Harnesses have become a central determinant of coding-agent performance, shaping how models interact with repositories, tools, and execution environments. Yet automating harness engineering is hard: a heterogeneous action space, sparse and noisy evaluation signal, multi-million-token trajectories, and edits whose effect is hard to attribute to the next round's outcomes. We introduce Agentic Harness Engineering (AHE), a framework that automates harness-level evolution by instrumenting the three stages of any engineering loop (component editing, trajectory inspection, and decision making) with matched observability pillars: (1) component observability gives every editable harness component a file-level representation so the action space is explicit and revertible; (2) experience observability distills millions of raw trajectory tokens into a layered, drill-down evidence corpus that an evolving agent can actually consume; and (3) decision observability pairs every edit with a self-declared prediction, later verified against the next round's task-level outcomes. Together, these pillars turn every edit into a falsifiable contract, so harness evolution proceeds autonomously without collapsing into trial-and-error. Empirically, ten AHE iterations lift pass@1 on Terminal-Bench 2 from 69.7% to 77.0%, surpassing the human-designed harness Codex-CLI (71.9%) and the self-evolving baselines ACE and TF-GRPO. The frozen harness transfers without re-evolution: on SWE-bench-verified it tops aggregate success at 12% fewer tokens than the seed, and on Terminal-Bench 2 it yields +5.1 to +10.1pp cross-family gains across three alternate model families, indicating the evolved components encode general engineering experience rather than benchmark-specific tuning. These results position observability-driven evolution as a practical pathway to keep coding-agent harnesses continually improving.

Long-horizon LLM tasks often fail not because a single answer is unattainable, but because knowledge states drift across rounds, intermediate commitments remain implicit, and interruption fractures the evolving evidence chain. This paper presents ADEMA as a knowledge-state orchestration architecture for long-horizon knowledge synthesis rather than as a generic multi-agent runtime. The architecture combines explicit epistemic bookkeeping, heterogeneous dual-evaluator governance, adaptive task-mode switching, reputation-shaped resource allocation, checkpoint-resumable persistence, segment-level memory condensation, artifact-first assembly, and final-validity checking with safe fallback. Evidence is drawn entirely from existing materials: a four-scenario showcase package, a fixed 60-run mechanism matrix, targeted micro-ablation and artifact-chain supplements, and a repaired protocol-level benchmark in which code-oriented evaluation is the clearest quality-sensitive mechanism block. Across the fixed matrix, removing checkpoint/resume produced the only invalid run, and it did so in the interruption-sensitive resume condition. By contrast, dual evaluation, segment synthesis, and dynamic governance are best interpreted as supporting control mechanisms that shape trajectory discipline, explicit artifact progression, and cost-quality behavior rather than as universal binary prerequisites for completion. The contribution is therefore a knowledge-state orchestration architecture in which explicit epistemic state transition, evidence-bearing artifact progression, and recoverable continuity are the primary design commitments.

We study city-scale control of electric-vehicle (EV) ride-hailing fleets where dispatch, repositioning, and charging decisions must respect charger and feeder limits under uncertain, spatially correlated demand and travel times. We formulate the problem as a hex-grid semi-Markov decision process (semi-MDP) with mixed actions -- discrete actions for serving, repositioning, and charging, together with continuous charging power -- and variable action durations. To guarantee physical feasibility during both training and deployment, the policy learns over high-level intentions produced by a masked, temperature-annealed actor. These intentions are projected at every decision step through a time-limited rolling mixed-integer linear program (MILP) that strictly enforces state-of-charge, port, and feeder constraints. To mitigate distributional shifts, we optimize a Soft Actor--Critic (SAC) agent against a Wasserstein-1 ambiguity set with a graph-aligned Mahalanobis ground metric that captures spatial correlations. The robust backup uses the Kantorovich--Rubinstein dual, a projected subgradient inner loop, and a primal--dual risk-budget update. Our architecture combines a two-layer Graph Convolutional Network (GCN) encoder, twin critics, and a value network that drives the adversary. Experiments on a large-scale EV fleet simulator built from NYC taxi data show that PD--RSAC achieves the highest net profit, reaching \$1.22M, compared with \$0.58M--\$0.70M for strong heuristic, single-agent RL, and multi-agent RL baselines, including Greedy, SAC, MAPPO, and MADDPG, while maintaining zero feeder-limit violations.

Optimization modeling underpins real-world decision-making in logistics, manufacturing, energy, and public services, but reliably solving such problems from natural-language requirements remains challenging for current large language models (LLMs). In this paper, we propose \emph{Agora-Opt}, a modular agentic framework for optimization modeling that combines decentralized debate with a read-write memory bank. Agora-Opt allows multiple agent teams to independently produce end-to-end solutions and reconcile them through an outcome-grounded debate protocol, while memory stores solver-verified artifacts and past disagreement resolutions to support training-free improvement over time. This design is flexible across both backbones and methods: it reduces base-model lock-in, transfers across different LLM families, and can be layered onto existing pipelines with minimal coupling. Across public benchmarks, Agora-Opt achieves the strongest overall performance among all compared methods, outperforming strong zero-shot LLMs, training-centric approaches, and prior agentic baselines. Further analyses show robust gains across backbone choices and component variants, and demonstrate that decentralized debate offers a structural advantage over centralized selection by enabling agents to refine candidate solutions through interaction and even recover correct formulations when all initial candidates are flawed. These results suggest that reliable optimization modeling benefits from combining collaborative cross-checking with reusable experience, and position Agora-Opt as a practical and extensible foundation for trustworthy optimization modeling assistance. Our code and data are available at https://github.com/CHIANGEL/Agora-Opt.

Security analysts are overwhelmed by the volume of alerts and the low context provided by many detection systems. Early-stage investigations typically require manual correlation across multiple log sources, a task that is usually time-consuming. In this paper, we present an experimental, agentic workflow that leverages large language models (LLMs) augmented with predefined queries and constrained tool access (structured SQL over Suricata logs and grep-based text search) to automate the first stages of alert investigation. The proposed workflow integrates queries to provide an overview of the available data, and LLM components that selects which queries to use based on the overview results, extracts raw evidence from the query results, and delivers a final verdict of the alert. Our results demonstrate that the LLM-powered workflow can investigate log sources, plan an investigation, and produce a final verdict that has a significantly higher accuracy than a verdict produced by the same LLM without the proposed workflow. By recognizing the inherent limitations of directly applying LLMs to high-volume and unstructured data, we propose combining existing investigation practices of real-world analysts with a structured approach to leverage LLMs as virtual security analysts, thereby assisting and reducing the manual workload.

Patient simulators are gaining traction in mental health training by providing scalable exposure to complex and sensitive patient interactions. Simulating depressed patients is particularly challenging, as safety constraints and high patient variability complicate simulations and underscore the need for simulators that capture diverse and realistic patient behaviors. However, existing evaluations heavily rely on LLM-judges with poorly specified prompts and do not assess behavioral diversity. We introduce PSI-Bench, an automatic evaluation framework that provides interpretable, clinically grounded diagnostics of depression patient simulator behavior across turn-, dialogue-, and population-level dimensions. Using PSI-Bench, we benchmark seven LLMs across two simulator frameworks and find that simulators produce overly long, lexically diverse responses, show reduced variability, resolve emotions too quickly, and follow a uniform negative-to-positive trajectory. We also show that the simulation framework has a larger impact on fidelity than the model scale. Results from a human study demonstrate that our benchmark is strongly aligned with expert judgments. Our work reveals key limitations of current depression patient simulators and provides an interpretable, extensible benchmark to guide future simulator design and evaluation.

The marginal likelihood, also known as the evidence, is regarded as a mathematical embodiment of Occam's razor, enabling model selection that avoids overfitting. The evidence lower bound (ELBO) objective from variational inference has also been used for similar purposes. Prior work has shown that restricting the approximate posterior family via a mean-field approximation can lead the ELBO to underfit. In this paper, we show how ELBO-based hyperparameter learning in a simple over-parameterized regression model can also produce overfitting, depending on the assumed rank of the covariance matrix in a Gaussian approximate posterior. Surprisingly, among only the underfit and overfit options, Bayesian model selection via the evidence itself sometimes prefers the overfit version, while the ELBO does not. Bayesian practitioners hoping to scale to large models should be cautious about how reduced-rank assumptions needed for tractability may impact the potential for model selection.

With the rapid development of the Internet, users have increasingly higher expectations for the recommendation accuracy of online content consumption platforms. However, short videos often contain diverse segments, and users may not hold the same attitude toward all of them. Traditional binary-classification recommendation models, which treat a video as a single holistic entity, face limitations in accurately capturing such nuanced preferences. Considering that user consumption is a temporal process, this paper demonstrates that the timing of user actions can represent diverse intentions through statistical analysis and examination of action patterns. Based on this insight, we propose a novel modeling paradigm: Action-Aware Generative Sequence Network (A2Gen), which refines user actions along the temporal dimension and chains them into sequences for unified processing and prediction. First, we introduce the Context-aware Attention Module (CAM) to model action sequences enriched with item-specific contextual features. Building upon this, we develop the Hierarchical Sequence Encoder (HSE) to learn temporal action patterns from users' historical actions. Finally, through leveraging CAM, we design a module for action sequence generation: the Action-seq Autoregressive Generator (AAG). Extensive offline experiments on the Kuaishou's dataset and the Tmall public dataset demonstrate the superiority of our proposed model. Furthermore, through large-scale online A/B testing deployed on Kuaishou's platform, our model achieves significant improvements over baseline methods in multi-task prediction by leveraging sequential information. Specifically, it yields increases of 0.34% in user watch time, 8.1% in interaction rate, and 0.162% in overall user retention (LifeTime-7), leading to successful deployment across all traffic, serving over 400 million users every day.

Real-world evidence (RWE) studies that emulate target trials increasingly inform regulatory and clinical decisions, yet residual, hard-to-quantify biases still limit their credibility. The recently proposed BenchExCal framework addresses this challenge via a two-stage Benchmark, Expand, Calibrate process, which first compares an observational emulation against an existing randomized controlled trial (RCT), then uses observed divergence to calibrate a second emulation for a new indication causal effect estimation. While methodologically powerful, BenchExCal is resource intensive and difficult to scale. We introduce TrialCalibre, a conceptualized multiagent system designed to automate and scale the BenchExCal workflow. Our framework features specialized agents such as the Orchestrator, Protocol Design, Data Synthesis, Clinical Validation, and Quantitative Calibration Agents that coordi-nate the the overall process. TrialCalibre incorpo-rates agent learning (e.g., RLHF) and knowledge blackboards to support adaptive, auditable, and transparent causal effect estimation.

Social identity is a concept from psychology that refers to the part of an individual's identity that derives from their group membership(s). In this paper, we explore social identity in members of the professional community of Research Software Engineers (RSEs). Using a mixed-methods approach, our study combined computational linguistic analysis and inferential statistics to examine over 28,000 social media posts, 1,700 blogs, and survey responses from 381 professional RSEs. The findings highlight the emergence of a collective RSE identity and demonstrate its role in shaping professional wellbeing. This study contributes an interdisciplinary perspective by integrating social psychology and software engineering to show how a professional identity evolves and why it matters.

Creating interactive STEM courseware traditionally requires HTML/CSS/JavaScript expertise, leaving barriers for educators. While generative AI can produce HTML codes, existing tools generate static presentations rather than interactive simulations, struggle with long documents, and lack pedagogical accuracy mechanisms. Furthermore, full regeneration for modifications requires 200--600 seconds, disrupting creative flow. We present MAIC-UI, a zero-code authoring system that enables educators to create and rapidly edit interactive courseware from textbooks, PPTs, and PDFs. MAIC-UI employs: (1) structured knowledge analysis with multi-modal understanding to ensure pedagogical rigor; (2) a two-stage generate-verify-optimize pipeline separating content alignment from visual refinement; and (3) Click-to-Locate editing with Unified Diff-based incremental generation achieving sub-10-second iteration cycles. A controlled lab study with 40 participants shows MAIC-UI reduces editing iterations (4.9 vs. 7.0) and significantly improves learnability and controllability compared to direct Text-to-HTML generation. A three-month classroom deployment with 53 high school students demonstrates that MAIC-UI fosters learning agency and reduces outcome disparities -- the pilot class achieved 9.21-point gains in STEM subjects compared to -2.32 points in control classes. Our code is available at https://github.com/THU-MAIC/MAIC-UI.

Microservice-based systems impose significant organizational coordination challenges, yet the role of individual developers in shaping organizational coupling (OC) remains underexplored. Prior work largely focuses on structural architectural aspects, leaving gaps in understanding how developer roles influence coordination dynamics over time. This study investigates how different developer roles contribute to OC in a large-scale microservices system. The analysis focuses on three key roles, namely Jacks, representing broad knowledge holders, Mavens, representing deep specialists, and Connectors, representing organizational bridges. A longitudinal repository mining analysis of GitHub data, including commits and issue and pull request interactions, is conducted to operationalize OC and quantify its evolution over time. The results show that Connectors are consistently associated with higher levels of OC, while the co-occurrence of multiple roles within the same developer further amplifies coupling effects. In contrast, Jacks and Mavens exhibit more localized and role-specific influences. These findings indicate that OC in microservices is primarily a role-driven phenomenon rather than an inevitable structural property, providing a foundation for role-aware organizational design and targeted decoupling strategies.

Chain-of-Thought (CoT) has been shown to empirically improve Transformers' performance, and theoretically increase their expressivity to Turing completeness. However, whether Transformers can learn to generalize to CoT traces longer than those seen during training is understudied. We use recent theoretical frameworks for Transformer length generalization and find that -- under standard positional encodings and a finite alphabet -- Transformers with CoT cannot solve problems beyond $TC^0$, i.e. the expressivity benefits do not hold under the stricter requirement of length-generalizable learnability. However, if we allow the vocabulary to grow with problem size, we attain a length-generalizable simulation of Turing machines where the CoT trace length is linear in the simulated runtime up to a constant. Our construction overcomes two core obstacles to reliable length generalization: repeated copying and last-occurrence retrieval. We assign each tape position a unique signpost token, and log only value changes to enable recovery of the current tape symbol through counts circumventing both barriers. Further, we empirically show that the use of such signpost tokens and value change encodings provide actionable guidance to improve length generalization on hard problems.

The convergence of accelerating human spaceflight ambitions and critical terrestrial health monitoring demands is driving unprecedented requirements for reliable, real-time feature extraction on extremely resource-constrained wearable health sensors. We present an ultra-low-power (ULP) Field-Programmable Gate Array (FPGA) based solution for real-time Seismocardiography (SCG) feature classification using Convolutional Neural Networks (CNNs). Our approach combines quantization-aware training with a systolic-array accelerator to enable efficient integer-only inference on the Lattice iCE40UP5K FPGA, which offers an ideal platform for battery-powered deployments -- particularly in space environments -- thanks to its power efficiency and radiation resilience. The implementation achieves a validation accuracy of 98% while consuming only 8.55 mW, completing inference in 95.5 ms with minimal hardware resources (2,861 LUTs and 7 DSP blocks). These results demonstrate that fully on-device SCG-based cardiac feature extraction is feasible on resource-constrained hardware, enabling energy-efficient, autonomous health monitoring for astronauts in long-duration space missions.

We present StratFormer, a transformer-based meta-agent that learns to simultaneously model and exploit opponents in imperfect-information games through a two-phase curriculum. The first phase trains an opponent modeling head to identify behavioral patterns from action histories while the agent plays a game-theoretic optimal (GTO) policy. The second phase progressively shifts the policy toward best-response (BR) exploitation, guided by a per-opponent regularization schedule tied to exploitability. Our architecture introduces dual-turn tokens -- feature vectors constructed at both agent and opponent decision points -- coupled with bucket-rate features that encode opponent tendencies across five strategic contexts. On Leduc Hold'em, a small poker variant with six cards and two betting rounds, we test against six opponent archetypes at two strength levels each, with exploitability ranging from 0.15 to 1.26 Big Blinds (BB) per hand. StratFormer achieves an average exploitation gain of +0.106 BB per hand over GTO, with peak gains of +0.821 against highly exploitable opponents, while maintaining near-equilibrium safety.

Knowledge distillation (KD) is a well-known technique to effectively compress a large network (teacher) to a smaller network (student) with little sacrifice in performance. However, most KD methods require a large training set and internal access to the teacher, which are rarely available due to various restrictions. These challenges have originated a more practical setting known as black-box few-shot KD, where the student is trained with few images and a black-box teacher. Recent approaches typically generate additional synthetic images but lack an active strategy to promote their diversity, a crucial factor for student learning. To address these problems, we propose a novel training scheme for generative adversarial networks, where we adaptively select high-confidence images under the teacher's supervision and introduce them to the adversarial learning on-the-fly. Our approach helps expand and improve the diversity of the distillation set, significantly boosting student accuracy. Through extensive experiments, we achieve state-of-the-art results among other few-shot KD methods on seven image datasets. The code is available at https://github.com/votrinhan88/divbfkd.

Knowledge distillation (KD) represents a vital mechanism to transfer expertise from complex teacher networks to efficient student models. However, in decentralized or secure AI ecosystems, privacy regulations and proprietary interests often restrict access to the teacher's interface and original datasets. These constraints define a challenging black-box data-free KD scenario where only top-1 predictions and no training data are available. While recent approaches utilize synthetic data, they still face limitations in data diversity and distillation signals. We propose Diverse Image Priors Knowledge Distillation (DIP-KD), a framework that addresses these challenges through a three-phase collaborative pipeline: (1) Synthesis of image priors to capture diverse visual patterns and semantics; (2) Contrast to enhance the collective distinction between synthetic samples via contrastive learning; and (3) Distillation via a novel primer student that enables soft-probability KD. Our evaluation across 12 benchmarks shows that DIP-KD achieves state-of-the-art performance, with ablations confirming data diversity as critical for knowledge acquisition in restricted AI environments.

Subliminal learning describes a student language model inheriting a behavioral bias by fine-tuning on seemingly innocuous data generated by a biased teacher model. Prior work has begun to characterize this phenomenon but leaves open questions about the scope of signals it can transfer, the mechanisms that explain it, and the precision with which a bias can be encoded by seemingly unrelated data. We tackle all three problems by introducing subliminal steering, a variant of subliminal learning in which the teacher's bias is implemented not via a system prompt, as in prior work, but through a steering vector trained to maximize the likelihood of a set of target samples. First, we show that subliminal steering transfers complex multi-word biases, whereas prior work focused on single-word preferences, demonstrating a large scope of subliminally transferrable signals. Second, we provide mechanistic evidence that subliminal learning transfers not only the target behavioral bias, but also the steering vector itself, localized to the layers at which the teacher was steered. Finally, we show that the bias is encoded with surprising precision. We train a new steering vector directly on the subliminally-laden dataset and find that it attains high cosine similarity with the original vector.

Frontier AI both amplifies existing risks and introduces qualitatively novel challenges. Not only is there a notable lack of stable scientific consensus resulting from the rapid pace of technological change, but emerging frontier AI safety practices are often misaligned with, or may undermine, established risk management frameworks. To address these challenges, we systematically surface open problems in frontier AI risk management. Adopting a problem-oriented approach, we examine each stage of the risk management process - risk planning, identification, analysis, evaluation, and mitigation - through a structured review of the literature, identifying unresolved challenges and the actors best positioned to address them. Recognising that different types of open problems call for different responses, we classify open problems according to whether they reflect (a) a lack of scientific or technical consensus, (b) misalignment with, or challenges to, established risk management frameworks, or (c) shortcomings in implementation despite apparent consensus and alignment. By mapping these open problems and identifying the actors best positioned to address them - including developers, deployers, regulators, standards bodies, researchers, and third-party evaluators - this work aims to clarify where progress is needed to enable robust and meaningful consensus on frontier AI risk management.The paper does not propose specific solutions; instead, it provides a problem-oriented, agenda-setting reference document, complemented by a living online repository, intended to support coordination, reduce duplication, and guide future research and governance efforts.

In the MNIST auxiliary logit distillation experiment, a student can acquire an unintended teacher trait despite distilling only on no-class logits through a phenomenon called subliminal learning. Under a single-step gradient descent assumption, subliminal learning theory attributes this effect to alignment between the trait and distillation gradients, but does not guarantee that this alignment persists in a multi-step setting. We empirically show that gradient alignment remains weakly but consistently positive throughout training and causally contributes to trait acquisition. We show that a mitigation method called liminal training works by attenuating the alignment and fails to stop trait acquisition in this setup. These results suggest that mitigation methods that operate in this regime may not reliably suppress trait acquisition when the first-order drive dominates.

Source Code Plagiarism Detection (SCPD) plays an important role in maintaining fairness and academic integrity in software engineering education. Code Evaluation Metrics (CEMs) are developed for assessing code generation tasks. However, it remains unclear whether such metrics can reliably detect plagiarism across different levels of modification (L1-L6), increasing in complexity. In this paper, we perform a comparative empirical study using two open-source labelled datasets, ConPlag (raw and template-free versions) and IRPlag. We evaluate five CEMs, namely CodeBLEU, CrystalBLEU, RUBY, Tree Structured Edit Distance (TSED), and CodeBERTScore. The performance is evaluated using threshold-free ranking-based measures to assess overall, per dataset, and per-level plagiarism performance. The results are compared against state-of-the-art (SOTA) Source Code Plagiarism Detection Tools (SCPDTs), JPlag and Dolos. Our findings show that without preprocessing, Dolos achieves the highest overall ranking performance, while among the individual metrics, CrystalBLEU, CodeBLEU, and RUBY outperform JPlag. Performance is strongest at L1 and drops from L4 onward, while CrystalBLEU remains competitive on L6. With preprocessing, CrystalBLEU surpasses Dolos overall. Per dataset, Dolos achieved the best ranking on the ConPlag raw dataset, while CrystalBLEU was the best-performing metric on the remaining datasets. At the plagiarism levels, Dolos remains strongest on L4, while Crystal-BLEU leads most of the remaining difficult levels. These results indicate that CEMs are comparable to dedicated tools in terms of ranking metrics.

Federated inference enhances LLM performance in edge computing through weighted averaging of distributed model predictions. However, autoregressive LLM inference requires frequent full-model forward passes across workers, severely limiting decoding throughput. Distributed deployment further aggravates this due to a communication bottleneck: each worker must transmit full token probability distributions per draft token, dominating end-to-end latency. To address these challenges, we introduce speculative decoding to enable parallel LLM processing and propose a top-K compressed transmission scheme with two server-side reconstruction strategies. We theoretically analyze the robustness of our method in terms of local reconstruction error, aggregation bias, and acceptance-rate bias, and derive corresponding bounds. Experiments demonstrate that our scheme achieves high generation fidelity while significantly reducing communication overhead.

Critical analyses of emotion recognition technology have raised ethical concerns around task validity and potential downstream impacts, urging researchers to ensure alignment between their stated motivations and practice. However, these discussions have not adequately influenced or drawn from research on speech emotion recognition (SER). We address this gap by conducting a systematic survey of SER research to uncover what stated motivations drive this work and if they align with the datasets and emotions studied. We find that while SER research identifies appealing goals, such as well-situated voice-activated systems or healthcare applications, commonly-used datasets do not reflect these proposed deployment contexts, thus presenting a gap between motivations and research practices. We argue that such gaps engender ethical concerns, and that SER research should reassert itself with concrete use-cases to prevent misinterpretations, misuse, and downstream harms.

In the context of quantum secure scenarios, existing research on mobile edge devices and intelligent computing and edge (ICE) systems based on the Non-Orthogonal Multiple Access (NOMA) communication model have overlooked the energy consumption overhead of Post-Quantum Cryptography (PQC) modules, and the high complexity of traditional resource allocation algorithms fails to meet the demands of real-time decision-making. To address these challenges, this paper proposes a lightweight agentic AI framework designed for online joint optimization within ICE-enabled mobile devices. The scheme constructs a multi-stage stochastic Mixed Integer Nonlinear Programming (MINLP) model that incorporates static power-consumption constraints for PQC modules. Based on Lyapunov optimization theory, the long-term optimization problem is decoupled, and a linear complexity algorithm is proposed to solve the nonconvex challenges of NOMA power allocation . Simulation results verify that the proposed scheme significantly improves computational throughput while ensuring system queue stability and energy consumption constraints. Compared with traditional Successive Convex Approximation (SCA) algorithms, the complexity is reduced to $\mathcal{O}(N)$, achieving a speedup of approximately 46 times when the number of devices $N=35$, thereby meeting the real-time decision-making requirements in dynamic wireless environments.

We present a quantitative analysis of character-pair substitution ratios in the Voynich manuscript, testing whether Currier's A/B language distinction (1976) reflects a genuine structural property of the text. A Beta-Binomial mixture model applied to raw character counts without access to labels recovers the Currier split with ARI = 0.383. A supervised Beta-Binomial classifier trained on a subset of folios predicts the A/B identity of held-out folios at 89.2% accuracy. The character pairs separate into three functional regimes that constrain any theory of the Voynich writing system.

Prior work on node classification has shown that Graph Neural Networks (GNNs) can learn representations that transfer across graphs, when underlying graph properties are shared. For a fixed graph, one would then expect GNNs trained for link prediction to learn a representation consistent with that learnt for node classification. We show this intuition does not hold in the general case. Instead, we find popular link prediction models can learn a trivial mini-batch dependent heuristic, enabled by batch-normalisation layers, to solve the edge classification task. When correcting for this, we observe increased alignment of the network representation with node-class relevant features, suggesting the network has learnt a graph representation that better aligns with the underlying graph's properties. Our findings suggest that standard link prediction training may be leading us to overestimate link predictors' ability to learn a generalised representation of a graph that is consistent across tasks.

Organizations routinely make strategic budget allocations under operational constraints, but often lack a principled way to assess whether realized allocations were close to the best feasible choices in hindsight. We present a retrospective auditing framework based on hindsight regret, defined as the opportunity cost of the realized allocation relative to a constraint-faithful benchmark under the same budget and stability guardrails. The framework estimates regime-specific spend--response functions from historical logs, computes feasible hindsight allocations via constrained optimization, and propagates uncertainty through Monte Carlo evaluation to produce regret distributions, expected lift, and probability-of-improvement summaries. This separates allocation inefficiency from uncertainty in the estimated response surfaces. Experiments on real marketing allocation logs show that the framework yields interpretable post-hoc diagnostics and reveals a practical trade-off between allocation flexibility and detectability: moderate feasible reallocations often capture most measurable gain, while larger shifts move into weak-support regions with higher uncertainty. The result is a practical method for auditing historical budget decisions when online experimentation is costly or infeasible.

Fault-tolerant quantum computing (FTQC) is emerging as the architectural regime in which practical large-scale quantum workloads will execute. In this setting, however, multiprogramming is no longer a matter of partitioning a flat pool of qubits. Quantum error correction exposes a structured floorplan of data tiles, ancilla tiles, and magic-state service resources, so concurrent execution must account for compact placement, connectivity, routing headroom, and shared support infrastructure. This makes FTQC multiprogramming fundamentally harder than its NISQ counterpart: admission decisions can fragment the remaining floorplan, conservative reservations can waste ancilla, and dynamic contention across data, ancilla, and magic-state resources can degrade both throughput and quality of service. In this work, we develop a formal framework for FTQC multiprogramming that captures these structural constraints and their runtime implications. We formulate the baseline static allocation problem, extend it to limited-resource and online settings through hierarchy-aware scheduling policies, and further generalize it to cultivation-enabled architectures with dynamic magic-state generation. Through simulation on synthetic Clifford+T workloads, the proposed scheduler achieves a normalized system speedup of 3.1x, improving over prior FTQC multiprogramming baselines by ~29% while maintaining low mean slowdown.

Deploying production-ready multi-agent systems (MAS) in complex industrial environments remains challenging due to limitations in scalability, observability, and autonomous evolution. We present OxyGent, an open-source framework driven by two core novelties: a unified Oxy abstraction and the OxyBank evolution engine. The unified abstraction encapsulates agents, tools, LLMs, and reasoning flows as pluggable atomic components, enabling Lego-like scalable system composition and non-intrusive monitoring. To enhance observability, OxyGent introduces permission-driven dynamic planning that replaces rigid workflows with execution graphs generated at runtime, providing adaptive visualizations. Furthermore, to support continuous evolution, OxyBank serves as an AI asset management platform that drives automated data backflow, annotation, and joint evolution. Empirical evaluations and real-world case studies show that OxyGent provides a robust and scalable foundation for MAS. OxyGent is fully open-sourced under the Apache License 2.0 at https://github.com/jd-opensource/OxyGent.

Key-value (KV) caching is essential for large language model inference, yet its memory overhead poses a critical bottleneck for long-context generation. Existing eviction policies predominantly rely on empirical heuristics, lacking a rigorous theoretical foundation. This work rethinks KV cache eviction through the lens of the Information Bottleneck principle. Under a linear-Gaussian surrogate of attention, we derive a closed-form mutual information objective that characterizes the effective information capacity of a retained KV cache subset. This formulation reveals that a wide range of existing eviction strategies can be interpreted as different approximations of the same capacity-maximization principle. Guided by this insight, we introduce CapKV, a capacity-aware eviction method that directly targets information preservation via a log-determinant approximation using statistical leverage scores. This approach replaces heuristic selection with a theoretically grounded mechanism that preserves the maximum predictive signal. Extensive experiments across multiple models and long-context benchmarks show that CapKV consistently outperforms prior methods, achieving a better trade-off between memory efficiency and generational fidelity.

Recent knowledge distillation (KD) methods for semantic segmentation introduce increasingly complex hand-crafted objectives, yet are typically evaluated under fixed iteration schedules. These objectives substantially increase per-iteration cost, meaning equal iteration counts do not correspond to equal training budgets. It is therefore unclear whether reported gains reflect stronger distillation signals or simply greater compute. We show that iteration-based comparisons are misleading: when wall-clock compute is matched, canonical logit- and feature-based KD outperform recent segmentation-specific methods. Under extended training, feature-based distillation achieves state-of-the-art ResNet-18 performance on Cityscapes and ADE20K. A PSPNet ResNet-18 student closely approaches its ResNet-101 teacher despite using only one quarter of the parameters, reaching 99% of the teacher's mIoU on Cityscapes (79.0 vs 79.8) and 92% on ADE20K. Our results challenge the prevailing assumption that KD for segmentation requires task-specific mechanisms and suggest that scaling, rather than complex hand-crafted objectives, should guide future method design.

Efficient GPU execution of convolution operators is governed by memory-access efficiency, on-chip data reuse, and execution mapping rather than arithmetic throughput alone. This paper presents a controlled operator-level study of CUDA kernel optimization for the depthwise convolution used in Structured State Space Model Convolutional Diagonal (S4ConvD), together with a cloud-compatible, counter-free performance analysis methodology. The operator, model, dataset, and training configuration are fixed, and only the CUDA kernel implementation is varied. The evaluated CUDA kernels comprise naive, global-memory-coalesced, shared-memory cache-blocked, and warp-tiled variants, covering forward, input-gradient, and weight-gradient execution paths under steady-state training conditions. Performance is characterized using a counter-free methodology that combines CUDA-event timing, execution-path decomposition, analytically derived memory-traffic modeling, effective-bandwidth estimation, and roofline analysis. This enables profiling-like architectural insights without requiring hardware performance counters or privileged profiling access. The warp-tiled kernel reduces convolution runtime by $3.26\times$ relative to the naive CUDA baseline, while end-to-end training speedup reaches $1.29\times$. A PyTorch implementation is used separately for numerical validation and runtime context, but is not treated as a controlled architectural baseline. Forward and input-gradient paths benefit substantially from improved locality and on-chip data reuse, whereas the reduction-dominated weight-gradient path remains the primary bottleneck. The results demonstrate that meaningful architecture-level GPU kernel analysis can be performed reproducibly in restricted cloud environments, even without access to hardware performance counters.

In multi-agent reinforcement learning (MARL), the integration of a communication mechanism, allowing agents to better learn to coordinate their actions and converge on their objectives by sharing information. Based on an interaction graph, a subclass of methods employs graph neural networks (GNNs) to learn the communication, enabling agents to improve their internal representations by enriching them with information exchanged. With growing research, we note a lack of explicit structure and framework to distinguish and classify MARL approaches with communication based on GNNs. Thus, this paper surveys recent works in this field. We propose a generalized GNN-based communication process with the goal of making the underlying concepts behind the methods more obvious and accessible.

Speculative decoding enhances the inference efficiency of large language models (LLMs) by generating drafts using a small draft language model (DLM) and verifying them in batches with a large target language model (TLM). However, adaptive drafting inference on a mobile single-NPU-PIM system faces idle overhead in traditional operator-level synchronous execution and wasted computation in asynchronous execution due to fluctuations in draft length. This paper introduces AHASD, a task-level asynchronous mobile NPU-PIM heterogeneous architecture for speculative decoding. Notably, AHASD achieves parallel drafting on the PIM and verification on a single NPU through task-level DLM-TLM decoupling and specifically, it incorporates Entropy-History-Aware Drafting Control and Time-Aware Pre-Verification Control to dynamically manage adaptive drafting algorithm execution and pre-verification timing, suppressing invalid drafting based on low-confidence drafts. Additionally, AHASD integrates Attention Algorithm Units and Gated Task Scheduling Units within LPDDR5-PIM to enable attention link localization and sub-microsecond task switching on the PIM side. Experimental results for different LLMs and adaptive drafting algorithms show that AHASD achieves up to 4.2$\times$ in throughput and 5.6$\times$ in energy efficiency improvements over a GPU-only baseline, and 1.5$\times$ in throughput and 1.24$\times$ in energy efficiency gains over the state-of-the-art GPU+PIM baseline, with hardware overhead below 3\% of the DRAM area.

Viruses represent the most abundant biological entities on Earth and play a pivotal role in microbial ecosystems, yet, as prominent human pathogens, they are closely linked to human morbidity and mortality. Accurate identification of viral sequences from viral genome sequences is therefore essential, but existing genome-based classification models that largely relying on composition- or frequency-based subsequence features often suffer from limited interpretability and reduced accuracy, particularly on complex or imbalanced datasets. To address these limitations, we propose GeneNSPCla (Genomic Negative Sequential Pattern-based Classification), a novel viral classification framework based on Negative Sequential Patterns (NSPs) that extracts discriminative absence-based features from nucleotide sequences of RNA viral genomes. By transforming these NSPs into numerical feature vectors and integrating them into multiple supervised classifiers, GeneNSPCla effectively captures both presence and absence signals in viral sequences. Furthermore, we propose a negative pattern mining algorithm adapted for processing genomic data: GONPM+, which can discover longer and more biologically meaningful negative sequential patterns. The experimental results demonstrate that the average accuracy of GONPM+ in 8 classifiers has improved by 10.03% compared to the original negative pattern mining algorithm and by 24.75% compared to the positive pattern mining algorithm. These findings highlight the effectiveness of incorporating absence-based sequential information, providing a new and complementary perspective for viral genome analysis and classification.

Vision-language models (VLMs) are increasingly used as automated judges for multimodal systems, yet their scores provide no indication of reliability. We study this problem through conformal prediction, a distribution-free framework that converts a judge's point score into a calibrated prediction interval using only score-token log-probabilities, with no retraining. We present the first systematic analysis of conformal prediction for VLM-as-a-Judge across 3 judges and 14 visual task categories. Our results show that evaluation uncertainty is strongly task-dependent: intervals cover ~40% of the score range for aesthetics and natural images but expand to ~70% for chart and mathematical reasoning, yielding a quantitative reliability map for multimodal evaluation. We further identify a failure mode not captured by standard evaluation metrics, ranking-scoring decoupling, where judges achieve high ranking correlation while producing wide, uninformative intervals, correctly ordering responses but failing to assign reliable absolute scores. Finally, we show that interval width is driven primarily by task difficulty and annotation quality, i.e., the same judge and method yield 4.5x narrower intervals on a clean, multi-annotator captioning benchmark. Code: https://github.com/divake/VLM-Judge-Uncertainty

Deep learning models are widely deployed in safety-critical domains, but remain vulnerable to adversarial attacks. In this paper, we study the adversarial robustness of NTK neural networks in the context of nonparametric regression. We establish minimax optimal rates for adversarial regression in Sobolev spaces and then show that NTK neural networks, trained via gradient flow with early stopping, can achieve this optimal rate. However, in the overfitting regime, we prove that the minimum norm interpolant is vulnerable to adversarial perturbations.

Dimensionality reduction methods such as UMAP and t-SNE are central tools for visualising high-dimensional data, but their local-neighborhood objectives can preserve sampling noise while distorting global topology. We show that standard local metrics reward this noise memorisation: top-performing embeddings invent cycles and disconnected islands absent from the data. We introduce a topology-faithfulness benchmark based on noisy manifolds with known homology, tune DiRe against it, and find Pareto-optimal configurations that match or beat GPU-accelerated UMAP on classification while recovering exact first Betti numbers on stress tests. On 723K arXiv paper embeddings, DiRe preserves 3-4 times more topological structure than UMAP at comparable wall-clock.

Recent self-supervised pre-training methods for electroencephalogram (EEG) have shown promising results. However, the pre-trained models typically require full fine-tuning on each downstream task individually to achieve good performance. In practical applications involving multiple tasks, utilizing a separate model for each task is not ideal regarding computational and spatial cost. In this study, we go one step further and explore the simultaneous adaptation of a pre-trained model to multiple different tasks. The EEG signals exhibit significant heterogeneity due to their collection from various subjects using diverse devices and experimental setups, resulting in potential conflicts among different tasks that impede joint optimization. To tackle this challenge, we propose MTEEG, a multi-task EEG analysis framework which incorporates task-specific low-rank adaptation (LoRA) modules to disentangle the parameter space and alleviate task conflicts. To investigate the trade-off between task specification and interaction, we propose three variants of MTEEG that integrate the LoRA modules in different ways and evaluate them on six downstream tasks, demonstrating that MTEEG can surpass state-of-the-art single-task methods on the majority of metrics. MTEEG shows the potential of multi-task EEG analysis and promotes the development of general-purpose brain-computer interfaces in the future.

Forecasting when AI systems will become capable of meaningfully accelerating AI research is a central challenge for AI safety. Existing benchmarks measure broad capability growth, but may not provide ample early warning signals for recursive self-improvement. We propose measuring AI's capability to autonomously implement end-to-end machine learning pipelines from past AI research breakthroughs, given a minimal task description. By providing a concise task description instead of the full prior work as reference, we hope to better elicit emerging AI research taste. We introduce a proof-of-concept benchmark in which frontier coding agents autonomously implement an AlphaZero-style machine learning pipeline for Connect Four on consumer hardware within a three-hour budget, and we evaluate the resulting game AIs in a round-robin tournament anchored to the Pascal Pons Connect Four solver. Across four agents with eight trials each, we find substantial differentiation: Claude Opus 4.7 won as first-mover against Pons in seven of eight trials, statistically significantly better than the other agents tested, none of which exceeded two of eight. The task, which no frontier agent could reliably complete when we began development in January of 2026, is now near-saturation. Our evaluation also surfaced anomalous behavior in GPT-5.4, which consistently used far less of its allocated time budget than other agents. A follow-up 16-trial probe using shorter, less evaluation-coded prompts substantially increased GPT-5.4's time-budget usage, consistent with but not diagnostic of sandbagging; Bradley-Terry ratings across probe conditions showed only directional differences, despite significant differences in time-budget usage. We release our data, code, and prompts to support reproduction and extension.

Large language models have transformed AI-assisted software engineering, but current research remains biased toward high-resource languages such as Python, with weaker performance in languages like Rust and OCaml. Since real-world systems are inherently polyglot, robust multilingual code intelligence is crucial. This survey focuses on two key tasks: multilingual code generation from shared natural-language requirements, and multilingual code translation that preserves semantics across languages. It reviews representative methods, benchmarks, and evaluation metrics, and highlights challenges and opportunities for trustworthy cross-language generalization.

Frontier AI companies first deploy their most advanced models internally, for weeks or months of safety testing, evaluation, and iteration, before a possible public release. For example, Anthropic recently developed a new class of model with advanced cyberoffense-relevant capabilities, Mythos Preview, which was available internally for at least six weeks before it was publicly announced. This internal use creates risks that external deployment frameworks may fail to address. Legal frameworks, notably California's Transparency in Frontier Artificial Intelligence Act (SB 53), New York's Responsible AI Safety And Education (RAISE) Act, and the EU's General-Purpose AI Code of Practice, all discuss risks from internal AI use. They require frontier developers to make and implement plans for how to manage risks from internal use, and to produce internal use risk reports describing their safeguards and any residual risks. This guide provides a harmonized standard for companies to produce internal use risk reports suitable for all three regulatory frameworks. It is addressed primarily to evaluation and safety teams at frontier AI developers, and secondarily to regulators and auditors seeking to understand what good reporting looks like. Given the pace of AI R&D automation and the limited external visibility into how companies use their most capable models internally, regular and detailed risk reporting may be one of the few mechanisms available to ensure that the risks from internal AI use are identified and managed before they materialize. Whenever a substantially more capable or riskier model is deployed internally, the developer should create a risk report and argue why the model is safe to deploy. We structure the reporting framework around two threat vectors -- autonomous AI misbehavior and insider threats -- and three risk factors for each: means, motive, and opportunity.

While preference optimization is crucial for improving visual generative models, how to effectively scale this paradigm remains largely unexplored. Current open-source preference datasets contain conflicting preference patterns, where winners excel in some dimensions but underperform in others. Naively optimizing on such noisy datasets fails to learn preferences, hindering effective scaling. To enhance robustness against noise, we propose Poly-DPO, which extends the DPO objective with an additional polynomial term that dynamically adjusts model confidence based on dataset characteristics, enabling effective learning across diverse data distributions. Beyond biased patterns, existing datasets suffer from low resolution, limited prompt diversity, and imbalanced distributions. To facilitate large-scale visual preference optimization by tackling data bottlenecks, we construct ViPO, a massive-scale preference dataset with 1M image pairs at 1024px across five categories and 300K video pairs at 720p+ across three categories. State-of-the-art generative models and diverse prompts ensure reliable preference signals with balanced distributions. Remarkably, when applying Poly-DPO to our high-quality dataset, the optimal configuration converges to standard DPO. This convergence validates dataset quality and Poly-DPO's adaptive nature: sophisticated optimization becomes unnecessary with sufficient data quality, yet remains valuable for imperfect datasets. We validate our approach across visual generation models. On noisy datasets like Pick-a-Pic V2, Poly-DPO achieves 6.87 and 2.32 gains over Diffusion-DPO on GenEval for SD1.5 and SDXL, respectively. For ViPO, models achieve performance far exceeding those trained on existing open-source preference datasets. These results confirm that addressing both algorithmic adaptability and data quality is essential for scaling visual preference optimization.

Word embeddings are fundamental to natural language processing, yet traditional approaches represent each word with a single vector, creating representational bottlenecks for polysemous words and limiting semantic expressiveness. While multi-anchor representations have shown promise by representing words as combinations of multiple vectors, they have been limited to small-scale models due to computational inefficiency and lack of integration with modern transformer architectures. We introduce Adaptive Dictionary Embeddings (ADE), a framework that successfully scales multi-anchor word representations to large language models. ADE makes three key contributions: (1) Vocabulary Projection (VP), which transforms the costly two-stage anchor lookup into a single efficient matrix operation; (2) Grouped Positional Encoding (GPE), a novel positional encoding scheme where anchors of the same word share positional information, preserving semantic coherence while enabling anchor-level variation; and (3) context-aware anchor reweighting, which leverages self-attention to dynamically compose anchor contributions based on sequence context. We integrate these components into the Segment-Aware Transformer (SAT), which provides context-aware reweighting of anchor contributions at inference time. We evaluate ADE on AG News and DBpedia-14 text classification benchmarks. With 98.7% fewer trainable parameters than DeBERTa-v3-base, ADE surpasses DeBERTa on DBpedia-14 (98.06% vs. 97.80%) and approaches it on AG News (90.64% vs. 94.50%), while compressing the embedding layer over 40x -- demonstrating that multi-anchor representations are a practical and parameter-efficient alternative to single-vector embeddings in modern transformer architectures.

Given the increased use of LLMs in financial systems today, it becomes important to evaluate the safety and robustness of such systems. One failure mode that LLMs frequently display in general domain settings is that of sycophancy. That is, models prioritize agreement with expressed user beliefs over correctness, leading to decreased accuracy and trust. In this work, we focus on evaluating sycophancy that LLMs display in agentic financial tasks. Our findings are three-fold: first, we find the models show only low to modest drops in performance in the face of user rebuttals or contradictions to the reference answer, which distinguishes sycophancy that models display in financial agentic settings from findings in prior work. Second, we introduce a suite of tasks to test for sycophancy by user preference information that contradicts the reference answer and find that most models fail in the presence of such inputs. Lastly, we benchmark different modes of recovery such as input filtering with a pretrained LLM.

Selectively programming large number of non-volatile synaptic weights without compromising scalability is a key challenge for in-memory computing. Here, we demonstrate remote programming of synaptic weights in series-connected chains of 11 vortex-based magnetic tunnel junctions using broadcast radiofrequency signals applied through a shared strip line. The programming relies on frequency-selective reversal of the vortex-core polarity and therefore does not require individual access lines or selector devices. By reconfiguring the binary states of these chains, we reshape the weighted sums they perform on frequency-multiplexed RF inputs. Using a 22-synapse network composed of two such chains, we remotely reconfigure the same hardware to perform two distinct tasks: handwritten-digit classification and drone RF-signature identification. The digit-optimized configuration reaches 94.91 +/- 0.26% accuracy on handwritten digits but only 13.17 +/- 0.47% on drone RF signatures, whereas the drone-optimized configuration reaches 97.33 +/- 0.62% on drones but only 47.59 +/- 1.5% on digits. Broadcast RF programming thus provides a compact and scalable route to rapidly reconfigurable spintronic neuromorphic hardware.

Edge computing environments impose strict constraints on energy consumption and latency, making the deployment of deep neural networks a significant challenge. Therefore, smart and adaptive inference strategies that dynamically balance computational cost or latency with predictive accuracy are critical in edge computing scenarios. In this work, we build on Adaptive Deep Neural Networks (ADNNs) that employ the Multi-Armed Bandit (MAB) framework. Current literature leverages the first version of the Upper Confidence Bound (UCB1) strategy to dynamically select the optimal confidence threshold, enabling efficient early exits without sacrificing accuracy. However, we introduce four additional Upper Confidence Bound strategies in ADNNs, namely UCB-V, UCB-Tuned, UCB-Bayes, and UCB-BwK, and perform, for the first time, a comparative study of these strategies with respect to trade-offs between accuracy, energy consumption, and latency. The proposed UCB strategies are employed on the ResNet and MobileViT neural networks, and are evaluated on the benchmark datasets of CIFAR-10, CIFAR-10.1, and CIFAR-100. Experimental results demonstrate that all strategies achieve sub-linear cumulative regret, with UCB-Bayes converging the fastest, followed by UCB-Tuned and UCB-V. Finally, UCB-V and UCB-Tuned dominate the Pareto Frontiers of accuracy-latency and accuracy-energy trade-offs.

The rapid growth in the size of large language models has necessitated the partitioning of computational workloads across accelerators such as GPUs, TPUs, and NPUs. However, these parallelization strategies incur substantial data communication overhead significantly hindering computational efficiency. While communication-computation overlap presents a promising direction, existing data slicing based solutions suffer from tail latency. To overcome this limitation, this research introduces a novel communication-computation overlap technique to eliminate this tail latency in state of the art overlap methods for distributed LLM training. The aim of this technique is to effectively mitigate communication bottleneck of tensor parallelism and data parallelism for distributed training and inference. In particular, we propose a novel method termed CommFuse that replaces conventional collective operations of reduce-scatter and all-gather with decomposed peer-to-peer (P2P) communication and schedules partitioned computations to enable fine-grained overlap. Our method provides an exact algorithm for reducing communication overhead that eliminates tail latency. Moreover, it presents a versatile solution compatible with data-parallel training and various tensor-level parallelism strategies, including TPSP and UP. Experimental evaluations demonstrate that our technique consistently achieves lower latency, superior Model FLOPS Utilization (MFU), and high throughput.

Federated learning enables a population of clients to collaboratively train machine learning models without exchanging their raw data, but standard algorithms such as FedAvg suffer from slow convergence and high communication and memory costs in heterogeneous, resource-constrained environments. We introduce FedSLoP, a federated optimization algorithm that combines stochastic low-rank subspace projections of gradients, thereby reducing the dimension of communicated and stored updates while preserving optimization progress. On the theoretical side, we develop a detailed nonconvex convergence analysis under standard smoothness and bounded-variance assumptions, showing that FedSLoP is guaranteed to converge to a first-order stationary point at a rate of $O(1/\sqrt{NT})$. On the empirical side, we conduct extensive experiments on federated MNIST classification with heterogeneous data partitions, showing that FedSLoP substantially reduces communication volume and client-side memory while achieving competitive or better accuracy compared with FedAvg and representative sparse or low-rank baselines. Together, our results demonstrate that random subspace momentum methods such as FedSLoP provide a principled and effective approach to communication- and memory-efficient federated learning. Codes are available at: https://github.com/pkumelon/FedSLoP.git.

On-policy distillation (OPD) has shown strong potential for transferring reasoning ability from frontier or domain-specific models to smaller students. While effective on static single-turn tasks, its behavior in multi-turn agent settings remains underexplored. In this work, we identify a key limitation of vanilla OPD in such settings, which we term Trajectory-Level KL Instability. Specifically, we observe that KL divergence increases together with a drop in success rate, and even after convergence, the KL remains high, leading to unstable training. This instability arises from inter-turn error compounding: as errors accumulate, the student is driven beyond the teacher's effective support, rendering the supervision signal unreliable. To address this, we propose TCOD (Temporal Curriculum On-Policy Distillation), a simple yet effective framework that controls the trajectory depth exposed to the student and progressively expands it from short to long with a curriculum schedule. Experimental results across four student-teacher pairs on three multi-turn agent benchmarks (ALFWorld, WebShop, ScienceWorld) show that TCOD mitigates KL escalation and enhances KL stability throughout training, improving agent performance by up to 18 points over vanilla OPD. Further evaluations show that TCOD can even surpass the teacher's performance and generalize to tasks on which the teacher fails. Our code is available at https://github.com/kokolerk/TCOD.

Deep learning approaches have shown remarkable promise in turbulence closure modeling for large eddy simulations (LES). The differentiable physics paradigm uses the so-called a-posteriori approach for learning by embedding a neural network closure directly inside the solver and optimizing its learnable parameters against ground truth time-series data which may be observed sparsely. This addresses a key limitation of a-priori learning where direct numerical simulation (DNS) data is used to approximate the subgrid stress with the assumption of a filter. However, closures that are trained in this manner frequently lead to unstable deployments due to the mismatch between the assumed filter and the effect of numerical discretizations. However, a-posteriori learning incurs high computational costs due to the need to backpropagate gradients through an LES solver. Furthermore, a-posteriori methods are challenging to apply broadly since they require significant modification of existing solvers. Finally, these approaches have also been observed to be limited when generalization is desired across different numerical schemes. In this work, we discuss a novel approach for the deep learning of turbulence closure models motivated by the continuous data assimilation (CDA) approach (also known as nudging). Our approach enables a-priori training of closures for coarse-grid LES, treating DNS data as sparse observations. This approach enables the deep learning model to successfully learn the required forcing to capture the ground-truth statistics while maintaining long term stability without needing adjoints or backpropagation through the solver. We train and evaluate the model's ability to adapt to different numerical and temporal schemes. Additionally, we analyse the model behavior with varying numerical discretization errors and compare its predictions to traditional closure models.

Sparse autoencoders (SAEs) extract millions of interpretable features from a language model, but flat feature inventories aren't very useful on their own. Domain concepts get mixed with generic and weakly grounded features, while related ideas are scattered across many units, and there's no way to understand relationships between features. We address this by first constructing a strict domain-specific concept universe from a large SAE inventory using contrastive activations and a multi-stage filtering process. Next, we build two aligned graph views on the filtered set: a co-occurrence graph for corpus-level conceptual structure, organized at multiple levels of granularity, and a transcoder-based mechanism graph that links source-layer and target-layer features through sparse latent pathways. Automated edge labeling then turns these graph views into readable knowledge graphs rather than unlabeled layouts. In a case study on a biology textbook, these graphs recover coherent chapter and subchapter-level structure, reveal concepts that bridge neighboring topics, and transform messy sentence-level activity containing thousands of features into compact, readable views that illustrate the model's local activity. Taken together, this reframes a flat SAE inventory as an internal knowledge graph that converts feature-level interpretability into a global map of model knowledge and enables audits of reasoning faithfulness.

The discovery of novel materials is critical for global energy and quantum technology transitions. While deep learning has fundamentally reshaped this landscape, existing predictive or generative models typically operate in isolation, lacking the autonomous orchestration required to execute the full discovery process. Here we present ElementsClaw, an agentic framework for materials discovery that synergizes Large Atomic Models (LAMs) with Large Language Models (LLMs). In response to varied human queries, ElementsClaw orchestrates a suite of LAM tools finetuned from our proposed 1-billion-parameter model Elements for atomic-scale numerical computation, while leveraging LLMs for high-level semantic reasoning. This shift moves AI-driven materials science from isolated processes toward integrated and human interactive discovery. Applied to superconductors, ElementsClaw screens 2.4 million crystals in just 28 GPU hours to identify 68,000 high-confidence candidates (The complete dataset of screened superconductors is available at https://developer.damo-academy.com/material), expanding known superconducting space by orders of magnitude compared to datasets curated over decades. Critically, ElementsClaw achieves a high success rate in identifying superconductors hidden in literature and discovers four novel experimentally verified superconductors, exemplified by Zr3ScRe8 with a transition temperature of 6.8 K and HfZrRe4 at 6.7 K. Together, our results establish a knowledge integrated, autonomously orchestrated, and experimentally grounded paradigm for materials discovery.

Silicon samples are increasingly used as a low-cost substitute for human panels and have been shown to reproduce aggregate human opinion with high fidelity. We show that, in the alignment-relevant domain of philosophy, silicon samples systematically collapse heterogeneity. Using data from $N = {277}$ professional philosophers drawn from PhilPeople profiles, we evaluate seven proprietary and open-source large language models on their ability to replicate individual philosophical positions and to preserve cross-question correlation structures across philosophical domains. We find that language models substantially over-correlate philosophical judgments, producing artificial consensus across domains. This collapse is associated in part with specialist effects, whereby models implicitly assume that domain specialists hold highly similar philosophical views. We assess the robustness of these findings by studying the impact of DPO fine-tuning and by validating results against the full PhilPapers 2020 Survey ($N = {1785}$). We conclude by discussing implications for alignment, evaluation, and the use of silicon samples as substitutes for human judgment. The code of this project can be found at https://github.com/stanford-del/silicon-philosophers.

The health condition of components in civil infrastructures can be described by various discrete states according to their performance degradation. Inferring these states from measurable responses is typically an ill-posed inverse problem. Although Bayesian methods are well-suited to tackle such problems, computing the posterior probability density function (PDF) presents challenges. The likelihood function cannot be analytically formulated due to the unclear relationship between discrete states and structural responses, and the high-dimensional state parameters resulting from numerous components severely complicates the computation of the marginal likelihood function. To address these challenges, this study proposes a novel Bayesian inversion paradigm for discrete variables based on Probabilistic Graphical Models (PGMs). The Markov networks are employed as modeling tools, with model parameters learned from data and structural topology prior. It has been proved that inferring this PGM produces the same probabilistic estimation as the posterior PDF derived from Bayesian inference, which effectively solves the above challenges. The inference is accomplished by Graph Neural Networks (GNNs), and a graph property-based GNN training strategy is developed to enable accurate inference across varying graph scales, thereby significantly reducing the computational overhead in high-dimensional problems. Both synthetic and experimental data are used to validate the proposed framework

In 2016, Viazovska famously solved the sphere packing problem in dimension $8$, using modular forms to construct a 'magic' function satisfying optimality conditions determined by Cohn and Elkies in 2003. In March 2024, Hariharan and Viazovska launched a project to formalize this solution and related mathematical facts in the Lean Theorem Prover. A significant milestone was achieved in February 2026: the result was formally verified, with the final stages of the verification done by Math, Inc.'s autoformalization model 'Gauss'. We discuss the techniques used to achieve this milestone, reflect on the unique collaboration between humans and Gauss, and discuss project objectives that remain.

Orbital energy splittings are important quantum dot parameters for the operation of hole spin qubits. They are known to depend on the lateral confinement of the quantum dots. However, when changing top, plunger gate voltages, which are the typical control parameter for qubit applications, such energy splitting changes are typically negligible, both as measured in experiment and as assumed in effective theories. Here, we study the singlet-triplet (ST) splittings, which depend on the orbital splittings, of a double quantum dot (DQD) in a Ge/SiGe heterostructure using photon-assisted tunneling (PAT) and pulsed-gate spectroscopy. We find that the ST splittings have a surprising, strong dependence on the top gate voltages, leading to anomalous PAT measurements. We combine data from both measurements in a model that well describes the linear gate-voltage dependence of the ST splittings. Finally, we show that the ST splittings of the two dots exhibit similar linear gate-voltage dependences when the device is retuned such that their ratio is significantly different.

The study of Entanglement Asymmetry has emerged in recent years as a powerful tool to characterise the symmetry properties of quantum states in relation to a given charge operator through the lens of entanglement. While extremely powerful and general, the standard definition of asymmetry introduces significant non-Gaussian features in free-fermionic systems, leading to certain analytical limitations. In this work, we introduce an asymmetry measure that remains strictly within the Gaussian manifold and analyse its properties. In particular, we show that it quantifies the minimal distance between a Gaussian state and the manifold of symmetric Gaussian states. We further demonstrate that this measure captures the established dynamical signatures of entanglement asymmetry, such as the Mpemba effect, symmetry restoration, and the lack thereof. The Gaussian structure allows these novel asymmetry measures to be computed exactly using correlation matrix techniques, and to be described asymptotically through the quasiparticle picture. We also comment on the possibility of using charge fluctuations to characterise the asymmetry of a Gaussian state.

Ten years after the experimental discovery of Weyl semimetals, theoretical and experimental work has pointed to the possibility of realizing surface-only superconductivity at relatively high temperatures in these materials. A consensus is developing that this unusual form of superconductivity is mediated by surface electronic states unique to Weyl semimetals, known as Fermi arcs. In this work, we show that the topological protection of these exotic states can be exploited to engineer high critical temperatures. Motivated by a real-material example (PtBi$_2$), we demonstrate that surface van Hove singularities can be induced by depositing a suitable additional layer on top of the Weyl surface. We also investigate the role of these singularities in raising the critical temperature, showing that it is significantly enhanced when the chemical potential lies in their vicinity. More generally, our results demonstrate how topological protection can be exploited to manipulate surface electronic states, thereby opening experimentally accessible routes toward engineering high-temperature two-dimensional superconductivity and other exotic phases.

Euclidean random matrices arise in a wide range of physical systems where interactions are determined by spatial configurations, including disordered media and cooperative phenomena in atomic ensembles. Unlike classical random matrix ensembles, their entries are strongly correlated through the geometry of the underlying random points, making their analytical treatment challenging. While global spectral properties such as the spectral density are relatively well understood, much less is known about extremal eigenvalues and the associated eigenvectors, despite their central role in applications. Here we address the problem of characterising the largest eigenvalue and the corresponding top eigenvector of large Euclidean random matrices with a quadratic kernel. For vectors in any dimension $d \geq 1$ drawn independently from a common distribution, we show that both quantities can be computed within a unified replica-based framework, leading to a set of self-consistent equations involving a small number of parameters. This approach yields an explicit expression for the average largest eigenvalue, fully determined by low-order moments of the underlying distribution, and an analytical characterisation of the distribution of top eigenvector's components in the large-$N$ limit. We find that the top eigenvector exhibits a non-trivial geometric structure, with components concentrating on a hypersurface determined by the same parameters controlling the largest eigenvalue. We further perform extensive numerical simulations that confirm these predictions. More broadly, our work provides a general framework to access extremal spectral properties of Euclidean random matrices.

The paper studies constitutive modelling of Korteweg fluids. Thermodynamic consistency, i.e. compatibility with entropy balance law, is achieved using Liu's method of multipliers. Appropriate constitutive assumptions facilitated inclusion of the capillary effects in the specific entropy. Korteweg stresses are derived from the equilibrium conditions -- vanishing of the entropy production and its minimization in equilibrium. Material parameter in Korteweg stresses is allowed to depend on temperature, which turns out to be consistent with kinetic-theory results and leads to cross-coupling of mechanical and thermal effects. The generalized Gibbs' relation, which inherits the capillary effects, is derived as consequence, which is a peculiar feature of the Liu's method.

Self-propelled particles serve as minimal models for emulating the dynamic self-organization of microorganisms, yet most synthetic systems remain limited to a single mode of motion, namely active Brownian particles (ABPs). Here, we present an experimental strategy to encode various persistent random walks in ABPs by combining light-modulated propulsion strength with magnetic control of propulsion direction. Our system enables programmable Levy walks with tunable step-length distributions, run-and-tumble dynamics, self-avoiding random walks, and Gaussian walks, with on-demand switching between motion modes within a single experiment. In addition, particles are steered along complex trajectories such as Fibonacci spirals and nested polygons. Beyond single-particle behavior, we show that propulsion modes influence clustering dynamics by comparing ABPs with chiral active particles undergoing circular motion. These results establish a versatile platform for investigating how encoded motion at the level of individual particles governs transport, search strategies, and emergent organization in active matter systems.

To study a nonlocal electron transport in an aluminum superconducting quasi-one-dimensional structure, we measured negative nonlocal (local) direct current voltages in the structure in a magnetic field near the critical temperature. The structure is a normal-superconducting at $T_{cn}<T<T_{cw}$ ($T_{cn}$ and $T_{cw}$ are the critical temperatures for narrow and wide wires, respectively, making up this structure). Negative voltage arises due to a quasiparticle current flowing through the N-S interface. We plotted the experimental and theoretical temperature and magnetic-field dependences of current, resistance and voltage corresponding to the peak of negative voltage, taking into account either equilibrium or nonequilibrium superconducting fluctuations.

We investigate quantum transport in an off-diagonal Aubry--André--Harper chain. The periodic hopping modulation generates effective internal boundaries that strongly influence the transmission characteristics. We show that edge, in-band bulk, and band-edge bulk states can be clearly distinguished through their transport signatures. In particular, bulk states near the band edges exhibit behavior similar to edge states, with weak dependence on system size, whereas in-band bulk states display pronounced size-dependent oscillations. We further demonstrate that the chain--electrode coupling strength controls the broadening of transmission resonances and drives a crossover from tunneling-dominated to nearly ballistic transport. In addition, dephasing introduces distinct sensitivity across different state classes, depending on their degree of spatial localization. These results highlight the key role of internal boundaries and quantum coherence in governing transport in modulated one-dimensional systems.

Ultrafast scintillators are indispensable for precise timing in high-energy physics and medical diagnostics. Fundamentally constrained by the trade-off between emission rate and light yield, conventional scintillators remain kinetically trapped in the sub-nanosecond regime, failing to break 50-picosecond limit. Here, we demonstrate a strategy to bypass this limitation by harnessing the coherent radiative acceleration in weakly confined CsPbCl3 perovskite nanocrystals to generate an ultrafast photon burst. This effect originates from the giant oscillator strength, which we unlock by suppressing exciton-phonon scattering at mild cryogenic temperatures. Consequently, our scintillator achieves an unprecedented dominant lifetime of 13.11 ps alongside a high light yield of 21,851 ph/MeV. The resulting prompt photon emission rate more than 100 times higher than that of state-of-the-art ultrafast scintillators. We validate this breakthrough in realistic detection scenarios, achieving a coincidence time resolution of 30.8 ps and accurately resolving 13.5 ps electron bunches and 16.6 ps single-shot gamma-ray pulses. Our findings establish a robust coherent framework for next-generation ultrafast scintillators, pushing extreme radiation diagnostics into the picosecond frontier.

We investigate electronic transport across a junction between a Weyl semimetal (WSM) and a layered Chern insulator (LCI) in the presence of a magnetic field perpendicular to the interface. The topological mismatch between the gapless Weyl semimetal and the momentum-resolved chiral edge modes of the layered Chern insulator leads to interface Fermi-arc states with a qualitatively distinct connectivity: unlike WSM-WSM junctions, the interface Fermi arcs are forced to reconnect through the Brillouin-zone boundary rather than terminating at the projections of the Weyl nodes. We analyze the spectrum and compute the magneto tunnel conductance mediated by the interface-localized states. We find that the conductance increases linearly with magnetic field at low fields and saturates beyond a critical field to a constant value that is independent of microscopic details such as interface coupling, arc geometry, and lattice-scale parameters. This universal saturation reflects a transport mechanism governed by the topological charge pumping associated with the Chern layers, rather than magnetic breakdown between Fermi arcs. We further show that, under specific conditions, a junction between two distinct Weyl semimetals can exhibit a similar saturation behavior, thereby clarifying the topological origin of the observed universality.

Modern information technologies rely on controlling the flow of electrons through their charge and spin. A rapidly emerging alternative is to use the orbital motion of electrons, the way they circulate around atomic sites as a new carrier of information. This orbital angular momentum (OAM) could enable more energy-efficient devices and reduce reliance on scarce heavy elements, but how orbital currents are generated and transported, especially on ultrafast timescales, remains largely unknown. In this review, we introduce terahertz optorbitronics, an approach that uses ultrafast femtosecond laser pulses and terahertz radiation to observe orbital transport in real time. On timescales of quadrillionth of a second, this technique allows us to track how orbital currents are launched, propagate, and convert into electrical signals in nanoscale thin-film materials. Surprisingly, recent experiments have revealed conflicting pictures such as orbital currents may travel over tens of nanometres like ballistic waves or instead decay within just a few atomic layers, highlighting a fundamental unresolved question in the field. We explain how these ultrafast measurements can disentangle orbital motion from conventional spin transport, and we highlight new materials from engineered graphene to altermagnets, that could act as tunable sources of orbital currents. We also discuss how light, electrical gating, strain, and interface design can be used to actively control orbital transport and improve its conversion into usable electronic signals. By revealing orbital transport as a dynamic, non-equilibrium process, terahertz optorbitronics opens a new direction for nanoscale science, the one that could lead to faster, more efficient technologies operating beyond the limits of conventional spin-based electronics.

In modern condensed matter theory, phases of electronic matter--such as metals and insulators-are fundamentally distinguished by the presence or absence of charge-carrying quasiparticles or excitations near the Fermi surface at low temperatures. Here, we show that this criterion breaks down in Berry-curvature-dominated systems, where transport is governed by interband coherence across the entire Fermi sea. We develop a microscopic theory of quantum transport in bulk topological insulators with a vanishing density of states at the Fermi energy, for which the conventional Drude contribution is absent. We demonstrate that impurity-scattering-induced coherence decay generates a distinct longitudinal transport channel even in the topologically trivial regime, with edge contributions rigorously excluded. This mechanism yields a finite longitudinal conductivity even in the absence of carriers at the Fermi level and exhibits an unconventional scaling linear in impurity density in the dilute limit, in stark contrast to Drude behaviour. Importantly, this decoherence-induced conductance is inversely proportional to temperature, reminiscent of strange-metal behaviour, most prominently observed in cuprate superconductors above their critical temperature. Our findings reveal quantum decoherence as a fundamental origin of longitudinal transport beyond the Drude paradigm, challenging the traditional distinction between metals and insulators.

Magnetic skyrmions are topologically protected spin textures that can act as reconfigurable nanoscale information carriers. In synthetic antiferromagnets (SAFs), interlayer exchange coupling offers an additional control parameter beyond the interfacial Dzyaloshinskii-Moriya interaction (DMI) and magnetic anisotropy. Here, we engineer a SAF composed of two chemically distinct ferromagnets (CoB and CoFeB), in which the external magnetic field and interlayer exchange act asymmetrically on the sublattices. The competition of these effects, acting as a resultant effective-field, gives rise to two distinct skyrmion families in different field regimes. In large fields, conventional-polarity skyrmions nucleate, with core antiparallel to the external field, whereas in smaller fields an inverse-polarity skyrmion state emerges as the effective-field reverses sign and almost saturates the CoFeB layers. Return-point memory measurements confirm independent nucleation pathways for the two families. Using element-resolved x-ray magnetometry, correlative magnetic force and Lorentz transmission electron microscopies, and parameter-matched micromagnetic modelling, we show that all textures reside only in the CoFeB layers, which experience a Ruderman-Kittel-Kasuya-Yosida (RKKY) exchange field originating from the CoB layers. This effective-field method provides a robust route to programmable three-dimensional spin textures with controlled polarity in selected layers of a multilayer with potential for applications in skyrmion-based computing and spin-logic architectures.

Moiré superlattices based on rhombohedral multilayer graphene have emerged as a highly tunable platform for engineering correlated topological phases. Here, we systematically investigate the transport properties of the hole-doped side in rhombohedral tetralayer graphene/ hexagonal boron nitride (hBN) moiré superlattices across a range of twist angles and alignment orientations. Notably, we observed multiple high-Chern-number Chern insulators, including the previously reported integer Chern insulator with Chern number C = -4 at moiré filling factor v = -1 and newly discovered symmetry-broken Chern insulating states with C = +3, $\pm$2, $\pm$1 at fractional moiré fillings of v = -2.5 or -2.6. These Chern insulating states emerge in both hBN alignment, but exhibit a sensitive moiré wavelength dependence. Our findings demonstrate the exceptional tunability of these high-Chern-number states via moiré wavelength, displacement electric field and external magnetic field, underscoring the distinct topological landscape realized in hole-doped RTG/hBN moiré superlattices.

We consider a general model of a Luttinger liquid with broken parity and time reversal symmetry, but with their composite symmetry intact. Such a scenario can be due to a combination of torsion and a Zeeman field in nanowires, or a result of bringing different helical Luttinger liquids into proximity. The broken symmetries result in a band structure with no axis of symmetry, and therefore with asymmetric velocities between left and right moving contributions. By taking a general spin-full model with all possible scattering and interaction terms in the bosonic model we show that generically the spin degree of freedom becomes gapped out, resulting in an effective spinless Luttinger liquid with asymmetric velocities. Our work generalizes and extends previous studies which focused on a minimal model of a spinless Luttinger liquid. We further demonstrate that a possible experimental signature of the asymmetry of such asymmetric models can be seen in the spectral function.

Stochastic thermodynamics is a framework for describing non-equilibrium processes at the level of fluctuating trajectories, where the state of a system evolves as a stochastic time series, allowing thermodynamic quantities such as work, heat, and entropy production to be defined along individual realizations rather than at the ensemble level only. Over the past three decades, the field has yielded fundamental results, including fluctuation theorems and several universal bounds, such as thermodynamic uncertainty relations, speed limit theorems, and many others. Many of them have been tested on a range of experimental platforms. This Perspective reviews recent developments in stochastic thermodynamics that extend its scope beyond its traditional domains, including systems with memory and hidden degrees of freedom, microscopic approaches to interacting and active matter, and geometric formulations based on optimal transport. Next, the Perspective surveys the challenges that arise when applying these ideas to macroscopic and complex systems, where the link between statistical irreversibility and thermodynamic dissipation becomes less direct. Finally, emerging applications in non-physical contexts are highlighted, including computation, biological systems, and social dynamics. Transcending the traditional boundaries of physics, these developments catalyze an unorthodox framework to tackle the thermodynamics of complex systems.

To construct dispersion relations for diffusion or oscillation processes on random networks, it is necessary to obtain effective length scales for the eigenvectors of a graph Laplacian matrix, whose eigenvalues represent inverse time scales. For this purpose, we adapt a method originally introduced in condensed-matter physics to estimate correlation lengths for disordered materials as the ratio of volume to interface area [P. Debye, H.R. Anderson and H. Brumberger, J. Appl. Phys. 28, 679 (1957)]. In a graph setting of vertices connected by edges, we interpret this as the ratio of twice the total number of edges to the number of edges connecting vertices bearing values of different sign on the particular eigenvector. After describing the method and the necessary concepts in pedagogical detail, we apply it to nine different graphs representing natural and artificial networks, including two tree graphs without and with random shortcuts, the nervous system of a roundworm, a food web, a social network of dolphins, an electrical power grid, and a model porous material. The results identify both distributed and localized eigenvectors. They are given in graphical format showing example eigenvectors, dispersion relations, and discrete densities of states, as well as tables summarizing the main numerical results.

Predicting species persistence within ecological communities is a fundamental challenge for both empirical and theoretical ecology. Existing methods span from mechanistic models, whose parameters are difficult to estimate from data, to statistical tools whose context-specific parameters are less interpretable. Here, we present a general framework, grounded in the statistical physics of complex systems, that integrates the key processes governing species survival into a single measurable quantity: the competitive balance. This metric quantifies a focal species' vulnerability beyond its abundance by incorporating the diversity of dispersal strategies and the structure of interspecific interactions within the community. Crucially, it can be inferred from spatial abundance data, thus circumventing the need to estimate species traits or dispersal parameters. Our results reveal that greater heterogeneity in dispersal strategies reduces vulnerability for a given abundance. Although we validate the framework using tropical and temperate forest data, it can be applied to a range of different ecosystems, providing a systemic and interpretable tool for assessing a context-dependent species vulnerability that accounts for its interactions with the entire community.

Momentum-space nonsymmorphic symmetries have recently attracted significant interest in both artificial and condensed-matter crystals, whereas real-space nonsymmorphic symmetries have long played an important role in the study of crystalline topological phases. Here, we establish a general theory of momentum-space crystallographic groups that emerge from projective representations of real-space crystallographic groups in the presence of gauge flux, applicable in particular to real-space nonsymmorphic groups. A central result is a flux-mediated ``bi-nonsymmorphicity'' relation that reveals a structural correspondence between real-space and momentum-space nonsymmorphicity mediated by gauge flux. This relation implies that, under a symmetric gauge flux, real-space nonsymmorphicity can enforce momentum-space nonsymmorphicity, and that in some cases a symmetric gauge flux requires nonsymmorphicity in both real and momentum space. Our work not only identifies a fundamental structure in projective crystal symmetries, but also provides guiding principles for designing artificial crystals and condensed-matter platforms that exhibit both real-space and momentum-space nonsymmorphic symmetries.

Two qubits strongly coupled to a common bosonic reservoir can become entangled with each other, despite having no direct interaction. In equilibrium, such coupling-induced coherences can be described by the mean-force Gibbs state. Here we derive approximate, analytic expressions for the two-qubit mean-force Gibbs state, and use these to characterize equilibrium qubit-qubit entanglement mediated by a thermal reservoir. Entanglement, which is highest at lowest temperatures, is a non-monotonic function of the system-reservoir coupling strength. Moreover, we find that broadening the reservoir spectral density beyond a single mode, as is realistic for typical baths, can enhance the qubit entanglement. Our results provide a comprehensive understanding of reservoir-mediated two-qubit entanglement in thermal equilibrium and provide a benchmark to compare with numerical methods, as well as demonstrating the utility of strong system-reservoir coupling as a resource.

This article presents a brief account of Amir O. Caldeira's contributions to the theory of quantum Brownian motion. Motivated by its importance, we outline the description of Brownian motion in the quantum regime following Caldeira's first works. In this context, we particularly highlight the effect of dissipation on the tunneling rate out of a metastable state. We then journey along the alternative ways to approach quantum Brownian motion developed by Caldeira during his career, which go beyond the so-called Caldeira-Leggett model. We conclude by summarizing some of Caldeira's contributions to contemporary fields such as the theory of quantum decoherence and quantum thermodynamics, that were strongly inspired by his eponymous approach to quantum Brownian motion.

Numerical simulation of the dynamics of planar two- and three-layer molecular structures formed by $β$-sheets of polyglycine peptide chains and systems of parallel Kevlar (para-aramid) molecules placed on a graphene sheet has been performed. It is shown that in these structures the $β$-sheets retain their shape, due to the presence of parallel chains of hydrogen bonds, up to a temperature of $T=800$K. An even higher stability is exhibited by the system of parallel Kevlar molecules. Here, the parallel chains of hydrogen bonds between peptide groups of neighboring molecules are preserved even at higher temperatures. The performed modeling allows us to conclude that the addition of graphene to Kevlar fibers can significantly increase their thermal stability.

When a hydrophilic and deformable porous material is immersed in a bath, it may absorb the solvent and expand by several times its volume, thus forming a highly soft and porous hydrogel. A stress applied on the soft hydrogel surface deforms it and forces the absorbed solvent to move by flowing through the network of pores. This coupled phenomenon sets the framework of poroelasticity. Moreover, polymeric gels are often used in ultra-thin coatings to tune surface properties. Together with the characteristic poroelastic coupling, this thinness challenges the modelling of their response. In this article, we derive the point-force mechanical response of a thin, permeable and poroelastic layer bounded to a rigid substrate. We show that the gel surface is only deformed around the indentation point, within a radius on the order of the layer thickness. The obtained Green's function can be directly used to predict the space- and time-dependent surface deformation of the gel. Our findings are relevant for a broad range of applications, such as indentation experiments on swollen gels, thin membranes or soft and living systems, as well as lubrication problems involving a soft and porous wall, for instance in microfluidics.

Natural materials achieve adaptive behavior through hierarchical organization and coupled mechanisms across scales. Their translation into engineering, however, remains largely heuristic. What is missing is a formal translation framework that carries biological design logic into engineered realization while preserving physical consistency across levels of abstraction. Here we present a category theoretic compositional framework for verified nature-derived design. The framework defines a category of stimulus response dynamical systems with natural and artificial subcategories. It introduces a structure preserving implementation functor from biological mechanics to engineered systems. It also formalizes a machine agnostic specification layer that links behavioral intent to executable fabrication programs. We instantiate the framework on the hygromorphic pinecone hierarchy as a representative biological case. We implement the full pipeline in Grasshopper, where formal specifications are translated into modular parametric scripts that preserve the compositional structure of the model. The resulting designs are fabricated by fused filament fabrication, evaluated experimentally, and tested against model predictions derived from the pipeline. The current implementation generates four actuator classes spanning two stimulus types and two kinematic responses. One actuator arises purely through composition from previously validated components, without additional manual derivation. The results show that compositionality can function not just as a descriptive language, but as a generative and system level verifiable method for mechanical material design. More broadly, the work provides a concrete route for embedding formal multiscale reasoning within increasingly computational, generative, and physics-driven design workflows.

We investigate the structural and dynamical properties of binary aluminum-titanium liquid metallic alloys, as a function of temperature and composition. We make use of MD-simulations, using a transferable machine-learning potential developed by Song et al. [Nature Communications 15, 10208 (2024)], and compare our results to experimental data. Although this potential was initially trained on solid properties, we find good agreement between the experimental data and the simulation results for the liquid state. The excess volume and compositional changes of the structure are captured well by the machine-learned potential. The simulation allows to disentangle local packing from chemical-ordering effects; the latter are found to be weak in Al-Ti. Dynamical quantities like the viscosity and the diffusion coefficients are also discussed.

Bias dependent oscillations in excitonic photoluminescence are observed in a mixed dimensional 0D 2D heterostructure. These oscillations arise from modulation by oscillatory DC photocurrent, which exhibits periodic negative differential resistance, indicating recurring charge accumulation within the heterostructure. The persistence of these oscillations across a macroscopic area of diameter around 200 microns suggests the presence of periodically correlated quantum phenomena over large length scales. Furthermore, bias dependent oscillations in the photo capacitance are observed, reflecting a periodic ordering and disordering of excitonic populations. Together, these observations point to a direct competition between coherent and incoherent electron tunnelling processes. The coupled oscillatory behaviour of photoluminescence, photocurrent, and photo capacitance highlights new opportunities for exciton-based quantum optoelectronic devices.

Chemically active mixtures exhibit complex patterns that emerge from the interplay of physical interactions and reactions among components. Individually, these two processes are well-understood: Physical interactions can give rise to phase separation, whereas reactions can form reaction-diffusion patterns. To understand the combination of both processes, we identify a Lyapunov functional for a class of chemical reactions. By minimizing this functional, we identify a generalized Gibbs phase rule that governs the number of coexisting patterns, and we demonstrate that complex patterns can be created by the modular combination of independent phases. Our theory unveils complex stationary patterns in chemically active mixtures and provides a framework for analyzing more complex systems.

This paper presents a numerical investigation of Turing patterns (TPs) utilizing the reaction-diffusion equation for the activator $u$ and the inhibitor $v$ on two- and three-dimensional lattices, discarding vertex fluctuations. The absence of vertex fluctuations means the absence of positional movement of $u$ and $v$. Consequently, $u$ and $v$ have values at spatially discrete points, such as the pigment cells in zebrafish and sea shells. Furthermore, the mechanical property is implemented through the Finsler geometry modeling technique. This technique incorporates the internal degree of freedom $\vecτ$, corresponding to the direction of mechanical stress. Additionally, a stress formula based on Gaussian bond potential is shown to be well-defined on the non-fluctuating lattices, and therefore, it enables the calculation of entropy for capturing the stress relaxation phenomenon in a manner analogous to that on fluctuating surfaces. The results of the study indicate that these biological patterns also exhibit responses to external mechanical forces similar to TPs on fluctuating membranes.

Experiments show that isochoric (constant-volume) conditions enhance supercooling stability relative to isobaric (constant-pressure) conditions. Here, combining Helmholtz equilibrium thermodynamics with a first-order perturbation methodology, we derive an inequality governing nucleation stability under volumetric constraint. The derivation provides a general thermodynamic proof that for any substance undergoing phase transformation in which the solid is less dense than the liquid, the Helmholtz driving force for solidification in isochoric systems is smaller than the Gibbs driving force in isobaric systems. Since nucleation rates depend exponentially on the inverse square of the driving force, this provides a thermodynamic basis for the observed suppression of nucleation rates. While a full stochastic treatment is beyond the scope of this work, the reduction in driving force implies a weakening of the bias toward growth of pre-critical fluctuations, increasing their probability of thermal dissolution. The analysis yields a dimensionless isochoric stability number. This number is computable from bulk thermodynamic data alone and provides a geometry-independent criterion for comparing metastable liquid stability across materials and conditions.

A dramatic difference between the vertical thermal conductance response of homogeneous and heterogeneous graphene/h-BN interfaces to external mechanical perturbations, is predicted. Homogeneous graphene and h-BN interfaces exhibit strong conductance reduction for both in-plane strain and interfacial twist. Conversely, the vertical thermal conductance of the heterogeneous graphene/h-BN junction is insensitive to twist deformations but shows significant increase or decrease under compressive or tensile strains, respectively. Our atomistic simulations predictions are rationalized by Fermi's golden rule and density of phonon modes analyses, indicating that vertical phonons and local stacking configurations have a central role in the interlayer heat transport behavior. A simple phenomenological model, based on local interlayer distance and stacking, captures well the dependence of vertical heat conductance on strain and twist deformations.

This work realizes a three-dimensionally (3D) nanoarchitected, all organic, "self-separating" lithium-ion electrochemical energy storage (EES) device that is cycled as a solid-state full cell. The device is enabled by a monolithic carbon anode with a co-continuous pore network, derived from the structure direction of resols by an ultra-large molar mass block copolymer (BCP), poly(styrene-block-2-dimethylaminoethyl methacrylate) (SA). Electropolymerization of a single-phase conductive and redox-active material, poly((2,3-dihydrothieno[3,4-b][1,4]dioxin-2-yl)methyl 9,10-dioxo-9,10-dihydroanthracene-2-carboxylate) (PAQEDOT), into the pore space provides the cathode of the cell. The device is electronically contacted to the relevant electrode network enabled by the co-continuous nature of each electrode. Electrochemical processing via cycling against external lithium in an electrolyte generates a solid electrolyte interphase (SEI) as a separator and lithiates the cell electrodes, after which the EES device is cycled in the solid state. While the full cell does not demonstrate high cyclability, the best full cell demonstrates a discharge capacity of 267 milliamp hours per gram (mAh/g). This work marks, to the best of knowledge of the authors, the first example of an all-organic materials derived 3D nanoarchitected EES device, as well as the first design of "self-separating" cell fabrication. Furthermore, generalization of the design to another co-continuous carbon form factor is demonstrated.

The transport properties of discrete-time random walks on ring networks with deterministic shortcuts are investigated through analytical and numerical methods. The network consists of a periodic chain where each node is connected to its nearest neighbors and to nodes located at a fixed distance $r$. Using the spectral properties of the transition matrix, we derive explicit expressions for the occupation probabilities and mean first-passage times (MFPTs). Contrary to the common expectation that shortcuts monotonically enhance transport, we find that the MFPT between distant nodes develops a highly non-monotonic dependence on the shortcut length. Beyond a threshold value, the MFPT landscape exhibits a hierarchy of maxima and minima organized in a self-similar pattern associated with commensurability relations between the shortcut length and the system size. The scaling behavior of these extrema reveals regimes where transport efficiency is either strongly enhanced or suppressed. We further analyze the mean squared displacement and the influence of stochastic resetting, showing that resetting amplifies the oscillatory MFPT structure and induces strongly nonuniform stationary distributions. These results demonstrate that the spatial organization of long-range connections plays a crucial role in determining transport efficiency in networks.

The novel functional dimensional regularization (FDR) scheme has proven capable of yielding results that are competitive with the state-of-the-art in the computation of critical exponents in $d=3$, while also reproducing those from the $\varepsilon$-expansion for the Ising and other universality classes. In this work, we show that this is not a mere coincidence: by applying the scheme to the $O(N)$ universality class, we explicitly derive the flow equations and obtain critical exponents that are comparable to those obtained with higher-order non-perturbative approaches. In this case, FDR retains the features already highlighted in previous works -- namely, its efficiency and rapid convergence.

We study experimentally and theoretically the Stokesian settling of a well-known class of porous shapes: Bravais lattice unit-cells, whose porosity we vary controllably by changing their lattice spacing. In our experiments, conducted in a square cuboidal container with its long-axis aligned along gravity, we find that the settling speed U and the solid fraction φ of these lattice units obey a power-law relationship U $\propto$ φ^γ , with an exponent γ = 0.43 independent of their shape. To understand the observed scaling exponent, we analytically and numerically investigate the settling of the simple cubic structure under different approximations. We find that the walls of the container, though far from the sinking object, have a defining effect. Our Stokesian boundary integral simulations show that the Faxen's boundary correction captures the wall-effects accurately and enables us to discount the wall-effect from the experimental data, yielding a power-law exponent γ = 0.30 for settling in an unbounded domain. The power-law relating sinking speed and porosity is a step towards predictively understanding the sedimentation fluxes of complex objects in the clouds and the oceans. However, the applicability of this universal scaling to irregular and biologically richer aggregates found in nature remains an open direction.

Strain engineering is a powerful tool for tuning the electronic, magnetic, and topological properties of two-dimensional (2D) materials and thin films - particularly at high values of strain (>3%) where many electronic, magnetic, and structural transitions have been predicted. However, most approaches to tuning strain in 2D materials are limited below 1.5%, with poor repeatability when cycling strain and low strain transfer when cooling to cryogenic temperatures. Here, we report a high-yield sample preparation and device strain platform that overcomes these limitations, enabling precise, reversible strain tuning up to the intrinsic strain-to-failure of the materials tested herein. In addition, we show that this platform can be used to controllably design uniform linear strain gradients across of 10's of $μ$m, providing a novel route to systematically investigate flexoelectric and flexomagnetic phenomena. Using CrSBr as a model system, we demonstrate uniform uniaxial strain, up to ~4%, with negligible slippage and linear strain gradients of up to 0.06%/$μ$m. We further show that our strain approach is applicable to a broad class of 2D materials, validating its performance for three different phases of transition metal dichalcogenides: 2H-MoTe$_2$, 1T$^\prime$-MoTe$_2$ and T$_\mathrm{d}$-WTe$_2$. In T$_\mathrm{d}$-WTe$_2$, verified by theoretical calculations, we show a continuous redshift of the A$_1^3$ mode, up to a record-breaking ~5.5% strain, with a clear separation of the A$_1^3$ and A$_1^2$ modes starting at 2% strain.

DFT and AIMD are used to investigate the structural, stability, electronic, thermal, and optical properties of the quasi-2D C2N2O structure. The structure exhibits thermal and energy stability, signifying robustness under ambient conditions, however less dynamical stability is observed. The electronic structure investigation reveals that C2N2O displays semiconducting properties with a moderate indirect band gap resulting from the hybridisation of p-orbitals of N, C, and O atoms, with band gap values of 2.3 eV (GGA) and 3.9 eV (HSE06). The optical properties, including the dielectric function, optical conductivity, and refractive index, are thoroughly analyzed to clarify the electronic transitions. The material exhibits considerable optical absorption in the visible and ultraviolet spectrum, with notable anisotropy between in-plane and out-of-plane polarizations. Furthermore, plasmon resonance occurs at around 3.8 eV, relating to the collective oscillations of charge carriers. The thermal properties indicate a heat capacity of around 382 J/mol.K at 300 K, which is close to and slightly above the standard Dulong-Petit limit for this structure, indicating near-complete excitation of lattice vibrational modes at room temperature. The lattice thermal conductivity is extremely low, reaching approximately 0.017 W/m.K at 300 K, primarily attributed to significant phonon scattering, evidenced by a scattering rate of roughly 3.2 1/ps in the phonon frequency ranges. The findings demonstrate that the C2N2O structure maintains structural stability while allowing for tunable electronic, optical, and thermal properties, making it a promising candidate for nanoscale optoelectronic and thermal control applications.

We consider a many-body Hilbert space with a fixed global charge and show that the typical entanglement entropy of a subsystem, at the leading and subleading order in the thermodynamic limit, can be expressed in terms of a single quantity which represents the local thermal entropy at fixed charge density. We find a general formula which applies both to abelian U(1) symmetry and non-abelian SU(2) symmetry, including the case of a local Hilbert space which transforms under a general reducible representation of the symmetry group. We illustrate the general formula with model systems and discuss the relevance of the results as a probe of quantum chaos for physical Hamiltonians.

We study the nearest-neighbor spin-ice model subjected to a magnetic field applied along the global [111] and [110] directions, focusing on the role of sample geometry in stabilizing topological phase transitions. While no Kasteleyn transition is expected for this field orientations in the thermodynamic limit, we show that constraining the transverse dimensions of the system qualitatively changes the behavior. For samples elongated along the field direction with finite transverse area, the divergence-free constraint quantizes the number of string excitations that can span the system. As a result, the magnetization evolves through a cascade of discrete transitions corresponding to the successive entry of individual strings. Using Monte Carlo simulations, we demonstrate that each transition is marked by sharp magnetization steps and peaks in the specific heat and susceptibility, whose amplitudes scale linearly with the system length. We complement the numerical results with an analytical treatment based on the entropy - energy balance on a system with reduced dimensionality, deriving the critical fields associated with each topological sector. In the isotropic limit these transitions merge into a smooth crossover, but for anisotropic samples they remain sharply resolved, illustrating an unconventional mechanism by which finite geometry stabilizes topological phase transitions in frustrated magnets.

Diffusion with multipole-moment conservation gives rise to transport laws that generalize Fick's law and has attracted growing attention following experimental advances in strongly tilted optical lattices. It was recently shown that conserving complete multipole-moment groups leads to subdiffusive dynamics governed by a nonlinear diffusion equation, raising the question of whether hydrodynamic equations would also be nonlinear when the conservation law is imposed only at the subsystem level. Here we show that subsystem symmetries generically produce nonlinear hydrodynamic equations with shear-only transport, in which any localization present in the initial marginal distributions is preserved at long times by the conservation of those marginals. A linear regime emerges only as a limiting case for small fluctuations around a uniform background. We derive the deterministic and fluctuating parts of the hydrodynamic equations in arbitrary dimensions and obtain the corresponding maximum-entropy equilibrium distributions under constrained marginals. We also show that marginal-conserving diffusion provides a concrete hydrodynamic realization of partial multipole-moment conservation, and we offer an information-theoretic interpretation in which total correlation decays monotonically even when pairwise mutual information does not.

Thin ferroelectric BaTiO$_3$ films often exhibit continuous transitions instead of the first-order behavior of bulk crystals, a discrepancy usually attributed to epitaxial strain or dimensionality. Using quasi-adiabatic nanocalorimetry on freestanding BaTiO$_3$ membranes-free of clamping and substrate heat sinking-we show that domain morphology, not thickness or boundary conditions, controls the transition order. Thick membranes with large, monodomain-like regions display clear latent heat, whereas thinner membranes with dense 180$^{\circ}$ domain patterns show a continuous transition despite undergoing the same tetragonal-cubic structural change confirmed by x-ray diffraction. Piezoresponse force microscopy links this behavior to domain-size evolution, and a Ginzburg-Landau analysis demonstrates how reduced domain size lowers the free-energy barrier, rounding a nominally first-order instability. These results identify domain morphology as the key determinant of ferroelectric transition order in oxide membranes and establish design guidelines for enhancing caloric effects through domain engineering.

In Escherichia coli (E. coli), entropic repulsion between the two daughter DNA ring polymers under cylindrical confinement is believed to be an important factor governing chromosomal segregation. The repulsion can be enhanced by topological modifications, i.e., by the introduction of internal loops at certain locations along the contour of the circular DNA. However, the effect of topological modifications on the rate of segregation of ring polymers remains unclear. Therefore, we systematically varied the number and the contour length of loops introduced at selected locations by crosslinking monomers. The appropriate crosslinking was motivated by observations that extruded loops are located mainly near the origin of replication (ori-proximal) region of the E. coli chromosome. This resulted in the chains becoming intrinsically anisotropic. Using Langevin dynamics simulations of these topologically modified bead-spring polymers, we calculated the time required for segregation under cylinder confinement. With certain caveats, we found that increasing the number of loops resulted in a decrease in the time of segregation. In line with past work, we propose that this is due to the increase in the entropic repulsion between the polymers upon increasing the number of loops. In addition to the number of loops, the contour length of the loops and the mutual orientation of the (anisotropic) chains in the initial configurations played a role in determining the time of segregation.

We unveil that non-Abelian multigap band topology characterized by nontrivial Euler class invariants induces observable magnetononlinear Hall transport phenomena. We demonstrate these effects in a highly-tunable class of recently synthesized two-dimensional kagome N-heterocyclic carbene (NHC) metal-organic frameworks. We showcase the controllability of the nonlinear effect upon applying external voltage, changing temperature, and chemical substitutions that preserve the bulk topology and associated edge states. Our findings therefore reveal an uncharted presence of Euler class topology in metal-organic materials that can be experimentally deduced through measurable magnetotransport.

We investigate the effect on a Quantum Hall (QH) liquid of its coupling to 3+1 dimensional dynamical electromagnetism, which renders the system gapless. We calculate both the Hall and longitudinal resistances, $ρ_H$ and $ρ_L$, in the context of a minimal model of the electromagnetic environment, with a small three dimensional conductivity ${\tildeσ}$, that allows for a counter-flow current. In the thermodynamic limit, we show that $ρ_H$ is quantized, while $ρ_L$ approaches a non-zero limit, $ρ_L \sim α\, R_K$, where $α$ and $R_K=2π/e^2$ are the fine structure and the Klitzing constant. In contrast, the QH conductance, $σ_H$, is smaller than the expected quantized value by a correction $\sim α^2/R_K$. The electromagnetic interaction also generates corrections of order $α^2$ to the quasiparticle charges and statistics, in a way that is consistent with general arguments based on gauge invariance. In addition, we present an intuitive argument that relates the flux attachment associated with the composite boson representation of the electron liquid to the empirically observed %persistence of approximate quantization of $ρ_H$, even in circumstances in which $ρ_L$, and the deviation of $σ_H$ from its quantized value, are substantial.

A robust zero-bias conductance peak in putative $p$-wave superconductors is often regarded as the primary signature of a Majorana zero mode. Yet similar features can also arise from trivial bound states. This ambiguity has limited the reliability of conventional spectroscopy as a diagnostic tool, raising a long-standing problem of how to detect such impostors. Here, we address this issue with an alternative approach, atomic-scale shot-noise spectroscopy, that goes beyond conductance measurements. Through a detailed investigation of multiple defect-bound zero-bias states in the widely studied superconductor Fe(Se,Te), we observe that differential conductance can exhibit an apparently `robust' zero-bias peak. However, shot-noise measurements consistently reveal the fingerprint of the individual particle- and hole character hidden in the tunnelling conductance, unambiguously exposing the trivial nature of the zero-bias peak. Our results establish shot-noise spectroscopy as a decisive diagnostic for ruling out false Majorana signatures in atomic-scale experiments.

Disordered one-dimensional interacting systems have long been characterized via conventional correlation functions. A complementary quantum-information perspective quantifies the randomness of the unitary ensemble dynamics generated by a quantum system through the frame potential, which serves as a practical diagnostic for quantum algorithmic performance. However, no analytical treatment has yet been achieved for experimentally accessible interacting one-dimensional systems. In this Letter, we derive a closed-form expression for the frame potential of a Tomonaga-Luttinger liquid with quenched Gaussian forward-scattering disorder. Exploiting the exactly quadratic structure of the disorder-averaged Keldysh action, we show that the frame potential decays as a power law at early times and saturates to a late-time plateau controlled by a single coupling parameter. Taking the random field XXZ spin chain as a specific microscopic realization, we show that the strongest randomness is achieved near the Heisenberg ferromagnetic point and can be exponentially enhanced through a multiple-quench protocol. We validate our results across the entire gapless phase, with direct implications for algorithm design in analog quantum simulation platforms.

We provide a route to constructing an electronic quantum charge liquid (QCL), a state made up of fermions at fractional filling of a lattice that does not break translation. Starting with spinless fermions at filling $ν=3/2$ we pair them to get bosons at filling $ν=3/4$ per unit cell. The tetramer model, a generalization of the dimer model, on the square lattice is evaluated as a candidate bosonic QCL at filling $ν= 3/4$. It is shown that these models exhibit a local $\mathbb{Z}_4$ symmetry. Upon numerical study of a family of tetramer wavefunctions it is found that while one is gapless due to $\mathrm{U}(1)^3$ symmetry at least one other can be definitively shown to be gapped. The gapped nature of this state, along with its $\mathbb{Z}_4$ symmetry, leads us to propose that it is an example of the elusive bosonic QCL displaying the minimal $\mathbb{Z}_4$ topological order. We conclude by discussing possible extensions to other lattice geometries, electronic QCLs, and to Rydberg atoms.

The equation of state and, more generally, the thermodynamics of the Lennard-Jones fluid have long served as a benchmark problem in the statistical theory of fluids. Among available theoretical approaches, first-order perturbation theory occupies a special position: only at this level does the correction to the Helmholtz free energy admit an exact statistical-mechanical expression. In this work, we present a systematic, simulation-based assessment of a non-classical first-order perturbation theory in which the reference system incorporates the entire short-range part of the interaction, while the perturbation is confined to the remaining long-range tail. We show that this range-based decomposition transforms the perturbation contribution into a small, smoothly varying, near-mean-field quantity over a broad supercritical thermodynamic domain. When its density and temperature derivatives are consistently retained, the resulting equation of state reproduces high-accuracy reference data with excellent fidelity. The results demonstrate that the success of first-order perturbation theory is governed primarily by the physical content of the reference system and by the consistent treatment of its state dependence, rather than by the formal truncation order of the expansion.

Hybrid circuit quantum electrodynamics (cQED) aims at coupling various quantum degrees of freedom, among which are spin and charge degrees of freedom in gate defined quantum dots, phonons or magnons... with quantized electromagnetic fields in superconducting microwave cavities to investigate fundamental physics questions or for quantum computation and simulation. However, low microwave losses, key for many hybrid cQED experiments, are challenging to achieve given the often exotic and/or complex material stacks (e.g. semiconducting material, ferromagnets, or piezoelectric materials) required to host the various quantum degrees of freedom. In this work, we present a 3D-integration process to overcome this challenge for semi-industrial silicon MOS spin qubits. The process is based on dense indium bump interconnects at a pitch of 10 μm and superconducting thin films of Niobium Nitride (NbN). First, we report on DC and RF interconnect properties that demonstrate a high galvanic interconnection yield and internal quality factors above 105 in the single photon regime for NbN resonators interrupted by a single indium bump interconnect. Eventually, we fabricated a 3D-integrated hybrid circuit quantum electrodynamics (cQED) device based on a semi-industrial MOS hole double quantum dot and a high impedance NbN resonator. For this device, we report a cavity internal quality factor above 10000 and demonstrate record sensitivity for gate-based dispersive readout of the charge degree of freedom with an SNR of 100 in 300 ns. Finally, we demonstrate strong spin-photon coupling of gs/{2π} = 75 MHz, which highlights the viability of 3D-integration for quantum dot based hybrid spin circuit quantum electrodynamics and opens to high-fidelity spin readout and microwave photon-based remote spin qubit entanglement.

We study the thermodynamic phase diagram of a one-dimensional quantum spin chain subjected to both mean-field and nearest-neighbor interactions, and to a transverse magnetic field $h$. The purpose is to determine the effect of the quantum fluctuations, due to the transverse field, on the phase diagram, in particular with respect to the occurrence of ensemble inequivalence. We denote our model as a quantum Nagle-Kardar model. To perform the calculation of the canonical partition function, we show that, due to the presence of the mean-field term, in the thermodynamic limit one can use the Hubbard-Stratonovich transformation in spite of the non-commutativity of the different operators appearing in the Hamiltonian, and we adopt a procedure of successive approximations that lead to the determination of the phase diagram thanks to a scaling property of the phase transition lines. The results show that the ensemble inequivalence, present in the classical Nagle-Kardar model, is removed above a threshold value $h_c$ for the transverse field. For $h$ larger than $h_c$ the phase diagram exhibits only second-order phase transition lines, implying therefore restoration of ensemble equivalence.

Organic Electrochemical Transistors (OECTs) stand out for their interplay between ionic and electronic conduction, making them ideal analogues to biological synapses for neuromorphic computing and biosensing applications. Furthermore, they can be printed into integrated circuits on flexible substrates, enabling low-cost and high-throughput fabrication of complete electronic systems. However, most OECT electrolytes for integrated circuits still lack biocompatibility and suffer from rheology-related printing challenges. This paper presents a novel material basis that can be combined with an ionic liquid to fabricate an electrolyte for OECTs that only contains biocompatible materials. It allows rheological adjustments to enable the use of electrolyte in both inkjet and screen printing. Furthermore, the electrolyte is UV-curable, enabling it to transition into solid-state structures after printing. Extended ink and device lifetimes for screen-printed structures enable the fabrication of advanced OECTs that can operate in ambient air for over 30 days after fabrication. Ultimately, a fully screen-printed transistor using only biocompatible materials on a leaf substrate is shown

Although random matrix theory provides a fundamental framework for characterizing quantum chaos, encompassing both ergodic and localized phases, a comprehensive understanding of the universal features governing the critical transition remains elusive in many disordered and quasi-random systems. In this study, we explore the ergodic-to-many-body localization transition in the generalized Aubry-André model with interacting spinless fermions. Using the concept of Frobenius norm of an adiabatic gauge potential, we construct a phase diagram that captures the sensitivity of the eigenspectrum to infinitesimal adiabatic gauge deformations. To examine the stability of the critical disordered strength with respect to system size, we perform an unbiased finite-size scaling analysis via cost-function minimization techniques. Additionally, by analyzing the adjacent gap ratio and spectral form factor, we determine the scaling behavior of the Thouless time as a function of the disorder strength.

We theoretically and experimentally study finite-time optimal control of a colloidal particle steered through a spatially inhomogeneous environment, modeled by a position-dependent energetic cost at the final state. The competition between this state-dependent penalty and path-dependent dissipation gives rise to a sharp transition in the control strategy at a critical control duration. We further show that this transition can be linked to a dynamical phase transition in nonequilibrium relaxation after a quench, where the control cost maps onto the rate function governing rare trajectories.

We study conductance fluctuations in random resistor networks with hyperuniform bond disorder, where the fluctuations of the number of bonds present in a test volume $V$ scale as $V^{-a}$ with $a > 1/2$. Since small changes in the concentration of bonds present in a local region give rise to a proportionate increase in the locally averaged conductance, one may expect that in hyperuniform disorder, conductance fluctuations will also show suppressed fluctuations. We argue that this is not the case: conductance fluctuations scale as $L^{-d/2}$ for a sampling size $L$. We show numerical results for $d=2$.

Sensing the direction of arrival and polarization of impinging signals is a key prerequisite for beamforming and interference mitigation in modern wireless communication systems. Dynamic metasurface antennas (DMAs) can multiplex direction- and polarization-dependent field information onto a single detector by sequentially switching between programmable configurations. This makes DMAs attractive for joint direction-of-arrival and polarization (DoA-P) estimation with a single radio-frequency chain. Experimental demonstrations have so far relied on random pre-measured configuration sequences because optimizing the configurations requires an accurate forward model of the fabricated DMA. Here, we use an experimentally calibrated model based on multiport-network theory (MNT) to optimize DMA configuration sequences for DoA-P estimation. Our experimentally calibrated MNT model predicts the dual-polarized far-field response of our 96-element DMA for arbitrary admissible configurations, enabling model-based optimization without additional radiation-pattern measurements. We optimize sequences using effective-rank-based surrogate objectives and compare them with random sequences as a function of the sequence length and the noise level. The optimized sequences yield the largest gains in the intermediate-SNR and intermediate-sequence-length regime, where the inverse problem is neither noise-limited nor already solved by random diversity. We also tackle a dual-source scenario involving a jammer and a desired transmitter. Our results illustrate some of the potential in the context of jamming-resilient communications that is unlocked by experimentally calibrated MNT models for fabricated DMAs.

The growing number of wireless communication bands has driven demand for compact, low-loss, and frequency adjustable RF filtering. Tunable acoustic resonators are well suited to address these needs, offering a path toward reconfigurable front ends with reduced component count. In this work, we extend upon previous conference results to investigate epitaxial barium titanate (BTO) grown on silicon as a platform for tunable acoustic resonators. We demonstrate lateral excitation of symmetric Lamb (S0) modes in 120 nm X-cut BTO membranes using a multi-cell electrode architecture that simultaneously achieves high electromechanical coupling and practical impedance levels. Devices are fabricated with laterally patterned electrodes on released BTO membranes. Under applied DC bias, ferroelectric domains align, allowing electrical excitation, frequency tuning, and quality-factor enhancement of acoustic modes. The primary resonance near 700 MHz exhibits a Bode quality factor of 175, electromechanical coupling up to 25.1%, and series and parallel resonance tunability of 2.3% and 5.6%, respectively. Voltage-dependent material parameters, including permittivity, stiffness, and piezoelectric coefficients, are extracted through a combination of modified Butterworth-Van Dyke modeling and finite-element simulation to explain the observed trends. These results highlight monolithic BTO on silicon as a promising material system for laterally excited, tunable acoustic resonators for reconfigurable RF applications.

This paper provides a concise yet comprehensive review of recent advancements in millimeter-wave (mm-wave) oscillators below 100 GHz and sub-terahertz (sub-THz/THz) oscillators above 100 GHz for next-generation computing and communication systems, including 5G, 6G, and beyond. Various design approaches, including CMOS, SiGe, and III-V semiconductor technologies, are explored in terms of performance metrics such as phase noise, output power, efficiency, frequency tunability, and stability. The review highlights key challenges in achieving high-performance and reliable oscillator designs while discussing emerging techniques for performance enhancement. By evaluating recent design trends, this work aims to offer valuable insights and design guidelines that facilitate the development of robust mm-wave and sub-THz/THz oscillators for future communication, computing, and sensing applications.

Monitoring a 118 km subsea cable using Distributed acoustic, state-of-polarization, and Brillouin sensing captured storm-induced strain up to $\approx 0.003 με$ (dynamic) and $\approx 180 με$ (static), demonstrating consistent yet distinct modal responses to environmental loading.

The implementation of the \ac{isac} feature in \ac{6g} networks is most likely to be based on the framework of \ac{ofdm}. Input distribution design, or constellation design, is a crucial technique in \ac{ofdm}-\ac{isac} systems enabling a favorable balance between communication rate and sensing performance. In this treatise, we propose a computationally efficient input distribution design approach for \ac{ofdm}-\ac{isac} under frequency-selective channels, following the theoretical framework of capacity distortion. We highlight that under practical sensing constraints, the optimal strategy is to treat the kurtosis of constellations as a resource, and allocate it appropriately over subcarriers.

Time-of-arrival (TOA)-based localization in mixed line-of-sight (LOS) and non-line-of-sight (NLOS) environments is challenging because conventional Euclidean range models do not capture diffraction-dominated propagation. We show that the diffraction path-length model smoothly transitions between LOS and diffraction-dominated NLOS conditions, eliminating the need for explicit path classification. Although this model provides a unified geometric description of mixed LOS/NLOS propagation, the resulting 3D maximum-likelihood problem is nonconvex, and a direct Gauss--Newton estimator based on this model can converge to suboptimal local minima. This motivates the development of a class of structure-exploiting estimators. For known target height, the model induces a virtual-anchor representation of the reduced 2D problem, enabling estimators that exhibit a clear complexity--performance tradeoff: surrogate formulations provide structure and computational efficiency, while a semidefinite-relaxation formulation more faithfully preserves the original likelihood at higher cost. Building on this same structure, we develop 3D sample--polish--select estimators that reduce the global search to one dimension, solve the associated fixed-height 2D subproblems, and then apply local nonlinear refinement in 3D. The proposed estimators achieve near-Cramér--Rao lower bound (CRLB) performance with substantially lower complexity than multistart Gauss--Newton, while also being far more robust to initialization than a direct single-start Gauss--Newton estimator.

In the Internet-of-Things (IoT) era, efficient functionality integration is essential to address the growing demands of communication, computation, and sensing. Signal-level integrated sensing, computing, and communication (Sig-ISCC) is envisioned, where a single waveform simultaneously supports sensing, computing and communication via over-the-air computation (AirComp). Meanwhile, federated learning (FL) is widely regarded as a promising distributed machine learning framework that enables network intelligence in a privacy-preserving and secure manner, and exhibits strong synergy with AirComp, which alleviates the communication bottleneck of FL. In this paper, we study uplink Sig-ISCC design for AirComp-FL with joint target detection. We formulate the joint power and receive-scaling control problem, where edge devices' transmitted signals should serve both sensing and AirComp purposes. The goal is to minimize the AirComp aggregation distortion subject to a joint target-detection requirement. Although the resulting problem is non-convex in the original variables, we show that it admits an equivalent convex reformulation after a suitable variable transformation. By exploiting analytical optimality properties, we develop a robust, optimal, and polynomial-time-complexity algorithm that efficiently achieves the optimal transmit powers and receive scaling factor. Simulation results validate the optimality and numerical robustness of the proposed algorithm and show its superior FL performance compared to baseline methods.

In this paper, a novel pinching antenna-aided spatial multiplexing (PASM) architecture is conceived, which intrinsically amalgamates the benefits of flexible radiating element placement with radio-frequency (RF) chain transmission. Specifically, we leverage the deterministic phase variation along dielectric waveguides as a zero-power phase-control mechanism, where each waveguide fed by a single RF chain drives multiple pinching antennas (PAs) acquiring position-dependent phase shifts. Then, the PASM propagation environment is characterized by a realistic channel model encompassing Rician small-scale fading, correlated shadowing, and large-scale path loss. Based on this, a low-complexity vector approximate message passing (VAMP) detector is conceived, which exploits a waveguide-structured prior for jointly processing the signals associated with all PAs. Moreover, we derive an analytical upper bound on the bit error rate (BER) for the maximum likelihood (ML) detector to quantify the achievable performance limits. Finally, our simulation results demonstrate that the proposed PASM architecture achieves substantial signal-to-noise ratio (SNR) gain over the conventional phase-shifter-aided spatial multiplexing (PSSM), while the VAMP detector strikes an attractive trade-off between the system performance and computational complexity.

The average symbol error probability (SEP) of a phase-quantized single-input multiple-output system with M-ary phase-shift keying modulation and maximum ratio combining (MRC) is analyzed under correlated Rayleigh fading and additive white Gaussian noise. Building on our prior framework for the independent and identically distributed case, we extend the analysis to spatially correlated channels by introducing an asymptotically equivalent MRC combiner that enables tractable SEP characterization. Using this approach, we derive closed-form expressions at high signal-to-noise ratio (SNR) that explicitly characterize the diversity and coding gains as functions of the receive correlation structure, phase-quantization resolution, and modulation order, up to a scaling factor bounded between 1 and 2. The results show that channel correlation primarily degrades the coding gain, leading to an SNR penalty, while the diversity gain is preserved when the channel covariance matrix is full-rank. The analytical findings are validated through Monte Carlo simulations, demonstrating a tight match across a wide SNR range.

Integrated sensing and communication (ISAC) holds great promise in expanding the applications of wireless communication networks. However, in current communication-centric systems, the time-frequency resources available for sensing may be limited, and also usually non-uniformly and sparsely distributed across the time-frequency domain. Such a non-uniformity destroys the "thumbtack-shaped" ambiguity function of the orthogonal frequency division multiplexing (OFDM) waveform, leading to degraded sensing performance. To this end, this paper explores the parameter estimation algorithm for OFDM-ISAC systems with non-uniform sparse resource allocation. Specifically, for the single target case, we derive the closed-form Cramer-Rao lower bound (CRLB) for parameter estimation as a function of resource indices. Furthermore, we show that simply filling unused resource locations with zeros and applying the classic periodogram estimation is equivalent to maximum likelihood (ML) estimation, which is asymptotically optimal. For the multi-target case, we generate a virtual resource using the autocorrelation function of the original signal, which exhibits a significantly larger virtual bandwidth compared to the original signal, at the cost of higher peak-to-sidelobe ratio (PSLR). Simulation results demonstrate that the proposed approach outperforms the conventional periodogram method for non-uniform sparse resource allocation.

Cooperative integrated sensing and communication (ISAC) based on orthogonal frequency-division multiplexing (OFDM) enables network-wide sensing by exploiting the spatial diversity of multi-base-station (BS). This paper studies performance analysis and time-frequency resource allocation for a multi-BS cooperative OFDM-ISAC network with fine-grained resource-element (RE)-level orthogonal coordination. Two fusion architectures are considered: signal-level fusion (SLF), which forwards raw echoes to a fusion center, and parameter-level fusion (PLF), which reports only local delay/Doppler estimates and their uncertainty information. For SLF, we derive the Cramér--Rao bound (CRB) for joint target position and velocity estimation. For PLF, we develop a two-stage CRB-like metric by combining local delay/Doppler uncertainty characterization with first-order geometric error propagation, and show that only an oracle ML-based PLF benchmark can asymptotically attain the SLF CRB under restrictive conditions. Based on these results, we formulate a joint RE-selection and power-allocation problem under network-wide RE exclusivity, per-BS power budgets, a communication sum-rate constraint, and a sidelobe-amplitude constraint on the delay-Doppler ambiguity function. An efficient solution is developed via Schur-complement reformulations and penalty-based alternating optimization. Numerical results validate the analysis, demonstrate effective ambiguity-sidelobe suppression and consistent localization/velocity gains over representative baselines, while revealing geometry-dependent SLF-PLF performance gaps.

Microwave linear analog computing (MiLAC) has recently emerged as a promising architecture for analog-domain beamforming. In particular, a hybrid digital-MiLAC architecture was proposed and was shown to achieve fully-digital beamforming flexibility in narrowband systems when the number of RF chains equals the number of data streams. However, its performance in wideband systems remains unexplored. This paper presents the first study of hybrid digital-MiLAC beamforming for wideband multi-user multiple-input single-output (MU-MISO) systems. We first characterize the minimum number of radio-frequency (RF) chains required for hybrid digital-MiLAC beamforming to realize an arbitrary set of fully-digital beamforming matrices across all subcarriers. It turns out that, unlike in the narrowband case, a larger number of RF chains is generally required in frequency-selective channels to achieve fully-digital beamforming flexibility, which may be unfavorable in practice. To study the performance of hybrid digital-MiLAC beamforming with a limited number of RF chains, we then formulate the average sum-rate maximization problem and develop an efficient weighted minimum mean-square error (WMMSE)-based algorithm for beamforming design. Simulation results show that hybrid digital-MiLAC beamforming consistently outperforms conventional hybrid digital-analog beamforming, and achieves $89.93\%$ of the fully-digital sum-rate while using only $12.5\%$ of the RF chains in highly frequency-selective channels.

Visual data compression is shifting from human-centered reconstruction to machine-oriented representation coding. In this setting, an image is often mapped to a compact semantic embedding, which is then compressed and transmitted for downstream inference. We propose an adaptive transform-coding method for semantic-feature compression motivated by the conditional rate-distortion function of a Gaussian mixture model. The scheme uses mode-dependent transforms and quantizers selected according to the inferred source component, enabling more efficient coding of heterogeneous feature distributions. Evaluations on features from widely used vision backbones and foundation models show that the proposed method outperforms or is competitive with state-of-the-art neural compression methods while preserving flexibility and interpretability.

The CONCERTO millimeter-wave spectral-imaging instrument was deployed on the Atacama Pathfinder EXperiment (APEX), where it acquired science data between April 2021 and May 2023. The instrument features two focal-plane arrays, each composed of 2400 Microwave Kinetic Inductance Detectors (MKIDs). Each array is divided into six feedlines containing 400 MKIDs each, with each feedline read out by a dedicated FPGA-based board, KID_READOUT. The next-generation instrument aims to double the detector count per feedline, increasing it from 400 to 800 MKIDs. Achieving this requires a substantial scaling of the readout architecture and poses two key challenges for KID_READOUT: maintaining readout signal integrity and constraining firmware resource usage, as a direct upscaling of the existing design would exceed the available FPGA capacity. To overcome these limitations, we developed a Python-based, cycle-and bit-accurate digital twin of the full FPGA digital signal processing chain. This model enabled a detailed investigation of internal signal behavior and provided quantitative guidance for firmware optimization. Leveraging these insights, we identified the source of two spurs present in CONCERTO data and significantly reduced their amplitudes. At the same time, we achieved substantial reductions in firmware resource usage-39.0%pt in LUTs, 20.3%pt in flip-flops, and 28.98%pt in DSP slices-without degrading readout performance. The resulting architecture supports more than 800 MKIDs per feedline on the same hardware platform while preserving readout signal quality, offering a scalable and resource-efficient solution for future high-resolution millimeter-wave astronomical instruments.

Interference alignment (IA) is a widely recognized approach for mitigating inter-cell interference in multi-user multiple-input multiple-output (MIMO) networks. Despite its effectiveness, practical deployment remains constrained by two major challenges, i.e., the need for global channel state information (CSI) at each transmitter and the complexity of deriving closed-form solutions for intricate MIMO systems. This work aims to maximize network throughput by effectively mitigating interference using an IA-inspired learning algorithm that addresses its aforementioned challenges. First, we propose a predictive, transformer-based IA framework that estimates CSI to reduce signaling overhead in small-scale MIMO systems. Next, we formulate the IA problem as a multi-objective optimization problem based on subspace coordination and develop two reinforcement learning-based algorithms to enhance the scalability of IA in large-scale MIMO systems. Simulation results demonstrate that the proposed methods significantly outperform conventional baselines with up to 30% average user throughput gains over the best performing baseline.

Terahertz (THz) communication has emerged as a key enabler for sixth-generation (6G) networks, offering ultrawide bandwidths to support data-intensive applications such as holographic telepresence and immersive extended reality. Recent advances have enabled both electronics-based and photonics-based THz front-ends, each with distinct advantages and hardware limitations. While electronics-based solutions leverage mature semiconductor platforms, they suffer from amplified oscillator phase noise, frequency offsets, and nonlinearities introduced by multiplier and amplifier chains. Photonics-based systems, in turn, enable highly tunable and spectrally pure carriers but are subject to laser intensity noise, amplified spontaneous emission, shot noise in photomixers, and thermal noise in RF mixers. This article provides a comprehensive review of experimental demonstrations in electronics-, photonics-, and hybrid-based THz links, highlighting their hardware architectures, performance metrics, and implementation trade-offs. We then survey theoretical modeling efforts, emphasizing how hardware impairments affect system reliability and identifying limitations in existing studies. Building on this, we develop comprehensive signal models for both approaches, derive analytical expressions for signal-to-noise ratio (SNR), and evaluate bit error rate (BER) performance under realistic system parameters. Comparative results demonstrate how distinct impairment mechanisms shape the overall link performance of electronics- versus photonics-based THz systems. The insights offered aim to guide the design of robust transceiver architectures and accelerate the integration of THz technologies into future 6G deployments.

This work presents the design and analysis of a sixth-order triple-coupled transformer-based tank, enabling third-harmonic expansion for mmWave VCOs. Unlike conventional fourth-order tanks, the proposed tank inherently supports three resonance modes, enabling wideband third-harmonic expansion without additional low-Q switched-capacitor tuning elements. In contrast to conventional class-F23 designs, the proposed VCO removes the head resonator and adopts a noise circulating core to maintain low phase noise with reduced area. Implemented in TSMC 65-nm CMOS, post-layout simulation results demonstrate a 21.03-23.99 GHz (13.5%) tuning range, minimum phase noise of -116.25 dBc/Hz at 1 MHz offset, and peak FoM/FoMT/FoMA of 195.86/198.24/212.31 dBc/Hz while consuming 5.4 mW and occupying 0.02268 mm2.

As healthcare systems become increasingly interconnected, ensuring secure and continuous device authentication in health information exchange (HIE) networks is critical to safeguarding patient data and clinical operations. In this context, this paper proposes a novel cross-layer authentication scheme for HIE networks that integrates cryptographic mechanisms with physical (PHY) layer-based authentication to ensure reliable communication while minimizing computational and communication overheads. The initial authentication phase leverages a traditional public key infrastructure (PKI)-based approach, employing elliptic curve cryptography (ECC) and digital certificates to verify the legitimacy of communicating devices. Simultaneously, it extracts unique hardware-level features such as carrier frequency offset (CFO) and quadrature skewness from the devices. These features are then used to train a machine learning (ML) model during an offline phase managed by a regional centralized authority (RCA). For re-authentication, the system re-extracts these PHY-layer features from incoming orthogonal frequency division multiplexing (OFDM) symbols and verifies the device identity in real-time using the trained ML classifier. This cross-layer strategy enables continuous, lightweight identity verification without the need to exchange and validate cryptographic signatures for each message, thereby reducing system overhead. The proposed scheme further enhances privacy through the use of encrypted, frequently refreshed pseudo-identities, ensuring unlinkability and resistance to identity tracking. A formal security analysis using Burrows-Abadi-Needham (BAN) logic demonstrates the scheme's robustness against various threats, including impersonation, man-in-the-middle (MitM), replay, and Sybil attacks.

Micro DC brushed motors are widely deployed in battery-powered biomedical systems, where limited energy budgets and variable physiological loading impose stringent efficiency and safety constraints. However, conventional actuation strategies rely on conservative voltage margins to avoid stalling, leading to systematic energy inefficiency. Furthermore, existing methods primarily optimize steady-state performance, neglecting the energy required to complete individual actuation cycles under dynamic conditions. This paper reveals that the energy consumption per mechanical cycle of a DC motor exhibits a non-monotonic dependence on driving voltage, with a load-dependent minimum that shifts with external loading. Based on this insight, we propose a real-time operating-point tracking method that enables the motor to autonomously converge to its minimum-energy condition. A lightweight load metric derived from current waveform features is introduced to detect load variation, and a two-phase adaptive voltage strategy is developed to track the optimal operating point online. Experimental results demonstrate that the proposed method can track the new minimum-energy operating region under both low-to-high and high-to-low loading transitions. With 3-cycle averaging, the mean response time is 11.55s for the low-to-high case and 11.16s for the high-to-low case, while the mean convergence voltage is 2.73V and 2.0V, respectively.

The integration of multimodal sensing and millimeter-wave (mmWave) communications is a key enabler for highly mobile vehicle-to-infrastructure (V2I) networks. However, continuous high-resolution visual sensing incurs prohibitive computational energy, while delayed sensing information causes severe beam misalignment. This paper establishes a physics-aware multimodal integrated sensing and communication (M-ISAC) framework that mathematically bridges network-layer queuing delays with physical-layer spatial uncertainty via the semantic age of information (AoI). Guided by this relationship, we aim to strike an optimal trade-off between the tracking posterior Cramer-Rao bound (PCRB) and system energy budgets, we formulate a stochastic mixed-integer non-linear programming (MINLP) problem. Addressing the coupled challenges of temporal computing congestion and non-convex constant modulus constraints, we propose a reinforcement learning (RL) framework empowered by a Lyapunov-driven heterogeneous mixture-of-experts (LD-H-MoE) architecture. By strictly decoupling temporal scheduling and spatial phase mapping into specialized subnetworks, the LD-H-MoE circumvents gradient conflicts prevalent in monolithic multi-task learning. Simulations demonstrate that the proposed LD-H-MoE achieves a highly-effective event-triggered sensing policy, yielding superior tracking accuracy and radio-frequency (RF) resilience while guaranteeing edge computing queue stability and long-term energy budgets.

Movable antennas (MAs) have attracted growing interest for their ability to improve channel conditions via adaptive antenna movement. Nevertheless, such movement inevitably introduces mutual coupling (MC), whose impact has been largely overlooked in existing MA literature. In this paper, we show that MC is not merely an unavoidable electromagnetic effect, but also a new source of capacity gains in MA-enabled multiple-input multiple-output (MIMO) systems. To leverage MC effects, we develop an optimization framework for both narrowband and wideband systems based on a rigorous circuit-theoretic model. For narrowband systems, capacity maximization is formulated as a non-convex optimization problem, which is solved via a block coordinate ascent (BCA) framework. Because optimizing MA positions is challenging due to analytically intractable MC matrices, we develop a trust region method (TRM)-based algorithm that utilizes Sylvester equations to compute the derivatives of the inverse square roots of the MC matrices. We further consider wideband systems and formulate a sum-rate maximization problem. To find a unified set of MA positions that balances varying subcarrier conditions, the BCA framework and the TRM-based MA position optimization algorithm are extended to wideband systems. Simulation results demonstrate that exploiting MC effects in MA-MIMO systems yields significant performance gains in both narrowband and wideband systems under various channel conditions. These gains highlight the benefits of MC-induced superdirectivity and designable MC matrices.

Integrated sensing and communication (ISAC) is increasingly expected to operate under aggressive spectrum reuse, where co-channel orthogonal frequency division multiplexing (OFDM) interference can be catastrophic for data recovery on the time-frequency (TF) grid. We show that supporting blind ISAC is feasible by exploiting a fundamental asymmetry in the impact of co-channel OFDM interference: while communication is fragile on the TF grid, sensing depends on structured physical parameters whose signatures remain identifiable by relying on higher-order statistics. Based on this observation, we construct a fourth-order measurement tensor from the received OFDM signal whose coherent component preserves the delay-, Doppler-, and angle-dependent phase evolution of each source. We then develop a three-dimensional higher-order-statistics (HOS) based periodogram for iterative peak search and refinement to jointly estimate both range, velocity, and angle in the presence of unknown co-channel interferers. We further exploit constellation asymmetry to resolve the remaining phase ambiguities of blind recovery, enabling blind coherent demodulation via minimum constellation fitting. We also benchmark the performance through matched data-aided and stochastic Cramer-Rao lower bounds. We then quantify the cost of signal blindness. Simulations and experimental validations demonstrate reliable radar parameter estimation together with effective communication demodulation even when the TF-domain link is severely interfered with.

Designing Doppler-resilient unimodular discrete phase-coded waveforms (DPWs) with low delay-Doppler sidelobes is critical for multiple-input multiple-output (MIMO) radar. Existing block coordinate descent (BCD) methods suffer from high computational cost for designing long sequences or large waveform sets. Meanwhile, learning-based alternatives such as the soft-quantization network (SQN) only address correlation optimization in the delay domain, without considering ambiguity function (AF) optimization in the joint delay-Doppler domain. To address these issues, this paper proposes a novel Doppler-resilient DPW design framework, termed SQNGD, for joint transmit-receive optimization that simultaneously optimizes the auto-AF, cross-AF (CAF), and signal-to-noise ratio loss (SNRL) under unimodular constraints. To solve the multi-objective optimization problem (MOOP), a joint transmit-receive design and an alternating optimization strategy are developed. The transmit waveforms are optimized via soft-quantization-based differentiable parameterization, while the receive filters are updated by gradient descent (GD) with an energy constraint and SNRL penalty. An FFT-accelerated evaluation of the AF and CAF is further incorporated, reducing the optimization time by 1.9x - 11x compared with the state-of-the-art (SOTA) majorization-minimization-coordinate descent (MMCD) method. Numerical results show that SQNGD achieves a peak sidelobe level (PSL) of approximately -43 dB over the Doppler range [-0.5,0.5] and -31 dB over [-600,600], respectively, outperforming MMCD by 5.85 dB and 3.45 dB, while maintaining the same SNRL of 0.5 dB.

Mobile robots that move between outdoor and indoor environments still struggle with consistent positioning. Satellite-based and terrestrial ranging each work well in their home domains, but combining them at the raw measurement level has received little attention, and the building boundary is precisely where both classes degrade. This paper reports preliminary observations from the HYMN dataset, which time-synchronizes raw measurements from GNSS, Ultra-Wideband (UWB), WiFi Fine Time Measurement (FTM), and Bluetooth Low Energy (BLE) against millimeter-level ground truth in an industrial setting. Per-zone measurement availability and ranging-residual behavior are characterised. The two technology classes turn out to be complementary, and the indoor-outdoor transition is where their weaknesses overlap. The dataset is publicly available.

Metasurfaces' ability to control electromagnetic wave propagation has led to a rapid paradigm shift in wireless operation. These metasurfaces are often called reconfigurable intelligent surfaces (RISs) due to active tuning elements distributed across the meta-atoms comprising the metasurface array. However, each of these dynamic meta-atoms requires additional DC power lines and biasing circuitry for active tuning. Additionally, achieving polarization diverse operations using compact metasurface configurations is challenging due to the complexity involved in polarization detection. To address these limitations, we propose a passive metasurface array architecture that is both polarization sensitive and capable of altering radiation patterns with frequency diversity. In particular, we designed a polarization-sensitive meta-atom model with added randomness in the scattering behavior and extended it to a polarization-diverse-frequency-selective array. By capturing the electric fields scattered off from the metasurface, we can numerically acquire the polarization information of the incoming signal. The proposed polarization-diverse array can simplify the polarization measurement techniques and may find its application in polarization sensitive sensing and imaging operations.

In this paper, a broadband 1-bit coding metasurface-based reconfigurable intelligent surface (RIS) is presented. The unit cell of the metasurface consists of a wide dipole modified with interdigital capacitors and loaded with an SMP 1340-040LF PIN diode. The proposed element offers cell miniaturization and a stable angular response. A phase difference of 180$\degree \pm$ 30$\degree$ is achieved for a frequency range of 4.85-6.05 GHz between the ON and OFF states for the normal incidence of the TE polarized wave, whereas it provides a fairly stable response with reflection loss of less than 3 dB and phase difference of 180$\degree$ $\pm$ 50$\degree$ for oblique incidence up to 45$\degree$. The RF is isolated from the DC on the bias lines using properly designed butterfly-shaped radial stubs. Using this unit cell, a prototype with an array of 16 $\times$ 10 elements is constructed. A low-cost microcontroller-based control circuit is designed, which can be plugged-in for biasing the PIN diodes of such array. The theoretically calculated and full-wave simulated radiation patterns of the array are validated using experiments inside anechoic chamber. Furthermore, the capability of the RIS for non-line of sight (NLOS) user equipment (UE) localization and robust uplink communication is demonstrated using LTE communication framework. This shows great potential of our RIS for applications, such as in unmanned aerial vehicle (UAV) localization and its uplink communication at NLOS or extended range.

Automatic accent identification (AID) remains a challenging task due to the complex variability of accents, the entanglement of accent cues with speaker traits, and the scarcity of reliable accentlabelled data. To address these challenges, we propose a speaker augmentation strategy using voice conversion (VC), with which we generate additional training data by converting original training utterances into different speaker voices while preserving accentual cues. For this purpose, we select two recent VC systems and evaluate their capability to preserve accent. Alternatively, we also explore the use of non-timbral embeddings in AID, for their ability to convey accent information among other non timbral cues. The effectiveness of both methods is demonstrated on the GenAID benchmark, achieving a new state-of-the-art F1-score of 0.66, compared to the previous score of 0.55. Beyond AID, we show that non-timbral embeddings enable accent-controlled Text-to-Speech, producing high-fidelity speech with accurate accent transfer.

This letter proposes a novel mathematical framework for the statistical characterization of reconfigurable intelligent surface (RIS)-mounted high-altitude platform station (HAPS)-assisted MIMO systems over cascaded Rician fading channels. Due to the inherent coupling introduced by the RIS, the resulting cascaded channel does not satisfy the independence assumptions required for conventional Wishart-based modeling, which motivates a tractable alternative approach. By adopting a line-of-sight (LoS)-aligned precoding strategy, the received signal-to-noise ratio (SNR) is represented as a non-central quadratic form with a structured covariance matrix. Exploiting this structure, a saddle point approximation (SPA)-based framework is developed to characterize the SNR distribution. Closed-form expressions for the probability density function (PDF), cumulative distribution function (CDF), and outage probability are derived. The proposed framework further incorporates practical RIS hardware impairments, including discrete phase shifts and phase-dependent amplitude responses. The accuracy of the proposed analysis is validated through Monte Carlo simulations.

Under-resourced languages remain underrepresented in quantitative rhythm research,particularly in systematic intra-branch analysis of acoustic differentiation within closely related linguistic groups.This study investigates acoustic differentiation within the Tani language subgroup by examining speech rhythm in Nyishi and Adi,two under-resourced Tani languages spoken in Arunachal Pradesh,North-East India,using a frequency domain framework based on amplitude modulation(AM) low-frequency(LF) spectrum analysis,commonly referred to as rhythm formant analysis(RFA).The analysis is designed to identify whether intra-branch differentiation follows a hierarchical pattern across rhythmic and spectral domains.From the LF modulation spectrum,three rhythm formant features were derived:Number of Dominant peaks(NDP),Mean Frequency of Dominant Peaks(MFDP),and Variance of Dominant Frequencies(VFDP).In addition,Discrete Cosine Transform (DCT)coefficients and Mel Frequency Cepstral Coefficient(MFCC) were extracted to characterise the spectral modulation structure and broad spectral organisation of the speech signal.Statistical modelling reveals a hierarchical pattern of differentiation,where rhythmic features show consistent but moderate separation,with Nyishi exhibiting higher dominant modulation frequencies as well as greater dispersion than Adi.Classification experiments further support this hierarchy,with rhythm-only features achieved approximately 84-85% classification accuracy.Fusion using MFCC representations improved performance to 90.9% classification accuracy using support vector machine (SVM) and 93.96% using multilayer perceptron (MLP).These findings demonstrate that rhythmic and spectral features encode complementary levels of linguistic variations,with low frequency modulation capturing constrained macro temporal structure and spectral features reflecting finer phonological differentiation.

As a pivotal usage scenario for 6G networks, integrated sensing and communication (ISAC) has emerged as a focal point of both academic and industrial research. To accommodate the heterogeneous connectivity requirements of future networks while jointly enhancing both the sensing and communication performance, this paper integrates the multi-static ISAC architecture with double-side fluid antenna system (DS-FAS) to fully exploit the available spatial degrees-of-freedom. Specifically, we establish a joint optimization framework for FA positions and transmit beamforming to maximize the target detection probability while satisfying the communication quality-of-service requirements. Recognizing the intricate coupling between the double-side FA positions and transmit beamforming, instead of trying to obtain an initial feasible point, we resort to the penalty-based mechanism to ensure the robustness against initial feasibility without introducing additional non-convexity. An alternating optimization-based algorithm is proposed to solve the decoupled subproblems. Specifically, the transmit beamforming is globally optimized via the semidefinite relaxation technique, while the transmit FA positions are determined using the majorization-minimization method. Finally, leveraging the analyzed FA mechanism, the feasibility subproblem for receive FA positions is transformed into a signal-to-interference-plus-noise ratio maximization one, solved efficiently via a gradient ascent-based approach, which yields superior performance over the feasibility-based benchmark with reduced complexity. Numerical results demonstrate the superiority of the considered DS-FAS-assisted multi-static ISAC systems in both noise-limited and interference-limited scenarios, while key insights for practical deployment are further extracted from the simulation analysis.

Low Earth orbit (LEO) satellite relays will significantly extend the coverage of mobile networks, enabling users in remote areas to transmit data of real-time events. Nevertheless, the limited power of user devices and the long distance to satellites lead to low signal-to-noise ratio (SNR), which results in high error rates and frequent retransmissions, severely hindering the transmissions of high-dimensional data such as videos. In this paper, we propose a novel method to achieve high error tolerance in satellite-relay video transmissions using generative semantic communications (GSC). For the transmitter, we design and optimize a semantic encoder integrating a pre-trained video encoder with a low-density parity-check (LDPC) encoder, efficiently achieving generalizability and enabling forward error correction. For the receiver, we fine-tune a generative video model using an efficient in-context adaptation algorithm, enabling it to reconstruct videos from error-corrupted semantic information. Simulation results show that our method achieves 2.5 dB higher video peak SNR than conventional semantic communications at an error rate of 45%, and remains robust when the error rate exceeds 80%.

Mathematical analysis has long underpinned wireless communication theory, yet the growing complexity of next-generation systems demands increasingly sophisticated reasoning from domain experts. Recent advances in AI mathematical reasoning, from formal theorem proving to large language model (LLM)-based derivation, offer a promising but largely unexplored path forward. Here we argue that wireless communications is a uniquely structured domain for formal AI reasoning, and propose a three-layer framework of verification, derivation, and discovery to rethink how wireless mathematical knowledge is established.

Weak constitutive fluctuations in dispersive subsurface media can induce distributed clutter that reshapes the observation structure of ground-penetrating radar (GPR). This paper analyzes this effect for single-snapshot frequency-diverse array multiple-input multiple-output GPR. Focusing on medium-induced clutter, rather than on general target--clutter joint modeling, it establishes a statistical propagation chain from Cole--Cole parameter perturbations to electromagnetic contrast, first-order Born channel snapshots, clutter covariance, and subspace descriptors. A medium-aware snapshot model and a covariance propagation framework are then derived to characterize how constitutive uncertainty alters observation-domain spectral structure under a local weak-fluctuation regime. Numerical experiments verify the consistency of the proposed propagation relation under the adopted first-order Born and constitutive-linearization approximations. Within the tested setting, medium-induced clutter reshapes the eigenspectrum and changes target--clutter overlap metrics. Spatial correlation length and background-scene variation act as consistently strong structural drivers, while the FDA frequency increment also produces measurable changes in the normalized covariance geometry.

Pinching-antenna systems (PASS) have recently attracted significant attention as a promising architecture for flexible and reconfigurable wireless communications. Despite notable advancements, research on energy efficiency (EE) maximization for PASS is limited as existing studies mainly focus on transmit power minimization or utilizing a simple power consumption model. This paper evaluates the impact of pinching antenna (PA) activation power on EE maximization in a downlink NOMA-assisted PASS by jointly optimizing PA activation and user power allocation under quality-of-service and transmit power constraints. To tackle the resulting mixed-integer nonlinear programming problem, we develop a two-layer iterative algorithm, where the outer layer performs matching-based PA selection and the inner layer computes a closed-form optimal power allocation solution. Numerical results demonstrate that the proposed solution achieves substantial EE gains over conventional fixed antennas systems and the considered benchmark schemes, approaches the exhaustive-search upper bound with significantly reduced complexity, while exhibiting fast convergence. It also demonstrates the significance of accounting for PA activation power in EE maximization problem.

Injecting artificial noise (AN) along the tangent space of a curved constellation makes each transmitted symbol induce a Gaussian observation with a symbol-dependent rank-one covariance, so the matched maximum-likelihood (ML) decoder differs from the Euclidean nearest-neighbor decoder by a single rank-one correction per candidate. We develop a baseband-demapper realization of this correction for the Fourier-curve constellation and instantiate a regular $(3,6)$ low-density parity-check (LDPC)-coded link at $(k,M){=}(20,64)$. Against four baselines (Euclidean-mismatched, flat-constellation isotropic-AN, no-AN, and same-spectral-efficiency narrowband), the matched decoder extends the BLER${=}10^{-1}$ operating range by approximately $5$\,dB over the Euclidean-mismatched counterpart on the same tangent-AN transmitter, at a cost of $2kM$ additional multiply-accumulate operations per symbol ($+50\%/+100\%$ under residual/template-correlation accounting) and a $20$\,KB constellation--tangent lookup table ($10$\,KB incremental over a Euclidean template-only LUT). A bit-interleaved coded-modulation achievable-rate (BICM-AIR) computation supports the same matched-metric advantage at the tested labeling and max-log demapper, indicating that the BLER gain is not merely an artifact of this particular LDPC simulation, and a Woodbury extension generalizes the rank-one correction to per-tone Ricean fading. In the tested Monte-Carlo runs, a design-aware bounded-search eavesdropper without the phase-key shows no successful LDPC decoding at any tested $k\in\{2,8,20\}$ within a $B{=}10^{3}$ non-code-aided search budget; code-aided, multi-frame, and known-preamble attacks are left to follow-up work. LUT quantization down to $6$ bits yields no measurable coded-BLER degradation at the tested operating points.

This paper investigates a joint hybrid digital-analog beamforming and trajectory design for a cell-free multi-static integrated sensing and communication (ISAC) system supported by multiple unmanned aerial vehicles (UAVs). Specifically, these UAVs cooperatively serve ground users and perform multi-static sensing to detect the target. We formulate a weighted sum-rate (WSR) maximization problem by jointly optimizing the hybrid beamformers and the UAV trajectories. This joint design explicitly accounts for practical constraints, including transmit power budgets, sensing signal-to-noise ratio (SNR) requirements, UAV kinematic constraints, and both continuous and discrete phase shifters. In particular, we reformulate the original complex problem into a solvable form that can be addressed using the penalty dual decomposition (PDD) method. Simulation results demonstrate that the proposed design achieves performance close to that of the fully digital (FD) scheme and significantly outperforms other schemes. Furthermore, leveraging UAV mobility and multi-static cooperation provides crucial spatial degrees of freedom, effectively avoiding WSR degradation under limited transmit power or strict sensing requirements.

Direct Device-to-Satellite (D2S) communications, which enable direct satellite connectivity with unmodified user equipment (UE), not only expand global coverage but also reshape the evolution of future access networks. However, D2S links face fundamental challenges due to inherently low signal-to-noise ratios (SNRs) and limited spatial multiplexing gains arising from near line-of-sight propagation, both of which severely constrain achievable spectral efficiency. Despite the lack of spatial multiplexing, this work shows that aggressive non-orthogonal transmissions, where multiple users (e.g., four) transmit concurrently over the same frequency resources, even to a single receive antenna, can unlock substantial capacity gains that remain entirely unexploited by existing systems. Realizing these gains in practice, however, requires receiver architectures that, to the best of our knowledge, have not yet been developed. To this end, we introduce NL-COMM-Sat, an efficient and flexible framework that overcomes this limitation by enabling aggressive non-orthogonal signal transmissions. In contrast to conventional non-orthogonal multiple access (NOMA) schemes, NL-COMM-Sat supports more than two UEs per receive antenna on the same frequency resource. The framework revisits optimal receiver design principles and proposes computationally efficient processing schemes that translate previously unexplored theoretical gains into tangible throughput improvements, even under realistic channel estimation errors and high-mobility Doppler conditions. Our evaluation shows that NL-COMM-Sat achieves up to a 2x increase in spectral efficiency compared to orthogonal multiple access and NOMA baselines across all considered SNR and Doppler regimes, even with a single-antenna receiver and user speeds of up to 500 km/h.

CONCERTO is a millimeter-wave imaging instrument that operated on the Atacama Pathfinder Experiment (APEX) telescope from April 2021 to May 2023. Its primary scientific objectives include the study of galaxy clusters through the Sunyaev-Zel'dovich (SZ) effects, the observation of Galactic star-forming regions, and the first measurements constraining the power spectrum of dusty star-forming galaxies. The instrument consists of two detector arrays, each comprising 2400 Microwave Kinetic Inductance Detectors (MKIDs). Each of the two arrays comprises six feed-lines and is read out by six KID_READOUT electronic boards, each capable of reading out one feed-line coupled to 400 frequency-multiplexed MKIDs. As the demand for higher-resolution millimeter-wave imaging continues to grow, future instruments aim to significantly increase the pixel count, with more than 800 detectors per feed-line. However, the MKID readout electronics chain is inherently complex, making it difficult to fully understand its performance limits and optimization margins. To address this challenge, we initiated a modeling effort that first focused on the digital section of KID\_READOUT. In this phase, we developed a digital twin of the FPGA-based signal processing chain, which led to substantial performance improvements. The present paper extends this modeling strategy to the analog readout chain. It presents the characterization and behavioral modeling of all analog components, allowing us to identify the elements that limit the frequency multiplexing factor, determine the dominant noise contributors, and highlight areas for improvement in signal conditioning for the future-generation board. Together, these developments establish, to the best of our knowledge, the first consolidated digital-and-analog behavioral framework of an MKID readout architecture, implemented in a unified Python-based modeling environment.

We introduce a quantum algorithm for simulating the dynamics of electrical circuits consisting of resistors, inductors and capacitors (aka RLC circuits) along with power sources. Given oracle access to the connectivity of the circuit and values of the electrical elements, our algorithm prepares a quantum state that encodes voltages and current values either at a specified time or the history of their evolution over a time-interval. For an RLC circuit with $N$ components, our algorithm runs in time $\textsf{polylog}(N)$ under mild assumptions on the connectivity of the circuit and values of its components. This provides an exponential speed-up over classical algorithms that take $\textsf{poly}(N)$ time in the worst-case. Our algorithm can be used to estimate energy across a set of components or dissipated power in $\textsf{polylog}(N)$ time, a problem that we prove is BQP-hard and therefore unlikely to be efficiently solved by classical algorithms. The main challenge in simulating the dynamics of RLC circuits is that they are governed by differential-algebraic equations (DAEs), a coupled system of differential equations with hidden algebraic constraints. Consequentially, existing quantum algorithms for ordinary differential equations cannot be directly utilized. We therefore develop a quantum DAE solver for simulating the time-evolution of linear DAEs. For RLC circuits, we employ modified nodal analysis to create a system of DAEs compatible with our quantum algorithm. We establish BQP-hardness by demonstrating that any network of classical harmonic oscillators, for which an energy-estimation problem is known to be BQP-hard, is a special case of an LC circuit. Our work gives theoretical evidence of quantum advantage in simulating RLC circuits and we expect that our quantum DAE solver will find broader use in the simulation of dynamical systems.

We develop the Schwinger-Keldysh path-integral formalism for open non-Abelian gauge theories that are gauge-fixed via the BRST method in covariant gauges. We focus on generic initial states, pure and mixed, specified at finite times suitable for non-equilibrium processes. We pay particular attention to the handling of the indefinite Hilbert space, the construction of BRST-invariant Schrodinger picture wavefunctionals, density matrices and inner product, the implementation of the Hata-Kugo prescription, and the role of boundary terms at both the initial and final times. We highlight the advantages of the Nakanishi-Lautrup field representation in dealing with initial/final conditions. The resulting Schwinger-Keldysh path integral is manifestly invariant under a diagonal (retarded) BRST symmetry for arbitrary physical initial states, whether pure or mixed. From this, we obtain the corresponding Ward-Takahashi-Slavnov-Taylor identities, valid perturbatively. Non-perturbatively the Gribov ambiguity is expected to break or modify the BRST symmetry. The naive advanced BRST symmetry is shown to be explicitly violated by the in-in boundary conditions. We show that the Feynman-Vernon influence functional derived by integrating out charged matter and/or hard gluon modes remains (perturbatively) BRST invariant. When the Open EFT action is expanded to second order in advanced fields it exhibits an exact symmetry under a contraction of the original BRST symmetry. This Keldysh BRST symmetry is equivalent to the BRST associated with the retarded gauge transformations together with a linearly realized BRST transformation of the advanced fields. These govern the structure of the leading terms in an Open EFT. We illustrate this with the explicit example of Hard Thermal Loop Effective Theory, and construct the general form of the Open EFT in a Higgs phase when all gauge symmetries are spontaneously broken.

This work investigates which sets of quantum states give rise to the highest achievable success probability in minimum-error state discrimination if multiple copies of the unknown state are given. Specifically, we consider uniformly distributed ensembles of the form $\left\{\frac{1}{N},ρ_i^{\otimes k}\right\}_{i=1}^N$, where $N$ states in dimension $d$ are provided in $k$ identical copies, and derive universal limits in this scenario. For pure state ensembles, we prove that whenever $N$ is large enough to support a state $k$-design, these designs will exactly give rise to the maximally discriminable sets. We further show that when $N$ exceeds the size required for a $k$-design, mixed states can outperform all pure state ensembles. We also analyse the analogue classical discrimination problems, in which states are replaced by probability distributions. We recognise that the problem of most discriminable classical states in the multi-copy regime is in one-to-one correspondence to the concept of the multiplicative Bayes capacity of independent uses of classical channels, a concept that emerges naturally in the context of classical information leakage. This connection allows us to completely solve the classical analogue of our problem when $N\geq \binom{d + k - 1}{k}$, and to prove that quantum systems offer a quadratic advantage (in number of copies $k$) over classical ones. Curiously, we also show that this quantum advantage is strongly reduced when one is restricted to real quantum states. Finally, we introduce computational techniques to find sets of most discriminable ensembles, and to obtain rigorous universal upper bounds on the maximal success probability for multi-copy state discrimination in cases that are analytically intractable.

We study the power of quantum witnesses under perfect completeness. We construct a classical oracle relative to which a language lies in $\mathsf{QMA}_1$ but not in $\mathsf{QCMA}$ when the $\mathsf{QCMA}$ verifier is only allowed polynomially many adaptive rounds and exponentially many parallel queries per round. Additionally, we derandomize the permutation-oracle separation of Fefferman and Kimmel, obtaining an in-place oracle separation between $\mathsf{QMA}_1$ and $\mathsf{QCMA}$. Furthermore, we focus on $\mathsf{QCMA}$ and $\mathsf{QMA}$ with an exponentially small gap, where we show a separation assuming the gap is fixed, but not when it may be arbitrarily small. Finally, we derive consequences for approximate ground-state preparation from sparse Hamiltonian oracle access, including a bounded-adaptivity frustration-free variant.

The widespread claim that violations of Bell inequalities establish the nonlocality of nature is critically reexamined. It is argued that this conclusion is not logically compelled by either the Einstein--Podolsky--Rosen (EPR) argument or Bell's theorem. The analysis highlights the central role of counterfactual reasoning -- the assumption that outcomes of unperformed measurements possess definite values -- in deriving Bell inequalities. It is shown that these inequalities follow not from locality alone, but from the conjunction of locality with a global assignment of values across incompatible measurement contexts. Their experimental violation therefore signals the impossibility of such a global assignment, i.e.\ contextuality, rather than necessarily implying nonlocal causation. This interpretation aligns with Bohr's emphasis on the contextual character of physical quantities and is naturally formulated within modern sheaf-theoretic approaches to contextuality.

Knots and links represent a fundamental motif of non-local connectivity that permeates the physical sciences from string theory to protein folds. While spectral braiding has been explored in two-band non-Hermitian models across various platforms, its direct simulation and characterization on programmable quantum hardware, particularly beyond two strands, remains a formidable challenge due to the limitations of variational optimization in these systems. Here, we introduce a family of non-Hermitian multi-band twister models and implement a non-variational protocol to characterize their complex braided band structures on a programmable superconducting quantum processor. By mapping the winding of eigenstates to the spectral topology, we devise an efficient measurement strategy that extracts braid information, including braid words and knot invariants like the Alexander and Jones polynomials, without requiring full spectral tomography or repeated optimization. We experimentally demonstrate the reconstruction of complicated knots and links such as the Hopf chain and Solomon's knot. Our approach provides a general framework for investigating exotic non-Hermitian topology on near-term quantum devices, opening a route to simulate more sophisticated topological structures in knot theory.

In this work, we investigate gravitational-wave signatures of a proposed subsolar-mass merger scenario resulting from fragmentation inside a collapsar accretion disk. This scenario has gained recent interest with the electromagnetic transient AT2025ulz, a possible superkilonova counterpart candidate to the sub-threshold gravitational wave event S250818k. One prediction of fragmentation is the formation of multiple smaller neutron-star fragments, some of which might merge hierarchically. Such mergers are expected not only to produce individual electromagnetic counterparts, but also, because of their repeated capture and merger dynamics, to impart kicks to the system and thereby drive orbital eccentricity. By performing numerical relativity simulations of hierarchical compact object mergers modeled as black holes in a disk-like geometry consistent with this scenario, we demonstrate the build-up of potentially large eccentricity for the final merger, of order $e \simeq 0.6$ initially, and show that, because of the short lifetime of the system, a substantial part of this eccentricity , up to $e\simeq 0.1$, can survive until merger in the general case. As a result, future detections of eccentricities in potential subsolar-mass gravitational-wave candidate events would be a strong indicator for a hierarchical formation scenario.

Long-range interactions between emitters give rise to collective phenomena, including superradiance, spin squeezing, and coherence protection, that are important to both fundamental physics and quantum technologies. Despite progress in cold atoms, coherent cavity-mediated all-to-all interactions have not yet been realized in a solid-state ensemble. Here we demonstrate such interactions in a $^{171}$Yb$^{3+}$:CaWO$_4$ crystal coupled to a microwave resonator, observing superradiant emission on resonance and unitary one-axis twisting dynamics in the dispersive regime. The same interaction also opens a many-body energy gap that suppresses inhomogeneous dephasing, extending the ensemble Ramsey coherence time from tens of microseconds to milliseconds without decoupling pulses. These results establish a solid-state platform for collective many-body physics with direct implications for quantum technologies. Specifically, the observed one-axis twisting dynamics opens a path towards spin squeezing for entanglement-enhanced quantum metrology, and the extended coherence due to gap-protection is relevant for both microwave photon storage and precision measurement.

We consider the problem of quantum channel certification to unitary, where one is given access to an unknown $d$-dimensional channel $\mathcal{E}$, and wants to test whether $\mathcal{E}$ is equal to a target unitary channel or is $\varepsilon$-far from it in the diamond norm. We present optimal quantum algorithms for this problem, settling the query complexities in three access models with increasing power. Specifically, we show that: (i) $Θ(d/\varepsilon^2)$ queries suffice for incoherent access model, matching the lower bound due to Fawzi, Flammarion, Garivier, and Oufkir (COLT 2023). (ii) $Θ(d/\varepsilon)$ queries suffice for coherent access model, matching the lower bound due to Regev and Schiff (ICALP 2008). (iii) $Θ(\sqrt{d}/\varepsilon)$ queries suffice for source-code access model, matching the lower bound due to Jeon and Oh (npj Quantum Inf. 2026). This demonstrates a strict hierarchy of complexities for quantum channel certification to unitary across various access models.

The pure-loss channel is a fundamental model for describing noise in bosonic quantum platforms. It is characterised by a single parameter, the transmissivity, which quantifies the fraction of the input energy that reaches the output of the channel. In realistic scenarios, however, such as free-space quantum communication, the transmissivity is not fixed but fluctuates from one channel use to another. In this setting, the overall channel is effectively described as a convex combination of pure-loss channels, known as a fading channel. Despite its practical relevance, the quantum Shannon theory of the fading channel has remained largely unexplored. Here, we address this gap, specifically investigating degradability, anti-degradability, entanglement breakingness, and capacities of the fading channel. Of particular relevance to practical quantum-internet applications, we prove that entanglement distribution and quantum key distribution can always be achieved at a strictly positive rate over any fading channel, no matter how noisy it is or how strongly the transmissivity fluctuates, provided the channel is not completely noisy. Moreover, we prove that thermal states, which are optimal for a broad class of static bosonic Gaussian channels, fail to achieve the entanglement-assisted classical capacity of fading channels: non-Gaussian Fock-diagonal states strictly outperform all Gaussian encodings. Most strikingly, we identify regimes where the coherent information of thermal inputs vanishes, while optimized non-Gaussian states achieve strictly positive values, thereby activating the channel for quantum communication. For a paradigmatic binary fading model we establish this result analytically, deriving the exact capacity-achieving state in closed form. For general fading distributions, we design an iterative variational algorithm to optimize the coherent and mutual information.

Simulating quantum dynamics is one of the central applications of quantum computing. For Hamiltonians written as a sum of many terms, deterministic Trotter--Suzuki product formulas can require applying a large number of term-wise evolutions at each time step, leading to high circuit costs for large or dense systems. Randomized methods such as qDRIFT offer an alternative: each step samples only one Hamiltonian term, giving a circuit depth with no explicit dependence on the number of terms. However, when qDRIFT is used for observable estimation, high precision requires many independent random circuit realizations, resulting in a total gate complexity that scales as $\mathcal{O}(\varepsilon^{-3})$. We introduce a multilevel Monte Carlo framework for qDRIFT that reduces this sampling overhead. The method constructs a hierarchy of qDRIFT estimators with increasing circuit depths and couples adjacent levels by sharing their random Hamiltonian-term samples. This coupling makes the variance of the level differences decay with depth, allowing most samples to be taken on cheaper, coarse circuits and only a few on expensive, fine circuits. We prove that the resulting MLMC-qDRIFT estimator reduces the total gate complexity for fixed-precision observable estimation from the standard qDRIFT scaling $\mathcal{O}(\varepsilon^{-3})$ to $\mathcal{O}(\varepsilon^{-2}\log^2(1/\varepsilon))$, while preserving qDRIFT's lack of explicit dependence on the number of Hamiltonian terms. Numerical experiments for spin-chain dynamics confirm the predicted variance decay and demonstrate the practical gate-count savings of the multilevel construction.

Non-local conductance experiments using tunnel junctions can provide valuable spectroscopic information on both the transport and relaxation of quasiparticles in superconductors, as these techniques directly probe the quasiparticle charge and energy imbalance even at mK temperatures. In this work, we employ mesoscopic three terminal Cu and Al NIS devices to study non-local quasiparticle transport over length-scales on the order of the superconducting coherence length in this regime. Via a dual-bias scheme, which utilizes detector biases both above and below the superconducting gap, we are able to extract the effect of quasiparticle energy imbalance via its impact on the self consistent pair potential by symmetry considerations. We observe non-local conductance features due to pair-breaking which are anti-symmetric with respect to the polarity of the voltage bias, with a sharp onset during single electron tunneling at energies around $3Δ$. We compare these findings with quasiclassical simulations including inelastic effects to obtain estimates of the energy dependent inelastic scattering time. In addition, we demonstrate kinetic effects due to a large applied supercurrent which can also be captured in this formalism and decomposed with respect to the particle-hole symmetry and supercurrent direction, and discuss further opportunities for the advancement of this method.

We report the largest trapped-ion hardware demonstration of lattice protein-folding optimization to date, using bias-field digitized counterdiabatic quantum optimization (BF-DCQO) on a fully connected 64-qubit Barium development system similar to the forthcoming IonQ Tempo line. Six peptide sequences with 14-16 amino-acid residues are encoded using a coarse-grained tetrahedral lattice model, yielding higher-order spin-glass Hamiltonians with long-range interactions involving up to five-body terms and mapped to 46-61 qubits. The resulting instances are demanding for near-term quantum hardware because low-energy configurations must satisfy backbone-geometry constraints while optimizing dense residue-contact interactions. BF-DCQO uses a non-variational bias-feedback mechanism, where low-energy samples from each round define longitudinal fields that guide subsequent quantum evolutions. Across the studied instances, BF-DCQO shifts raw sampled energy distributions toward lower energies than uniform random sampling, with the strongest improvements appearing in residue-contact variables. To preserve this signal, we introduce a consensus-based post-processing pipeline that combines quantum-learned contact information with feasible backbone geometries. The resulting hybrid workflow reaches the classical reference energy in multiple instances and improves over the corresponding random-seeded pipeline. These results show that BF-DCQO can generate structured samples for dense protein-folding Hamiltonians at previously unexplored trapped-ion scales.

We propose a method to evaluate general thermodynamic fluctuations in open quantum systems, based on performing a two-point measurement scheme on the system using dynamics-dependent thermodynamic observables. Our approach allows one to obtain exact equalities for fluctuations of path-dependent thermodynamic quantities such as work and heat, and to isolate correction factors to Jarzynski's equality, requiring only access to the system degrees of freedom. This framework is flexible and can be applied to the limiting case of closed systems, recovering previous, yet seemingly contradictory, results from the literature. Moreover, the formalism admits a straightforward extension to strongly coupled open quantum systems. We investigate the effect of specific dynamical classes on the fluctuation relations, and show that the pure decoherence case is particularly special, as it deterministically does not contain any heat contribution and thus constitutes a class of open system dynamics for which the Jarzynski equality for work fluctuations is identically true at any coupling strength. Finally, we look explicitly at the shape and size of the correction factors to Jarzynski's equality for a qubit undergoing phase covariant dynamics, both in the weakly-coupled regime and in the deep non-Markovian regime.

We establish the rate at which the renormalized stress--energy tensor of a massless minimally coupled scalar field in the in-vacuum state of a collapsing null-shell spacetime approaches the corresponding Unruh-state value. At finite exterior radius, we establish the upper bound \[ |Δ\langle T_{μν}\rangle|\leq C(r)\,t_s^{-3} \] from the Cauchy-surface decomposition of the Hadamard difference and the branch-cut structure of the retarded Green function. At future null infinity, we show that the leading coefficient in the late-time expansion \[ Δ\langle T_{uu}\rangle\sim C_{uu}\,u_s^{-3} \] is nonzero, by computing the branch-cut residue explicitly at small frequency and using the Planck suppression of the thermal spectrum at large frequency to show that the dominant contribution to $C_{uu}$ has a definite sign. The result gives \[ Δ\langle T_{uu}\rangle\big|_{\Iscr^+}(u_s) \sim C_{uu}\,u_s^{-3}, \qquad u_s\to\infty, \] with $C_{uu}\neq 0$. The exponent is determined by the $ω^2\lnω$ branch-point singularity in the Wronskian of the $\ell=0$ radial wave equation, the same structure responsible for Price's law. The sign $C_{uu}<0$ is supported by a physical argument and by the numerical mode data of Gholizadeh Siahmazgi, Anderson, and Fabbri. The result confirms their conjecture that the approach is a power law. We conjecture that the same mechanism gives an analogous $t_s^{-7}$ bound for gravitational perturbations ($\ell_{\min}=2$), though the extension to the spin-2 case involves gauge issues not addressed here.

In this work, we demonstrate the effect of self-splitting spatial quantum correlations induced by Gouy phase engineering. In the process of spontaneous parametric down conversion the pump beam is structured with a mode superposition that produces a dynamical splitting and recombination of the light beam. This structure is transferred to the quantum correlations between signal and idler photons. As a result the joint two-photon probability distribution propagates like a self-splitting and recombining light beam, implementing a Mach-Zehnder-like interferometer. We observe heralded single-photon interference and two-photon NOON state interference. These results open new avenues for applications in quantum metrology.

Establishing the precise computational boundary between classically tractable fermionic systems and those capable of genuine quantum advantage is a central challenge in quantum simulation. While injecting non-Gaussian ``magic" inputs into free-fermion circuits is widely expected to generate intractable complexity, we identify a physically motivated intermediate regime. Supported by rigorous bounds and numerical evidence, we show that for a class of paired non-Gaussian fermionic states, essential quantum simulation primitives -- transition amplitudes, overlaps, and arbitrary-weight number correlators -- can be efficiently approximated to additive error under free-fermionic dynamics. This tractability stems from an algebraic reduction that compresses exponentially large multiparticle interference into a single coefficient of a multivariate Pfaffian polynomial. Because these classical estimators match the intrinsic $O(1/\sqrt{K})$ statistical uncertainty of quantum hardware utilizing $K$ measurement shots, they constitute a practical benchmark. Building on this foundation, we construct an additive-error estimator for high-weight Wilson observables in the noninteracting quench of recent trapped-ion experiments, providing a rigorous classical benchmark. Extending this to quantum chemistry, we demonstrate that core overlap-based subroutines for antisymmetrized products of strongly orthogonal geminals admit exact Pfaffian reductions. Ultimately, these results sharpen the boundary of quantum advantage, establishing that the paired-electron scaffold is effectively dequantized and clarifying exactly where quantum resources are indispensable.

The Euclidean bounce for vacuum decay enjoys an $O(4)$ symmetry that is lost in the presence of impurities than can catalyze the decay. We present a formulation for the calculation of the tunneling decay action, that is explicitly positive definite, for impurities whose effects are spherically symmetric so that the bounce symmetry is reduced to $O(3)$. The action constructed can be regarded as a generalization of the tunneling potential method, which implicitly assumed $O(4)$ symmetry. We show that the action obtained reduces to the tunneling potential for $O(4)$-symmetric cases and provide analytic examples with $O(3)$ symmetry and arbitrary wall thickness.

Finite local Hilbert-space truncations arise naturally in quantum simulations of lattice field theories and motivate qudit encodings, but their fault-tolerant advantage over qubit encodings remains unclear. We compare the non-Clifford cost of implementing quadratic diagonal evolutions, exemplified by $U=e^{-itφ_x^2}$ in a uniform field-amplitude discretization of a real scalar field, using either one logical $d$-level qudit or $n_b=\lceil \log_2 d\rceil$ logical qubits. We analyze two standard settings: product-formula simulation and LCU/block encoding, taking the resource metric to be the number of non-Clifford gates after synthesis into a discrete logical gate set. Because tight synthesis bounds for general single-qudit rotations are not known, we express the qudit constructions in terms of embedded two-level $SU(2)$ rotations and derive explicit finite-$d$ break-even conditions for their synthesis cost; these serve as compiler targets for when qudit encodings can outperform the qubit baseline. Within the constructive models studied here, product-formula implementations would require an exponentially stronger per-primitive synthesis advantage for qudits to win asymptotically, while in the LCU setting the qubit encoding is asymptotically cheaper in $d$. Nevertheless, the finite-$d$ threshold analysis identifies low dimensional regions in which qudits can yield meaningful constant-factor savings, particularly for LCU-based implementations. As a secondary analysis of the LCU construction, we use an idealized negligible-overhead qubit-qudit code-switching model to give an absolute $T$-count comparison, and reinterpret the savings as an allowable per-switch overhead budget.

Long-range quantum entanglement is essential for building large-scale quantum networks and unconditionally secure cryptographic systems based on quantum key distribution (QKD). While photonic integrated circuits offer a highly scalable platform, the fragility of phase coherence between spatial modes has prevented the distribution of path-encoded entanglement over long distances. Here, we report chip-to-chip distribution of path-encoded entangled states over 80 km between fully integrated silicon photonic transmitter and receiver chips. Telecom-band entangled photon pairs are generated via spontaneous four-wave mixing in on-chip spiral waveguides and distributed between chips over a dual-core, actively stabilized fiber link. Upon distribution, we measure a Bell state fidelity of $85.7 \pm 0.2 \%$. Implementing the BBM92 protocol with the same source, we obtain a secure key rate of 2.03 bit/s in the infinite-key regime. These results establish silicon photonic chips as a viable platform for long-distance path-encoded entanglement-based quantum key distribution, paving the way toward scalable, device-independent quantum networks.

We demonstrate optical squeezing below the shot-noise level generated through the interaction of an optical cavity field with two center-of-mass modes of a levitated nanoparticle, simultaneously cooled to occupation numbers well below unity. By analyzing the quadrature fluctuations of the cavity output through heterodyne detection, we resolve the full spectral covariance matrix of the optical field and map regions of sub-shot-noise squeezing as a function of detection phase and frequency. Operating in the resolved sideband and strong coupling regime where mechanical modes hybridize with the optical mode, we observe consistent squeezing in the band 70-95 kHz with a lowest variance of 0.98 (2$\%$ below vacuum fluctuations). We thus demonstrate optical squeezing mediated by multiple mechanical oscillators in their quantum ground state, bridging mechanical quantum control with non-classical light and establishing levitated optomechanics as a platform for multimode quantum interactions.

Near-degenerate resonance poles arise widely in open-wave systems, including gravitational-wave ringdowns, when two neighboring modes have almost the same complex frequency. On a finite observation window, the physical waveform is then dominated by a common carrier with a slowly varying interference envelope, so that attempting to treat the signal as a sum of two independently resolved damped sinusoids can become unstable in practice. Mathematically, an exact coalescence of poles leads to a double pole and a Jordan-chain (associated-vector) time dependence with a $t\,e^{-iωt}$ factor; here we show that the same carrier-plus-first-jet structure emerges more generally from a near-degenerate neighboring-pole \emph{pair} on a finite window, without requiring a true pole merger. We show that the local response is more naturally organized as a centered two-pole block about the shared carrier and, in the time domain, as a carrier-plus-first-jet waveform. This centered organization leads to a finite-window two-scale hierarchy: $κ$ controls when the leading correction to the carrier must be included, while $η^2$ controls the remaining error once that correction is retained. Toy two-pole numerics verify this scaling, and Kerr quasinormal modes provide a representative gravitational realization of the same finite-window centered organization.

We use a $^{87}\text{Rb}$ atomic vapor, suitable for an optically-pumped magnetometer (OPM) in Earth-field conditions, to study the noise properties of three strategies for generating pulsed optical pumping. We compare a frequency-modulated (FM) laser, amplitude modulation (AM) via an acousto-optic modulator (AOM), and amplitude modulation via a semiconductor optical amplifier (SOA). Pumping the ensemble to operate as a Bell-Bloom OPM, and with an equal degree of spin polarization, the three methods give nearly identical sensitivity, showing that the SOA, despite being an active device, can introduce negligible additional noise. Pumping the ensemble to operate as a free-induction-decay OPM, we observe longer unpumped coherence times with the SOA-AM method than with the FM method. Finally, using the higher power available from the SOA, we demonstrate an environment-limited sensitivity of $80\text{fT}/\sqrt{\text{Hz}}$ at $600\text{Hz}$ and 200fT$200\text{fT}/\sqrt{\text{Hz}}$ at $4\text{kHz}$, one to two orders of magnitude beyond what was achievable with the other pumping methods.

In the Schrödinger evolution of a quantum state time enters as a real parameter representing the coordinate. In a more consistent approach time should be defined as a quantum observable, with the evolution taking place in a four-dimensional spacetime. This is possible in the projection evolution model in which the wave function is defined in both space and time. This allows to construct the time operator and discuss the temporal structure of quantum processes. In this paper we discuss a photon travelling through a Mach-Zehnder interferometer, focusing the description on the temporal profile of the wave function. We show that in this approach the delayed-choice experiments can be explained by the temporal overlap of the photon and the devices in the interferometer.

The memory effects in open quantum systems can induce information backflow and revive quantum correlations, thereby providing a powerful way to protect and recover useful quantum resources in realistic noisy environments. However, such dynamics remains experimentally unexplored in multipartite quantum steering. Here we observe different non-Markovian evolution of tripartite quantum steering using Greenberger-Horne-Zeilinger-type mixed states, covering both death and revival processes. In particular, we experimentally demonstrate the more intricate asymmetric steering structure of tripartite quantum steering through different bipartitions, which do not arise in bipartite systems. Our results provide foundational insights into the hierarchical and directional structures in multipartite quantum steering, and highlight its potential as a useful resource for asymmetric quantum information processing.

The containment of malware in computing networks may be naturally formulated as a network influence minimisation problem, in which one seeks to limit the expected spread of an infection while balancing the operational cost of disabling network connections. Classical approaches often rely on Monte Carlo simulation of stochastic diffusion processes and greedy optimisation over candidate edge removals, resulting in significant computational overhead due to repeated influence evaluations. In this work, we propose a hybrid quantum approach which combines Quantum Amplitude Estimation (QAE) and Grover Minimum Finding (GMF) to provide quadratic improvements in both the estimation and optimisation components of the problem. Specifically, QAE replaces classical Monte Carlo simulation, reducing the sampling complexity of influence estimation from $O(1/\varepsilon^2)$ to $O(1/\varepsilon)$ for a target additive error $\varepsilon \ll 1$, while GMF reduces the number of candidate evaluations required to identify optimal edge removals from $O(|E_C|)$ to $O(\sqrt{|E_C|})$. We present a formal problem definition, describe the construction of the corresponding quantum oracles, and analyse the resulting complexity improvements under standard oracle assumptions. Preliminary experiments, including classical simulation of QAE and small-scale execution of Grover search on real quantum hardware, support the expected theoretical scaling. While practical implementation at scale requires fault-tolerant quantum devices, our results demonstrate that quantum algorithms offer a promising long-term direction for accelerating stochastic network optimisation problems such as malware containment.

Using a fully gauge-invariant approach, we compute for the first time in the literature relativistic effects on the redshift drift up to second order in cosmological perturbation theory. This is achieved by employing a set of light-cone coordinates that simplify the description of light propagation in an inhomogeneous and anisotropic universe. We show that redshift-space distortion occurs only as a second-order effect whereas, as known, it is not present among the linear perturbations. We then derive analytical expressions of the bispectrum for the leading-order perturbative contributions on sub-Hubble scales, providing some numerical evaluations. Our finding is that, at low redshift and for large momenta, the non-linearities in the bispectrum are enhanced more than the squared power spectrum.

In the present work, we study the causal structure of spherically symmetric black holes immersed in a Chaplygin-like dark fluid, emphasizing the impact of the fluid parameters on curvature and horizon formation. We show that the spacetime curvature is significantly stronger than in its similar counterpart, the Reissner-Nordstrom-de Sitter geometry with the same mass and charge, leading to modifications of the internal causal structure. For the presence of horizons the Chaplygin black hole possesses an upper bound $Q \approx 0.556219 M$, which is much smaller than that for Reissner-Nordstrom spacetime $Q_{\text{critical}} = M$ or of the Reissner-Nordstrom-de Sitter case $Q_{\text{critical}} = 3M/(2\sqrt{2})$, indicating that the black holes immersed in a Chaplygin-like dark fluid reach the extremal regime more easily. We derive a second critical condition for the Chaplygin cosmological parameter $B$, $B_c Q_c^4 = 4/3^9$, setting an upper bound on $B$ for a multi-horizon solution.

Compiling time-evolution operators of the form $U(t)=e^{-iHt}$ into hardware-native gate sequences is a central bottleneck for digital quantum simulation on noisy intermediate-scale quantum (NISQ) devices. Generic transpilation treats $U(t)$ as an arbitrary unitary, discarding the structure of Hamiltonian dynamics and producing circuits whose depth exceeds hardware coherence limits. We introduce a structure-aware compilation framework that treats product-formula decompositions as synthesis primitives rather than simulation approximations. The method combines (i) native placement of Hamiltonian terms onto the hardware coupling map, (ii) adaptive selection of Trotter blocks via a greedy discretization procedure, and (iii) variational refinement using a Trotter-initialized ansatz. Across Heisenberg, Ising, and XY models with $n=3$--$8$ qubits, the compiled circuits achieve fidelities $F>0.996$ with approximately linear scaling in the number of entangling gates, while generic synthesis produces circuits that are orders of magnitude deeper. On IBM Torino hardware, we observe a regime in which shorter approximate circuits outperform deeper exact decompositions: a 27-CX circuit achieves higher hardware fidelity ($F_{\mathrm{hw}}=0.987$) than a 187-CX exact circuit. These results demonstrate that, in the NISQ regime, structure-aware approximate compilation can outperform exact structure-agnostic synthesis, providing a practical pathway for executing Hamiltonian dynamics without requiring pulse-level control.

Quantum state discrimination is a fundamental information processing task that serves as a building block for numerous applications and provides implications at the foundational level. In this work, we consider minimum error discrimination of multi-copy states, where instead of preparing a single system we assume that multiple instances of the same state are prepared. Now the discrimination allows for measurements from multiple parties with different measurement strategies varying from global measurement strategy to ones restricted to different forms of local operations and classical communication strategies. By comparing the average success probabilities in quantum and classical cases, we find a qubit strategy that outperforms all the bit strategies. However, we find that there are other bit-like operational theories which can outperform the best qubit strategies even with a classical measurement strategy and we are able to identify instances of different theories where different measurement strategies are optimal. In this way, we are able to find instances of nonlocality without entanglement as well as provide general bounds for bit-like operational theories.

We formulate a variational principle for non-relativistic quantum mechanics inspired by Gauss's principle of least constraint. We define a quantum constraint functional as the probability-weighted square deviation between the actual motion and the unconstrained motion that would arise from external forces alone. In this functional, the quantum potential plays the role of an intrinsic constraint that modifies the acceleration. Minimizing this quantum constraint functional with respect to the acceleration field yields the quantum Euler equations, which together with the continuity equation are equivalent to the Schrödinger equation. The principle is instantaneous and provides a differential characterization of quantum evolution. We demonstrate that this formulation is not a mere rewriting of existing dynamics: it provides a unified and technically economical treatment of geometric constraints and velocity-dependent dissipative forces, neither of which admits a straightforward global variational formulation. Potential applications to a broad range of quantum phenomena are also indicated.

There exists a geometric phase for a quantum state during the adiabatic evolution of the system. If the adiabatic procedure happens between the system and the environment interacting with it similar to Born-Oppenheimer (BO) approximation, we can introduce a temperature into the environment, which can be regarded as in an equilibrium state. Then a temperature-dependent geometric phase can be obtained for the system, which originates from the Abelian gauge potential induced by the BO approximation. This gauge potential contributes to the effective potential of the system, which is temperature dependent, too. Finally, we demonstrate them using an example of H_2^+ ion system.

We present a genuinely non-radial quantum-mechanical route by which $π$ emerges from equatorial localization on the sphere. For the highest-weight branch of spherical harmonics, this localization is captured by a natural geometric rigidity index, whose exact finite-quantum-number value is a Wallis partial product. The mechanism is realized in two settings: the standard rigid rotor and the surface sector of a thin spherical shell, where radial freezing reduces the dynamics to the same angular problem. In the large-quantum-number limit, the probability cloud collapses toward the equator, the rigidity index approaches its classical value, and the Wallis formula is recovered through the correspondence principle. The result shows that Wallis-type structures in quantum mechanics can arise as exact signatures of semiclassical localization encoded by a simple geometric observable.

We study a projection-type gradient flow for equality-constrained maximisation of a smooth bilinear control objective on $\mathcal{H}=L^2(0,T;\mathbb{R})$, eliminating Lagrange multipliers through an $(M{+}1)\times(M{+}1)$ moving Gram matrix $Γ(s)_{\ell\ell'}=\int_0^T S(t)\,c_\ell(s,t)\,c_{\ell'}(s,t)\,\mathrm{d}t$. The flow generates monotonic ascent in continuous time but becomes unstable on discretisation; existing implementations rely on heuristic step-size safeguards lacking rigorous justification. We close this gap by replacing $Γ$ with $Γ_{\varepsilon}:=Γ+\varepsilon^{2}I$ and prove: (i) an exact spectral identity giving $κ(Γ_{\varepsilon})=(σ_{\max}^{2}+\varepsilon^{2})/(σ_{\min}^{2}+\varepsilon^{2})$; (ii) objective monotonicity $\mathrm{d}J/\mathrm{d}s\ge 0$ for all $\varepsilon\ge 0$; (iii) constraint drift $|h_{m}-C_{m}|=\mathcal{O}(\varepsilon^{2})$ with a computable prefactor; (iv) convergence of the regularised trajectory to the unregularised one in $L^{2}(0,T)$ at rate $\mathcal{O}(\varepsilon^{2})$ under uniform invertibility of $Γ$; and (v) a discrete CFL criterion $Δs\,G\,\|Γ_{\varepsilon}^{-1}\|\leα<2$ guaranteeing objective monotonicity of the forward-Euler scheme up to $\mathcal{O}(Δs^{2})$ local truncation error. The theory is validated on a three-level bilinear benchmark for all-optical Bell-state preparation, where $κ(Γ)\in[10^{9},10^{11}]$, the predicted $\varepsilon^{2}$ rate is confirmed over eight decades, and moderate regularisation eliminates step rejections and reduces constraint drift by more than an order of magnitude at unchanged final fidelity.

It is a well-established fact that some quantum correlations can be nonlocal, meaning that they cannot be described by a local hidden variable model. Certain quantum correlations have a form of nonlocality so strong that they cannot be reproduced even by models having an arbitrarily small local hidden variable component. These correlations are called fully nonlocal and lead to Bell inequalities in which the maximum quantum value saturates the non-signaling bound. A well-known example of this effect, which is also referred to as quantum pseudo-telepathy or all-versus-nothing proofs of nonlocality, is the quantum distribution fulfilling the Peres-Mermin square, in which the underlying state is a $4\times4$ dimensional maximally entangled state. Other examples of full nonlocality are known but, so far, all of them are for maximally entangled states and it is an open question whether maximal entanglement is necessary for full nonlocality. In this work, we first establish a link between full nonlocality and the concept of antidistinguishability of quantum states. We use this connection to show that in every bipartite $d\times d$ Hilbert space, with $d\geq3$, there are non-maximally entangled states that are fully nonlocal. In fact, we derive simple sufficient conditions for full nonlocality that are only based on the smallest and largest Schmidt coefficients. We also show that in every dimension there exist pure entangled states that do not exhibit full nonlocality. Finally, we show that all pure entangled states can be activated to show full nonlocality in the many-copy scenario.

We study thermodynamic properties, in particular the Temperature~(T), Angular velocity~($Ω_h$) and Entropy~(S) of the of magnetically charged slowly rotating (with rotation parameter $a \lsim 0.10$) Kerr black-hole(BH) with the inclusion of cosmological constant ($Λ$) in the background of nonlinear electrodynamics (NLED). At first we calculated the nonlinear electromagnetic magnetic charged density $ρ_{NLED}$$(r)$ which is needed to calculate the magnetic mass of the slowly rotating Kerr-BH. We showed the mass profile $M(r)$ of the BH for different combinations of magnetic charges~($q_m$) and non-linear parameter ($β$) presence in the Lagrangian density. We found that $M(r)$ attains a plateau for values of $r$ close to the the cosmological length~($L$), where $L^2$= $\pm \fracΛ{3}$, irrespective of the combinations of $q_m$ and $β$. The $\pm$ sign corresponds to the de-Sitter(dS) and Anti-de-Sitter(AdS) respectively. Afterwards we showed the allowed parameter spaces in $a$-$M$ plane using sharkfin diagram for different values of $L$ with positive values of the horizon function, $Δ(r)$, and explain the extremal criterion and asymptotic limit. We showed the values of $r$ where the horizon function of the quadratic polynomials becomes zero and called them as the inner(Cauchy), outer(Event) and large cosmological horizons with different values of $a$. We showed that the horizon structure depends on $a$, $L$ and the mass profile $M(r)$. Finally, we tabulated the numerical values of three thermodynamic parameter, i.e., $T$, $Ω_h$ and $S$ at those horizons surfaces. Our results demonstrate that NLED with cosmological constant significantly modifies both the internal structure and thermodynamic properties of Kerr-BH.

Identifying compact binary coalescences buried within the non-Gaussian and non-stationary data taken by gravitational-wave interferometers requires sophisticated search pipelines, such as the PyCBC analysis. A critical task for these pipelines is determining the statistical significance of candidate events by comparing a "ranking statistic" against a large background set. Currently, PyCBC's ranking statistic incorporates the joint probability of the relative arrival times, phase delays and amplitude ratios of the signals seen in different detectors. These parameters are tightly constrained for physical signals but are more broadly distributed for noise. PyCBC currently relies on precomputed binned histogram-based density estimators using Monte-Carlo simulations to obtain these probabilities. However, the storage requirements for these histograms scale prohibitively with the size of the detector network, preventing PyCBC from effectively analyzing four or more detectors. In this paper, we demonstrate that these histograms can be replaced with normalizing flows, a machine learning approach to density estimation. Applying this method to data from the third observing run of Advanced LIGO and Virgo, we demonstrate that normalizing flows reduce storage requirements by over three orders of magnitude. Furthermore, our approach maintains high sensitivity, with less than a 0.05% drop in the recovery of simulated signals at a fixed false-alarm rate. By relaxing several simplifying assumptions previously required by Monte-Carlo methods, we also achieved up to a 6.55% increase in recovered signals for specific detector combinations. These results suggest that normalizing flows provide a scalable, flexible framework for the PyCBC pipeline as it expands to include four or more detectors, or to extend to searches for precessing or higher-mode signals, in future observing runs.

We present a method for generating a laser beam with flat intensity and phase profiles in the focal region where the beam interacts with neutral $^{87}$Rb atoms in an array of optical dipole traps. We synthesize the beam as a superposition of Hermite--Gaussian or Laguerre--Gaussian modes. Then we give analytical expressions for the coefficients of such a superposition, an analysis of beam propagation along the $z$ axis in the vicinity of the waist, and several other related theoretical issues. Rydberg two-qubit dynamics driven by this flat-top profile are analyzed through numerical solutions of the Lindblad master equation using our in-house Julia package. Beam preparation is demonstrated on a neutral-atom experimental platform. Measurements reveal a difference in the visibility of Rabi oscillations for addressed atoms compared with neighboring ones, confirming the effective spatial selectivity provided by the flat-top beam profile.

Random numbers sequences (RNSs) play a vital role in various scientific and engineering applications. They are critical to the integrity of classical and quantum cryptography, the accuracy of mathematical modeling and Monte Carlo simulations, and the core mechanics of applications in fields as diverse as gambling and statistical sampling. While the primary criteria for RNSs sources are their quality and generation rate, their integration into experimental designs is equally significant for many fundamental physical tests and applications. This work presents a comparative analysis of optical random number generation architectures, which can be seamlessly included into various advanced classical and quantum optical experimental schemes. In particular, we evaluate the trade-off between the high generation rate of an attenuated laser (a quasi-single-photon source) and the superior statistical quality of a heralded single-photon source operating at a much lower frequency. To overcome the limitations of each individual source, we propose and examine a novel hybrid architecture that utilizes their mixed radiation, enabling the generation of highquality RNSs at an enhanced rate. Furthermore, we demonstrate that the raw sequences generated by such a source can not only exhibit but, in some cases, even surpass the degree of randomness achieved by sequences processed through powerful randomness extractors.

Quantum annealing targets low-energy solutions of Ising/QUBO problems, but reliable assessment requires more than best-energy comparisons. This dissertation develops a benchmarking framework for D-Wave quantum annealers that combines strong classical baselines, sampling and diversity metrics, and thermodynamic cost. Its first contribution, SpinGlassPEPS.jl, is a topology-aware tensor-network heuristic for optimization and sampling on Pegasus/Zephyr-like graphs. It maps Ising instances to local Potts clusters, represents the partition function with PEPS, and performs branch-and-bound search in probability space. Benchmarks show that it is a physically interpretable reference solver, though approximate contractions limit its competitiveness on the largest instances. The second contribution treats quantum annealers as effective thermal machines, relating success probability and solution quality to dissipation, entropy production, and effective temperature. Carefully placed pauses can improve performance while reducing thermodynamic cost, although longitudinal fields may become harmful in paused schedules. The thesis also introduces reinforcement-learning post-processing to improve returned samples and exact small-system simulations to probe annealing dynamics. Overall, it argues for quantum-annealing benchmarks that jointly measure algorithmic performance and physical expenditure.

Combinatorial optimization is widely regarded as a primary application for near-term quantum processors, although a definitive demonstration of the practical quantum advantage remains elusive. Recent studies have reported that both gate-based quantum circuits and quantum annealers can outperform state-of-the-art classical heuristics on multi-objective optimization (MO-MaxCut) problems. However, these studies did not fully account for the substantial pre- and post-processing overheads intrinsic to quantum solvers, leading to incomplete comparisons between quantum and classical approaches. In this work, we re-examine the same benchmark suite using GPU-based quantum-annealing-inspired algorithms (QAIAs), which, analogously to quantum processors, generate probabilistic samples and thus serve as formidable classical contenders. Our results show that QAIAs can sample candidate solutions approximately two orders of magnitude faster than previously studied quantum processors. In terms of end-to-end runtime, QAIAs also surpass industry-leading classical solvers, thereby establishing themselves as the superior performers among the quantum and classical solvers evaluated thus far for the MO-MaxCut instances.

The Fock-Darwin (FD) quantum system describes the motion on the plane of a charged particle under the action of an isotropic oscillator potential together with a perpendicular constant magnetic field. When the isotropic oscillator is suppressed, the FD system leads to the Landau Hamiltonian with infinitely degenerate Landau levels. The Fock-Darwin-Darboux (FDD) system is the generalisation of the FD system to a particle moving on the Darboux III space, which is a conformally flat surface with non-constant negative curvature. We present a systematic study of some information-theoretic entropy and dispersion-like measures for these quantum systems. Since both systems are exactly solvable, analytical expressions for Shannon, Rényi and Tsallis entropies, among others, can be obtained. We show that for the FD system, its information-theoretic measures are formally the same as the ones for the harmonic oscillator, provided a modified effective frequency depending on the magnetic field is introduced. In the FDD case, the nonlinear nature of the underlying manifold precludes the existence of a simple closed form for the wave-function on momentum space, which is numerically analysed. We compare the numerical behaviour of the different entropy measures and we analyse the interplay arising in the FDD system between the curvature parameter and the magnetic field. In particular, it is shown that the Landau system on the Darboux III space has no infinitely degenerate Landau levels.

The optimization of front-end crude oil scheduling is a critical determinant of refinery profitability and operational stability. However, the coupling of discrete logistics events (e.g., vessel berthing) with continuous material flows (e.g., pipeline transfers) renders this problem an NP-hard Mixed-Integer Linear Programming (MILP) challenge, often intractable for classical solvers at industrial scales. This study proposes a novel hybrid quantum-classical framework to address these computational bottlenecks. We employ Benders Decomposition to decouple the monolithic model into a discrete Master Problem (MP) and a continuous Subproblem (SP). To exploit the search capabilities of quantum computing, the MP is reformulated as a Quadratic Unconstrained Binary Optimization (QUBO) model and solved via a hybrid quantum solver, while the SP enforces mass balance and quality constraints through iterative optimality and feasibility cuts. Extensive experiments on 15 multi-scale instances demonstrate that the proposed framework significantly outperforms traditional metaheuristics (e.g., Genetic Algorithms, Tabu Search), reducing total operating costs by approximately 73--80% and achieving computational speeds comparable to state-of-the-art commercial solvers (Gurobi). By effectively leveraging global optimality cuts, the method overcomes the tendency of heuristic approaches to trap in local optima, providing a robust and scalable solution for complex refinery logistics.

We derive integrable deformations of the 2d Breitenlohner-Maison (BM) sigma model that describes the stationary, axisymmetric sector of 4d general relativity, as well as higher-rank generalisations thereof, using the framework of 4d Chern-Simons theory. In particular, we consider deformations of the boundary conditions and action of the 4d Cole-Weck model, which lead to deformations of the BM model associated with solutions to the homogeneous and inhomogeneous classical Yang-Baxter equations respectively.

We investigate spherical collapse in Eddington-inspired Born--Infeld (EiBI) gravity in the subhorizon, pressureless, and quasi-static regime, emphasizing the matter-gradient correction that appears in the weak-field limit of the theory. Starting from the nonlinear continuity and Euler equations, we derive the evolution equation for the density contrast and show that the EiBI contribution depends explicitly on spatial derivatives of the matter density. This feature makes the ideal discontinuous top-hat construction ill-defined, since gradient terms become singular at the boundary, and requires a regularized overdensity profile together with a coarse-graining prescription. We adopt an effective physical-gradient closure for the EiBI source term and compare two matched initial configurations: a regularized Tanh profile and a peak-based profile, calibrated to share the same characteristic radius and cumulative mass proxy. Within this framework, we compute the linear collapse threshold $δ_c(z_{\rm coll})$, the turnaround overdensity $δ_t(z_{\rm coll})$, the turnaround radius $R_t(z_{\rm coll})$, and the virial overdensity $Δ_{\rm vir}(z_{\rm coll})$. Relative to the $Λ$CDM reference case, the EiBI correction lowers $δ_c$, enhances both $δ_t$ and $Δ_{\rm vir}$, and produces a more modest reduction of $R_t$, with deviations increasing with the dimensionless coupling $\hatκ_{\rm BI}$ over the range considered. The nonlinear overdensity observables show the strongest response to the EiBI correction and retain a residual dependence on the internal shape of the matched profile, whereas the turnaround radius is comparatively less affected. These results identify spherical collapse as a sensitive probe of EiBI matter-gradient couplings and motivate applications to halo statistics and nonlinear structure formation.

Ensuring the integrity of quantum circuits is a significant challenge in the Noisy Intermediate-Scale Quantum (NISQ) era, where circuits are subject to compilation transformations, hardware constraints, and potential adversarial modifications. Existing validation approaches typically rely on either structural analysis or behavioral evaluation, leading to incomplete assessment of circuit correctness. In this work, we investigate the relationship between structural, interaction-level, and behavioral perspectives of circuit integrity, demonstrating that a single aspect of integrity is insufficient to guarantee circuit integrity; structural similarity alone does not ensure behavioral equivalence. To address this problem, we use a three-layer metric framework that combines the Structural Integrity Score (SIS), the Operational Integrity Score (OIS), and the Interaction Graph Semantic-Logical Score (IGS). SIS captures global structural properties, OIS quantifies behavioral divergence using Jensen-Shannon distance, and IGS models interaction patterns and dependencies in a pre-execution setting. Through controlled anomaly injection on benchmark quantum circuits, we demonstrate that each metric captures a different aspect of circuit deviation. In particular, structural blind-spot cases (SIS >= 0.95) reveal a clear limitation of structural analysis, where OIS detects anomalies in 93.85% of instances, while IGS detects 72.58%. These results highlight that the metrics provide complementary insights and that a single metric is insufficient for reliable circuit validation.

Recent advances in quantum computing have enabled the development of quantum processors with hundreds of qubits. However, noise continues to limit the amount of useful information that can be extracted from these systems, making it essential to identify the regime in which experimental outputs remain reliable. In this work, we benchmark Quantinuum Helios-1, a 98-qubit trapped-ion quantum processing unit, using the linear ramp quantum approximate optimization algorithm (LR-QAOA). To this end, we perform large-scale noiseless simulations on JUPITER, Europe's first exascale supercomputer, for circuits of up to 48 qubits and 3,384 two-qubit gates. These simulations, executed on 4,096 nodes equipped with 16,384 GH200 superchips and high-bandwidth CPU-GPU interconnects, provide a reference for validating experimental results at the edge of classical tractability. We find that, up to 48 qubits, Helios-1 remains in a noise-tolerant region, i.e., its samples cannot be clearly distinguished from those coming from a noiseless simulation. We then extend the analysis to larger system sizes using experimental data only, and apply a mean-of-means resampling procedure with a 3$σ$ threshold to determine whether the QPU output is statistically distinguishable from random sampling. This analysis identifies a regime of coherent performance up to 93 qubits (12,834 two-qubit gates), beyond which, at 95 qubits, the outputs become statistically indistinguishable from random sampling. These results demonstrate how exascale classical simulation can be used to validate quantum processors, and provide a quantitative boundary between noise-tolerant and random regimes in quantum processors.

Motivated by recent evidence that equal-time correlators can be simpler than the corresponding wavefunction coefficients, we study de Sitter correlators in conformally coupled $φ^3$ theory directly. By inverting the momentum-space dressing rules, we derive a time integral representation for generic graphs and show that its natural building blocks are flat space correlators of fields and conjugate momenta. Among other things, this representation gives two useful recursive structures, one obtained by collapsing leaves and one by fusing lower-point graphs. In this representation several simplifications also become immediate. Graphs with an odd number of conjugate momentum insertions vanish, explaining the weight drop of odd-point correlators, melonic insertions collapse to lower complexity graphs and the leading behavior near total and partial-energy singularities is manifest, closely paralleling the flat space story. We then take a first step beyond the integrand and study integrated answers. For tree level families, in particular chains and stars, we find that the symbol alphabet of the correlator is smaller than that of the corresponding wavefunction, with the missing letters admitting a natural interpretation in terms of tubing data. These results support a correlator-first viewpoint for de Sitter observables: part of their simplicity appears to be intrinsic to the correlator itself, rather than inherited indirectly from the wavefunction.

Ultra-faint dwarf galaxies serve as powerful testing grounds for wave dark matter models through dynamical stellar heating. Previous simulation-based work derived a lower bound on the fuzzy dark matter particle mass using a diffusion approximation valid only when the de Broglie wavelength is much smaller than the galaxy's half-light radius. We simulate the dynamical evolution of stellar clusters in FDM halos across a wide mass range and find that for sufficiently low masses, where the de Broglie wavelength is much larger than the cluster size, tidal heating is the main mechanism. We also find that a reduced soliton mass and tidally stripped halo can suppress the heating. We demonstrate that in order to constrain FDM mass from cluster heating, the structure and environment of the FDM halo must be carefully considered.

Imaginarity, stemming from the complex structure of quantum mechanics, has recently emerged as a fundamental resource, yet its dynamical generation remains largely unexplored. In this work, we introduce the notion of imaginarity-generating power (IGP) of unitary dynamics, which quantifies the ability of unitary operations to produce imaginarity from initially real quantum states. To quantify imaginarity, we employ a measure based on the Hilbert--Schmidt norm, which we show to be monotone under real unital operations. Within the framework of dynamical resource theories, we derive an exact expression for the purity-constrained IGP in arbitrary dimensions and show that, for pure real input states, it depends solely on intrinsic and experimentally accessible properties of the unitary. We further analyze its average behavior over ensembles of states with varying purity under both uniform and Hilbert--Schmidt distributions. We prove that it satisfies the essential properties of a valid resource monotone within the dynamical resource theory of imaginarity. We also characterize the unitaries that maximize the IGP and determine the corresponding bounds. Moreover, for Haar-random unitaries, we show that the IGP concentrates near its maximal value in high dimensions with small fluctuations, indicating that typical high-dimensional quantum dynamics are highly effective at generating imaginarity.

From the perspective of quantum information science, we investigate tree-level Bhabha scattering between an incident electron $A$ and a positron B, where $B$ is initially entangled with a spectator electron $C$, which does not participate in the scattering interaction.We find that the quantum electrodynamics (QED) scattering between $A$ and $B$ can drive the global $ABC$ system into a genuine tripartite entangled (GTE) state. Using four canonical tripartite entanglement metrics, we systematically characterize and quantify the GTE of the composite system, and demonstrate that the scattering momentum of the $A$-$B$ pair and the initial $B$-$C$ entanglement are the key resources governing GTE generation.We further analyze the monogamy of quantum correlations, which imposes fundamental constraints on the shareability of quantum resources in multipartite systems. Specifically, we systematically study the monogamy relations for the squared entanglement of formation and squared quantum discord in our scattering model, and find that monogamy constraints are markedly relaxed in the non-relativistic regime, enabling enhanced shareability of quantum correlations across the three particles. This work uncovers novel quantum correlation properties of fundamental QED scattering processes, and provides direct theoretical guidance for the development of QED-based quantum information processing protocols.

Advances in quantum information science (QIS) are providing transformative insights into the complexity of quantum many-body systems, potentially defining new frontiers in nuclear and high-energy physics. This review explores how QIS-derived techniques are fostering new analytic frameworks and algorithms - both classical and quantum - to tackle (some of the) present barriers to discovery in fundamental physics, with applicability to other science domains. We highlight how these techniques are shedding new light on the structure and dynamics of hadrons, nuclei, matter in extreme conditions, and beyond. Importantly, they are expected to play an essential role in the development of large-scale quantum simulations of such systems, particularly in setting the balance among quantum and classical computational resources.

Fluxonium qubits combine long coherence times with strong anharmonicity, making them a promising platform for scalable superconducting quantum processors. Recent experiments have demonstrated high-fidelity operations in multi-qubit processors while suppressing stray qubit interactions using fluxonium-transmon-fluxonium (FTF) architectures. However, scaling such systems to larger arrays is constrained by a trade-off between achievable coupling strength, crosstalk suppression and qubit-qubit spacing required for wiring in a two-dimensional architecture. Multimode couplers, such as the double-transmon coupler (DTC), provide a promising pathway to overcome this limitation by enabling stronger interactions without compromising qubit spacing and isolation. Here, we develop a quantitative design framework for fluxonium-based quantum processors employing DTCs. Central to this work is a frequency-partitioned architecture that places qubit transitions, tunable-coupler excitations, and resonator modes in well-separated spectral regions. This structured allocation reduces parameter interdependence and enables the concurrent optimization of gate operations, readout, and qubit reset. By formulating device design as a multi-objective optimization problem under realistic experimental constraints and fabrication-induced disorder, we develop a tractable sequential workflow and determine a feasible parameter regime that simultaneously supports high-fidelity single- and two-qubit gates, fast qubit reset, and robust dispersive readout. These results establish a system-level architectural methodology that links circuit parameters to processor-level performance, and provide an experimentally actionable pathway toward scalable fluxonium quantum processors.

We investigate the quantum nature of naked curvature singularities in Einstein-Yang-Mills (EYM) theory using the Horowitz-Marolf (HM) criterion, which assesses quantum singularities via the evolution of quantum scalar fields. Focusing on timelike singularities in spacetime dimension $ D \geq 5 $, we analyze both pure Yang-Mills and Einstein-Maxwell-Yang-Mills (EMYM) solutions. We then incorporate higher curvature corrections through Gauss-Bonnet (GB) terms. From positivity requirement the expression under square root that arises in GB may create a secondary singularity that shall be scrutinized carefully. Our analysis reveals that while EYM and EMYM spacetimes remain quantum mechanically singular, the inclusion of GB corrections can, in general, render the singularity quantum mechanically regular for specific values of the mass parameter $m$, which is related to the YM charge $Q$ in $ D = 5 $ space-time dimensions. Contrary, for space-time dimension $D \geq 6$, although the outer (secondary) singularity may be healed quantum mechanically for certain values of the mass parameter $m$, the central singularity remains quantum mechanically singular.

Schroedinger's equation gave early quantum theory a visual language that looked like physics again: a wave evolving by a linear differential equation. This essay argues that the same success also seeded a recurring impulse to keep quantum theory "classical-looking" by treating the wave function as a physical wave. Schroedinger quickly realized that, for many-particle systems, the wave function is naturally defined on configuration space rather than ordinary physical space, blocking any straightforward reading of it as a literal classical wave. Read through Mach and Boltzmann, who shaped his intellectual outlook most deeply, his achievement appears double-edged: it provided an extraordinarily powerful picture for calculation and discovery, while also warning against taking that picture too literally. I argue that this tension never fully disappeared. It still reappears in modern physics whenever the wave function, or in quantum field theory the field itself, is treated as ontology rather than as part of a representation tied to measurement and observational context, a point sharpened by Bell-type no-go theorems. The centenary moral is: use pictures boldly, but demote them ontologically.

We discuss a super bra-ket formalism for quasi-Hermitian Liouvillian operators within the framework of rigged Hilbert spaces (RHS). An RHS in terms of the Liouville space, referred to as a rigged Liouville space (RLS), is reconstructed by exploiting the mathematical fact that the space of Hilbert-Schmidt operators is unitarily equivalent to the tensor product of Hilbert spaces. The obtained RLS endows a rigorous foundation of the construction for the super bra-ket and for the spectral decompositions of both Hermitian and quasi-Hermitian Liouville operators, which are characterized by the generalized eigenvectors in the dual spaces. Furthermore, within this framework, the non-Hermitian Liouvillian operator and its adjoint can be constructed symmetrically, with their symmetric structure preserved. As an application of our RLS methodology, we examine the Liouville operators corresponding to Hermitian and non-Hermitian harmonic oscillators and elucidate the differences between their spectral decomposition forms.

Slater-type orbitals (STOs) provide the physically correct description of atomic wavefunctions but have been largely replaced by Gaussian-type orbitals in computational chemistry due to the lack of closed-form multi-center integrals. We present a systematic study of amplitude encoding of STOs on quantum computers using matrix product states (MPS). For one-dimensional orbital functions of the form $p_d(x) e^{-ζx}$, we derive analytical MPS constructions with constant bond dimension $χ= d + 1$, requiring $O(n)$ classical and quantum resources for $n$-qubit registers with no grid sampling. We demonstrate a complete one-electron integral pipeline -- overlap, kinetic energy, and nuclear attraction -- in one dimension, validating the overlap and kinetic energy on IBM Heron processors at 5~qubits with 0.67\% hardware-induced error using Zero-Noise Extrapolation. In three dimensions, we compute multi-center overlap integrals between 1s and 2s orbitals in Cartesian coordinates with 0.02\% discretization error at 18~qubits. A systematic entanglement analysis reveals that the MPS bond dimension of three-dimensional STOs in Cartesian coordinates saturates with increasing grid resolution -- reaching $\sim$138 for the hydrogen 1s orbital at 12~qubits per coordinate -- establishing bounded encoding complexity rather than the exponential scaling initially expected. The SVD truncation threshold provides a practical resource parameter, reducing the bond dimension to 39 at threshold $10^{-6}$ with negligible accuracy loss. These results map the entanglement landscape for amplitude encoding of atomic orbitals and establish MPS-based state preparation as a viable path toward exact STO basis sets on quantum computers.

Variational quantum algorithms (VQAs) constitute a prominent framework for exploring the capabilities of near-term quantum computers. As the effectiveness of VQAs depends heavily on the design of variational quantum circuits, Quantum Architecture Search (QAS) has emerged as a critical research area to automate the discovery of high-performing circuit structures. This paper reviews key advancements in current research on QAS, including its core concepts, representative methodologies, and applications. The content is structured to ensure broad accessibility for a diverse audience of researchers while preserving the core principles of complex methodologies. In addition, we discuss remaining challenges and suggest potential research directions to offer perspectives on future exploration in this rapidly evolving field.

We extend the established Galilean/relativistic structural divider in algebraic quantum field theory, namely, the absence of Reeh-Schlieder and of Tomita-Takesaki modular flow on local algebras of any Galilean Haag-Kastler net satisfying a natural axiom set augmented by the Bargmann-charge hypotheses (G7$^*$)(a) and (G7$^*$)(d) to curved backgrounds via the Newton-Cartan ($c\to\infty$) limit. We show, for the free Klein-Gordon field on Minkowski and on static globally hyperbolic spacetimes admitting a Post-Newtonian expansion, that a position-independent rest-energy rescaling produces in the limit a Galilean Haag-Kastler net satisfying the axioms of Ref. [1] in flat-space form (Minkowski) or in a curved-space modification (Killing-flow invariance and uniqueness of the vacuum replacing full translation invariance) appropriate to the static case. The Bargmann central charge equals the Klein--Gordon mass~$m$; the gravitational potential $V(x)$ enters the limiting Schrödinger Hamiltonian but not the algebraic structure obstructed by the Galilean Reeh-Schlieder no-go theorem. The strengthened obstruction theorem of Ref. [1] extends to the modified curved-space setting on Fock representations, and the limiting net carries no modular flow on local algebras. Schwarzschild is treated as a worked example: the Killing horizon shrinks to a point, the Hartle-Hawking thermal state has no $c\to\infty$ limit, and the Boulware vacuum limits to the gravitational hydrogenic ground state. The Reissner-Nordström metric is included as a sanity check confirming that leading Post-Newtonian misses the electromagnetic content of the background. We discuss how Newton's constant~$G$ enters the present (background-metric) framework only at the level of the limiting Hamiltonian, and indicate where dynamical-metric extensions would require $G$ to play a structural role.

Quantum computing has advanced rapidly in recent years and has shown advantages in a variety of domains. In this paper, we investigate its potential for discrete simulation optimization in the fixed-confidence setting, a fundamental problem in the simulation literature. We first introduce a quantum simulation oracle that prepares a coherent superposition over all candidate solutions and provides the foundation for quantum algorithm design. Based on this oracle, we develop the first Grover-search-based quantum algorithm for discrete simulation optimization, called SOGAS. In particular, SOGAS uses a binary-search framework to progressively eliminate suboptimal solutions while carefully controlling the error probability, and eventually identifies a set of near-optimal solutions. We prove that SOGAS returns a near-optimal solution with probability at least the prescribed confidence level and achieves a quadratic speedup in the dependence of query complexity on the number of candidate solutions. Numerical experiments further show that SOGAS substantially outperforms classical benchmarks and provide empirical evidence for quantum advantage in discrete simulation optimization.

The measurement of the Hubble constant from gravitational wave (GW) sources is one of the independent avenues to shed light on the Hubble tension, which is associated with about an $8\%$ mismatch in the value of the Hubble constant inferred from low-redshift and high-redshift cosmological probes. Such a key measurement is expected from GW sources as it is a direct measurement of the Hubble constant using the luminosity distance without the need for any luminosity distance calibration. However, such a measurement relies strongly on the reliability of the independent inference of the source redshift of the GW source. As a result, it becomes pertinent to gauge the accuracy and precision of techniques in understanding their reliability in inferring redshifts of GW sources. In this work, we show the requirement of the spectral siren technique in knowing the mass distribution of BBHs across cosmic redshifts in order to make a reliable inference of the Hubble constant. We show by a blinded mock data challenge analysis the criticality in capturing the underlying metallicity dependence of the BBH mass distribution and its interplay with time-delay distribution for a robust inference of the Hubble constant using the spectral siren technique. In order to have a reliable measurement of the Hubble constant at the level required to resolve the Hubble tension in the future, the mass distribution of the BBHs needs to be independently inferred at all relevant redshifts with an accuracy less than the statistical uncertainty. Otherwise, a mismatch of the true model and the underlying assumption made in the analysis can lead to a best-fit model for the wrong value of both BBH population parameters as well as the Hubble constant.

We prove that the standard Galilean Haag--Kastler axioms, augmented by Bargmann mass superselection, are inconsistent with the Reeh--Schlieder property: no such net admits a vacuum that is cyclic and separating for every local field algebra. Two ingredients combine: Galilean Schrödinger fields annihilate the Fock vacuum, and Bargmann mass superselection forbids the Hermitian-combination evasion that keeps relativistic axioms consistent. The result extends beyond the Fock-representation hypothesis: any Galilean Haag--Kastler net whose canonical fields carry definite Bargmann mass charges and admit time-zero restrictions on a field-algebra-stable common dense domain is incompatible with Reeh--Schlieder. The bounded-below mass spectrum and the vacuum-at-spectral-minimum, usually imposed as separate axioms, are derived consequences -- of positive-energy boost positivity and a Bose-CCR algebraic descent, respectively. The Tomita--Takesaki modular flow is consequently unavailable on Galilean local field algebras. We identify the Reeh--Schlieder property as the precise structural ingredient distinguishing relativistic from Galilean algebraic quantum field theory: relativistic AQFT has it as a theorem, Galilean AQFT cannot.

We develop, as the first of a six-paper series, an operator-algebraic framework relating non-relativistic quantum mechanics and special relativity. Three structural facts organize the framework. (i)~The photon sector of free QED is a transparent realization: classical Fourier--Maxwell theory supplies a complex Hilbert-space scaffold (inner product, symplectic form, Schrödinger-form mode equation, polarization $\mathbb{C}^2$) with no quantum input, and a single canonical commutator with scale $\hbar$ on the mode amplitudes promotes it to single-photon QED, with photon indivisibility, the Planck relation $E=\hbarω$, and the spin spectrum $\pm\hbar$ as theorems. (ii)~The constants $c$ and $\hbar$ play non-interchangeable roles: $c$ is intrinsic to each Fourier-conjugate space, while $\hbar$ acts \emph{between} them, converting kinematic phase rates into dynamical observables. (iii)~Lifting the framework to a Haag--Kastler net with sharper-than-equal-time microcausality and positive-energy spectrum is structurally obstructed in the Galilean case; we state this as the SR-selection conjecture and identify three strands of evidence (Hegerfeldt spreading; absence of a known Galilean multi-particle resolution; Reeh--Schlieder failure on Galilean Haag--Kastler nets), the third established as a precise no-go theorem in the second paper of the series. The framework's modular content (Tomita--Takesaki, Bisognano--Wichmann, Unruh as KMS, type-$\mathrm{III}_1$ universality) is collected as the algebraic substrate on which the curved-background, dynamical-metric, and crossed-product extensions of the third through sixth papers operate.

Quantum simulation is a cornerstone application for quantum computing, yet standard methods face a trade-off between circuit depth and accuracy: Trotterization depth scales with the number of Hamiltonian terms $L$, while sampling-based qDRIFT is restricted to $O(t^2)$ error scaling. Here, We introduce qSHIFT, an adaptive sampling protocol that overcomes these limitations. By adaptively updating sampling distributions, qSHIFT maintains $L$-independent gate complexity while achieving an improved error scaling of $O(t^{1+r})$ for an adjustable parameter $r$. This performance is enabled by a classical subroutine solving $L^r$ linear equations per sampling round. Numerical demonstrations confirm the $O(t^{1+r})$ scaling, showcasing qSHIFT as a resource-efficient framework for high-precision quantum simulation. Furthermore, the protocol's reduced circuit depth enhances its compatibility with physical error mitigation, making it a promising candidate for implementation on near-term quantum devices. In addition to its role as a standalone algorithm, qSHIFT can provide a high-precision foundation for modular quantum frameworks such as qSWIFT or Krylov quantum diagonalization.

We demonstrate that ultra-relativistic black hole encounters reveal a new regime of the two-body interaction in general relativity. Evolving equal-mass, nonspinning black holes with initial center-of-mass Lorentz factors up to $γ\approx 5.1$ using numerical relativity, we find that the resulting waveforms defy the standard expectation of a post-Newtonian description followed by a smooth transition to a prompt Kerr ringdown. Instead, at nonzero impact parameter, the system can exhibit prolonged, highly irregular emission and significant horizon absorption, even without coalescence. We show these phenomena are driven by transient null trapping and repeated lensing of radiation in the binary interaction region. Furthermore, our simulations indicate that over $65\%$ of the initial ADM energy can be radiated as gravitational waves at $γ\approx 5.1$, which is substantially larger than previously estimated by extrapolating from lower boost data.

Quantum simulations of scalar quantum field theories (QFT) provide important benchmarks for demonstrating quantum advantage. We revisit digitization in the occupation basis, which is typically hindered by unfavorable circuit depth scaling. We present an approach that achieves exponential reductions in circuit depth and significantly mitigates Trotter errors by diagonalizing field operators prior to their decomposition into Pauli strings. Focusing on a scalar QFT in 2+1 dimensions, we show that this method substantially reduces circuit depth and CNOT gate counts for time evolution. Using the Lorentzian energy-energy correlator as a benchmark observable, we find parameter regimes in which occupation-basis digitization converges more rapidly with respect to local truncation than the amplitude-basis approach of Jordan, Lee, and Preskill. These results provide both algorithmic advances and phenomenological benchmarks for studies of light-ray observables on near-term quantum devices.

Axionic cosmologies constitute a class of models with phenomenologically rich symmetry breaking in the early universe. In the case where monopoles are present in such a background, the axion profile may be deformed; it is possible to construct a ``monopole bag" state composed of a central monopole within a closed axion domain wall. We consider the gravitational properties of this hybrid defect, and find a both horizon-less and a black hole-like final state can result as remnants of the monopole-domain wall system after gravitational collapse for different input parameters. We demonstrate that the latter classifies as dyonic regular black hole, evading the usual singular gravitational collapse and retaining a non-trivial axionic profile through exotic electromagnetic properties of an axionic Chern-Simons term.

Transportation systems such as urban logistics, vehicle routing, and infrastructure planning require solving large-scale combinatorial optimization problems under complex constraints. Problems such as the vehicle routing problem (VRP), traveling salesman problem (TSP), and facility location problem (FLP) involve large discrete search spaces and the need to generate multiple feasible solutions in real time. In this work, we develop a hardware-grounded hybrid quantum optimization framework that uses Approximate Quantum Compilation (AQC) to compress early segments of digitized adiabatic evolution into shallow circuits. The compressed prefix is combined with variational layers, enabling a systematic study of how initialization, circuit depth, and expressivity interact on near-term quantum hardware. All experiments are performed on an IBM gate-based quantum computer, and circuits are evaluated as stochastic generators of candidate transportation plans. Results show that moderate prefix compression reduces two-qubit gate depth while maintaining or improving feasible solution discovery, particularly for routing problems. These benefits depend on compatibility between the compressed prefix and the variational ansatz: while standard QAOA effectively leverages AQC initialization, linear-chain QAOA shows limited improvement. Overall, this work demonstrates that hybrid AQC-QAOA methods provide a practical pathway for hardware-efficient quantum optimization, positioning quantum algorithms as candidate generators within transportation decision-making workflows.

We investigate the absorption cross sections of electromagnetic perturbations propagating on a four-dimensional brane in the background of higher-dimensional regular black holes arising in quasi-topological gravity. Employing a WKB-based approach for the computation of greybody factors, we analyze the impact of higher-curvature corrections and spacetime dimensionality on the scattering properties of the system. We show that regularity effects lead to systematic deviations from the singular Tangherlini solution, manifested in shifts of the absorption spectrum and modifications of the effective photon sphere radius. Increasing the number of spacetime dimensions suppresses these deviations, driving the system toward the classical limit and reducing the number of multipoles required for convergence. In the high-frequency regime, the absorption cross section approaches the geometric-optics limit, while at low frequencies it is strongly suppressed due to diminished transmission probabilities. Our results demonstrate that the interplay between regularity and higher-curvature effects leaves distinct imprints on the absorption characteristics, providing a sensitive probe of the underlying gravitational theory.

In this work, we investigated several inflationary scenarios within the framework of modified $f(Q,φ)$ gravity with a nonminimal coupling between the scalar field and the nonmetricity scalar. We focused on the impact of the coupling parameter $ξ$ on the inflationary observables, namely the scalar spectral index $n_s$ and the tensor-to-scalar ratio $r$. In the case of De Sitter inflation, we showed that the model can reproduce observationally viable predictions only within a restricted range of the coupling parameter. Specifically, we found that $n_s$ increases with $ξ$, while $r$ decreases, leading to a narrow allowed region $10^{-3} \lesssim ξ\lesssim 10^{-2}$ compatible with Planck data. Outside this range, the model either predicts excessively large tensor modes or an unphysical blue-tilted spectrum. We also derived theoretical constraints on $ξ$ from the consistency of the model, leading to an upper bound $ξ< \fracκ{2p}$. For $κ= 1$ and $p = 60$, this implies $ξ< 0.00833$, with a preferred region around $ξ\sim \mathcal{O}(10^{-3})$. Furthermore, we analyzed the Cosh-type inflationary model and showed that it provides a robust and consistent description of inflation. In this case, the tensor-to-scalar ratio decreases while the scalar spectral index increases with the number of e-folds $N$. For $N = 60$, the model predicts $n_s \approx 0.965 - 0.967, \qquad r \approx 0.017 - 0.018$, in excellent agreement with current observational constraints. Overall, our results highlight the crucial role of the nonminimal coupling in shaping the inflationary dynamics and ensuring compatibility with cosmological observations.

One of the main difficulties in preparing many-body ground states is achieving the target state through simple counterdiabatic controls. For critical systems crossing a transition to a topological phase, this task becomes even more challenging due to the closing of the gaps in multiple symmetry sectors. This is the case of the Kitaev chain, whose transition between the trivial and topological phases involves states belonging to different symmetry sectors. In this work, we apply the recently introduced minimal action shortcut to adiabaticity (MA-STA) to a Kitaev chain and propose a multi-step strategy to obtain the optimal control protocol to drive the system across its different phases. Our results show that high fidelities can be achieved through the adapted MA-STA at time scales much shorter than those of linear ramp protocols. We also compare the performance of both controls in suppressing work fluctuations. These findings may guide the design of STA protocols in many-body systems where competing energy scales and symmetries shape the global dynamics.

We study a minimal model of collective light emission from an incoherently driven ensemble of atoms where incoherent drive is applied only to a part of atoms and show that both the linewidth and the photon statistics can be controlled within a single framework. In this setting, collective dissipation induces correlations between the pumped and unpumped parts of the system, leading to interference between their emission contributions. By introducing a relative phase between these contributions and tuning the pump rate, we demonstrate that the properties of the emitted light can be varied over a broad range. In particular, the linewidth can be made either independent of system size or scale extensively with it, while the photon statistics can be tuned from antibunched or quantum to bunched. We further show that the role of the relative phase in controlling the interference can alternatively be played by the coherent interaction. By tuning the interaction strength together with the pump rate, one can access the same regimes as in the dissipation-only model. In addition, coherent interactions stabilize regimes of coherent emission with narrow linewidth, reminiscent of superradiant lasing. Our results illustrate how interference in partially driven collective systems provides a flexible mechanism for tailoring both spectral and statistical properties of light.

Quantum Machine Learning (QML) has recently emerged as a highly promising research frontier. Within this domain, Quantum Neural Networks (QNNs),characterized by Variational Quantum Circuits (VQCs) at their core and featuring layers of quantum gates optimized by classical algorithms, have garnered significant attention. However, a rigorous and exhaustive evaluation of their practical performance remains largely incomplete. In this study, we conduct a comprehensive comparative analysis of three prominent hybrid classical-quantum architectures: Quantum Convolutional Neural Networks (QCNN), Quantum Recurrent Neural Networks (QRNN), and Quantum Vision Transformers (QViT), focusing on the critical dimensions of generalization, accuracy, and robustness. Our findings provide novel insights that address previous evaluative gaps. Notably, while these models exhibit exceptional performance on low-feature datasets such as MNIST, their learning efficacy degrades significantly when transitioned to high-feature datasets. Furthermore, convolutional-based models like QCNN appear less effective on high-dimensional data than other machine learning architectures. Additionally, while all models are susceptible to adversarial noise, traditional architectures, such as recurrent and convolutional networks, demonstrate superior resilience. Conversely, in the presence of quantum noise, the transformer-based architecture proves its strength by maintaining high robustness against measurement noise, channel noise, and finite-shot effects, whereas other architectures suffer marked performance declines. These results provide a granular perspective on the current state of the field and underscore the critical importance of tailoring model selection to the constraints of contemporary Noisy Intermediate-Scale Quantum (NISQ) environments.

Tricationic protactinium monofluoride ($^{229}$PaF$^{3+}$) has been proposed as a candidate for probing physics beyond the Standard Model of particle physics. Since studies with $^{229}$PaF$^{3+}$ require significant experimental advances, we exploit the stable, valence-isoelectronic dicationic cerium monofluoride (CeF$^{2+}$) as a surrogate. Gas-phase fluorinated-cerium molecular ions are formed and identified using the Off-Line Ion Source and TITAN mass measurement facilities at TRIUMF. Quantum chemical calculations are performed on the electronic structure of CeF$^{2+}$, revealing a parallel to that of $^{229}$PaF$^{3+}$. Moreover, these calculations provide estimates on the sensitivity of CeF$^{2+}$ itself to various $\mathcal{P,T}$-odd properties. A brief discourse on the specifics of the quantum control of CeF$^{2+}$ is presented which anticipates future searches for symmetry violations.

Can quantum computers effectively simulate the propagation and scattering of electromagnetic waves in a classical plasma? This chapter introduces some of the basic concepts in mathematics and physics essential to answering that question. The numerical simulations of Maxwell equations for wave propagation in dielectrics are constrained by technological limitations of the present-day computers. In contrast, there has been ample fanfare around quantum computers and their potential to far exceed the performance of traditional computers. Whether the enhanced capabilities of a quantum computer can be put to use for simulating topics in classical physics is a source of intrigue and curiosity.

We present a variational algorithm for fault tolerant quantum computing to solve a system of linear equations which directly maximises the parameters of the target fidelity. This so-called measurement test algorithm can be applied to any computational task with a solution that is represented as eigenvector of a self-adjoint matrix. The solution is prepared as state of a register in the quantum computer by a von Neumann measurement of a corresponding observable, which is implemented using the phase estimation algorithm. The probability to project the system thus into the unknown target state, which equals the target fidelity, is measured in terms of relative frequencies and iteratively optimised to read out the target state. The new algorithm overcomes three issues of previous variational quantum algorithms: i) It does not rely on a decomposition in terms of Pauli strings and therefore can compute eigenvectors of dense matrices. ii) The accuracy is not limited by the condition number $κ$ of the matrix, provided a logarithmic number ($O(\logκ)$) of qubits is used to encode the eigenvalues and iii) the target fidelity $F_T = 1-ε$ can be reached with an accuracy $ε$ that scales with $1/N$ for $N$ measurements per iteration. We demonstrate this by numerical simulations for dense random real-valued $16\times 16$ matrices with non-vanishing determinant.

If dark matter consists of ultralight bosons, on galactic scales it can be effectively described as a coherent classical field experiencing oscillations. Such a field could perturb the dynamics of celestial bodies via a direct coupling to ordinary matter, introducing signatures detectable through high-precision pulsar timing analysis. In this work, we extend a two-step Bayesian inference framework, originally developed for linearly coupled scalar ultralight dark matter (ULDM), to probe a quadratic scalar coupling and spin-1 vector dark matter. By explicitly marginalising over nuisance orbital parameters, our approach provides robust sensitivity limits that avoid the artificial overestimation often associated with direct fitting techniques. For quadratic scalar ULDM, we establish new constraints on the coupling $β$ in the range between $2 \times 10^{-22}$ eV and $2 \times 10^{-21}$ eV inaccessible to other experiments, while identifying mass regimes where the sensitivity is dominated by the orbital phase $Ψ'$ or the projected semi-major axis $x$. For vector ULDM, we characterize resonant signatures present even in circular orbits and obtain bounds on the coupling $g$ within the $10^{-23}$ eV to $10^{-18}$ eV range, yielding results within the same orders of magnitude as current laboratory and space-based experiments.

Zero-field (ZF) nuclear magnetic resonance (NMR) spectroscopy probes scalar J-couplings between nuclei while dispensing with large homogeneous magnetic fields, enabling low-cost and geometrically flexible detection, including through conductive enclosures. Despite these advantages, its broader use for chemical analysis has been limited by sensitivity and by the difficulty of predicting the dense spectral multiplets that arise at zero field. Here we demonstrate natural-abundance (1.1%) 13C ZF spectroscopy on off-the-shelf liquids using a compact commercial 87Rb magnetometer for the first time, without hyperpolarization or special sample preparation. Instrumental advances yield improved sensitivity, <250-mHz linewidths and >week-long stability, enabling isotopomer-resolved fingerprint spectra across a 13-molecule library, including the ability to discern rare (0.0121%) doubly 13C-labelled species. In parallel, we demonstrate vibrationally corrected density-functional theory (DFT) based prediction of ZF NMR spectra for chemically diverse molecules with few-hertz accuracy. Comparing experiment with these calculations renders residual deviations as chemically informative, reporting on hydrogen bonding, hydration and ion pairing at high ionic strength. Together, these results contribute towards DFT-assisted ZF NMR as a general platform for field-constraint-free molecular identification and for extracting transient solution-state structure from responsive J-coupling observables.

Approximate combinatorial optimization is a promising use case for quantum computers. The quantum optimization algorithms often employ a fixed ansatz that evolves an unbiased initial state towards states with better values of the optimand, then samples the states to determine an approximately optimal solution. However, promising alternative approaches have considered ``warm-start" and sampling-based methods that instead begin from the best known solution, which can be directly optimized with the quantum computer and updated as new information becomes available, potentially outperforming the fixed ansätze. Here we use these ideas to design a nonvariational quantum algorithm for combinatorial optimization. At each step the algorithm begins with a state superposed around the best known solution, then drives it to lower energy using quantum imaginary time evolution. These nonvariational, initial-state-dependent circuits are determined using analytic equations that are evaluated using only a conventional computer. After implementing the circuits, the state is sampled, potentially obtaining a new best-known solution to use as the initial state at the next iteration. Using simulations of the algorithm solving MaxCut on 3-regular graphs with 30 or fewer vertices and a shot budget of 100 total shots, the approach obtains median solutions within 95\% of the global optimum and finds optimal solutions in 11\% or more of cases, significantly outperforming random and simplified classical search procedures. We discuss several future directions.

We give a counterexample to a recently conjectured variant of the Penrose inequality.

The Euclidean formulation of quantum gravity can be interpreted in terms of a probability distribution over Riemannian manifolds. In the context of de Sitter gravity, the statistics of the total volume according to this distribution is encoded in the dependence of the partition function on the cosmological constant. We use this observation to obtain a probability distribution for the volume from known results and proposals for the de Sitter partition function, in several levels of approximation: saddle point, one loop, an all-loop and a non-perturbative proposal in 3 dimensions, and an exact result in 2 dimensions, in the context of Liouville theory. In all cases we find a reasonable behavior: in the classical limit the distribution concentrates around the classical volume, and it spreads as quantum effects are turned on. We also find as a common trend that, as quantum effects are increased, the probability distribution favors increasingly smaller universes.

Broad claims about whether adaptivity helps in quantum state tomography can be misleading unless the state family, measurement architecture, and error metric are specified carefully. We study a restricted but physically important regime: single-copy quantum state tomography under local Pauli basis measurements, where the allowed measurement settings are tensor-product measurement operators built from local single-qubit Pauli operators, and performance is measured in trace distance with high probability in a minimax sense over a known structured family. We construct an explicit discrete prefix/tree family of states for which adaptive measurement selection achieves polynomial copy complexity, while every non-adaptive design requires exponentially many copies in the worst case. The adaptive upper bound comes from stagewise prefix recovery using hierarchical breadcrumb information revealed by partial prefix matches. The non-adaptive lower bound is based on a rare-prefix mechanism: every fixed design under-samples some deep prefix subset, and outside that subset the competing hypotheses induce identical one-shot laws, so only an exponentially small fraction of the measurement budget contributes to the KL divergence between the full data distributions. The result isolates a concrete regime in which adaptivity provably changes the sample-complexity scaling under the experimentally common local Pauli measurement architecture.

Variational Quantum Algorithms, including the Quantum Approximate Optimization Algorithm (QAOA), have shown promise in solving optimization problems but rely on costly variational loops that can themselves be hard optimization problems. Many methods have been proposed to mitigate this variational cost, with one of the most common being parameter transfer and concentration where variational parameters for one problem instance or for an average over problem instances can be used as a good set of parameters for another instance. Methods exist for reweighting these parameters based off graph degree and edge weights, but there has been little work on how to do this reweighting to handle higher locality problems where the graph structure turns into a hypergraph structure. In this paper, we analytically derive parameter reweighting rules to transfer parameters between different locality hypergraphs, resulting in a reweighting for the mixing terms in the Hamiltonian which have previously not been considered. These analytics rely on three cycle-free and low-circuit-depth assumptions, but numerics indicate that the results can be used even when these assumptions are not satisfied. The numerics obtain high quality results across a diverse set of hypergraphs with locality less than or equal to five, improving on previous relations that do not reweight the mixing terms.

The cosmological constant (CC, $Λ$) problem represents a remarkable discrepancy of about 120 orders of magnitude between the observed dark energy and its natural expectation from quantum field theory. This paper synthesizes two paradigms - holographic naturalness ($\mathcal{HN}$) and pre-geometric gravity (PGG) - to propose a unified resolution. The $\mathcal{HN}$ framework posits that CC stability is not a matter of radiative corrections but of quantum information and entropy. The large entropy $S_\text{dS}\sim M_\text{P}^2/Λ$ of the de Sitter (dS) vacuum acts as an entropic barrier, exponentially suppressing destabilizing quantum transitions. This explains why the universe remains in a high-entropy, low-CC state. We embed this within PGG, where spacetime geometry and the Einstein-Hilbert action emerge dynamically from the spontaneous symmetry breaking SO($1,4$)$\rightarrow$SO($1,3$), driven by a Higgs-like field $φ^A$. Both $M_\text{P}$ and $Λ$ are generated from more fundamental parameters. Crucially, we establish a direct correspondence between the VEV $v$ of the pre-geometric Higgs field and the de Sitter entropy: $S_\text{dS}\sim v$ (or $v^3$). Thus, the field generating spacetime also encodes its information content. The smallness of $Λ$ follows directly from the largeness of $S_\text{dS}$, a manifestation of a large $v$. The CC is stable because a large-entropy state's decay is exponentially suppressed. Our study shows new semi-classical quantum gravity effects dynamically generate "hairons", particles whose mass is tied to the CC. The instability of the dS space, driven by a condensate evolution, points to a dynamical origin for dark energy. This framework inextricably links the emergence of geometry, the hierarchy of scales and the quantum-information structure of spacetime, providing a novel path toward solving the CC problem.

In this paper, we reformulate the Schrodinger equation in gauge-theoretic terms. Starting from the Madelung representation, we rewrite the conserved probability-current using gauge fields, namely a one-form gauge field in the $(2+1)$-dimensional theory and a two-form gauge field in the $(3+1)$-dimensional theory. This gives a local equivalence between the Schrodinger equation, quantum hydrodynamics, and a gauge formulation, while the global information is carried by the quantization of phase winding around zeros of the wavefunction. We then explore how this correspondence organizes structures on both sides of the duality. On the gauge side, BF couplings to additional one-forms describe electromagnetic coupling, Berry connections, spinor dynamics, projected non-abelian adiabatic connections, and intrinsic holonomy, while Chern--Simons term generate Hopf functionals, charge-flux attachment, and anyonic sectors. In the presence of boundaries, the topological terms produce edge degrees of freedom and boundary charge algebras. Finally, in the nonlinear regime with a Bogoliubov sound mode, the dual two-form description relates acoustic memory, large gauge transformations and the corresponding soft theorem, organizing them into an infrared triangle.

Given the multi-frequency nature of relativistic orbits, transient orbital resonances are expected to be ubiquitous during an extreme-mass-ratio inspiral (EMRI). At a resonance, the orbital dynamics is modified in a nontrivial way, imprinting an overall dephasing in the emitted gravitational waves and potentially impacting both the detection and parameter estimation of these sources. In this work, using a Fisher-matrix approach, we investigate the bias induced by transient orbital resonances in EMRI parameter estimation. We focus on the most dynamically significant low-order resonances, 3 : 2 and 2 : 1, as well as on the high-order, subdominant resonances 3 : 1 and 4 : 3. We find that, for most of the orbits considered, neglecting the effect of a resonance crossing leads to significant losses in signal-to-noise ratio and induces bias in parameter recovery. Furthermore, both the sign and the amplitude of the resonance-induced modifications to the integrals of motion play a crucial role and must be modeled accurately. Our results provide further evidence that failing to model transient orbital resonances accurately can hinder EMRI detection and parameter estimation, thereby limiting their scientific potential.

Quantum noise is a central challenge in quantum computing across many applications. Extensive work has examined how qubits couple to their environment, leading to decoherence and relaxation, which is irreversible. Current studies focus on coherent gate errors caused by control misalignment, which accumulate with circuit depth but can, in principle, be corrected. This work studies a continuous coherent noise model for quantum circuits and compares it with a discrete Pauli model. The focus is on small coherent gate errors that build up across circuit depth. These errors are modeled as random rotations on the Bloch sphere using a von Mises-Fisher distribution. In the small-angle limit, the model reduces to an isotropic Gaussian distribution. We test the model on quantum error-correction circuits based on the [[5,1,3]] and [[7,1,3]] codes. A variant of Grover's search circuit with different qubit counts is also examined. To enable fair comparison, we introduce a model-independent matching scheme. Pauli and continuous noise channels are aligned using the binary entropy at readout. This isolates the effect of noise structure at fixed uncertainty. An approximate analytical method for coherent error propagation is also developed. The method tracks error distributions through Clifford circuits without full Monte Carlo sampling. It reduces simulation cost while preserving accuracy for circuit-level error estimates. The approximation is validated against brute-force simulations, identifying its regime of validity with Clifford circuits and limits under error correction. Our results show that continuous coherent noise can degrade logical performance more strongly than Pauli noise. They also clarify when simplified propagation models succeed and when they break down.

We study a quantum mechanical system whose spectrum coincides with that of bilinear operators of the Sachdev-Ye-Kitaev model. The standard positivity-based quantum mechanical bootstrap is degenerate with respect to the boundary data: it does not distinguish the boundary conditions that select the SYK spectrum, and hence is insufficient to determine the eigenvalues. Instead, by considering fractional powers of operators, we obtain constraint equations that determine the spectrum without imposing positivity. The resulting roots converge to exact eigenvalues as the truncation order increases. We call this the direct bootstrap.

Topological defects play a fundamental role in the investigation of symmetries in quantum field theories. For conformal field theories in two space-time dimensions, it is possible to construct these defects using lattice models allowing ab-initio analytical and numerical computations of their characteristics. In this work, topological defects are investigated in non-unitary conformal field theories using appropriate variations of the restricted solid-on-solid models. The relevant impurity models and the corresponding defect operators are constructed for the lattice system. Numerical computations are performed for the energy spectrum, eigenvalues of the defect operators as well as thermodynamic characteristics and compared with analytical predictions. Finally, renormalization group flows between the different fixed points are analyzed using numerical methods.

Creation and manipulation of entanglement with low error is essential in quantum information systems. In practice, two-qubit entangling gates constitute a dominant error source, limiting circuit depths and performance in fault-tolerant architectures. Using a neutral-atom quantum processor, we realize entangling CZ gates with a high Rabi frequency smooth-amplitude pulse, employing state-selective readout and qubit reuse for fast calibration, and achieve state-of-the-art fidelities of 99.854(4)% which improve to 99.941(3)% upon loss postselection, with stable performance for 10 hours. We then use these low-error gates in quantum circuits with coherent atom rearrangement. We first benchmark performance by creating and disentangling cluster states, and subsequently implement scrambling circuits featuring longer-range connectivity to study non-locally entangled states generated through chaotic dynamics. These results pave the way towards deep-circuit, efficient fault-tolerant quantum computation.

Using the worldline effective field theory formalism, we derive the nonlocal-in-time conservative contributions arising from tail effects in gravitational scattering to fifth Post-Minkowskian (5PM) and tenth self-force (10SF) orders. The result features multiple polylogarithms of up to weight three. This challenging computation relies on state-of-the-art integration techniques, including a novel integration-by-parts algorithm: the Sparse Integral Reducer (SpideR). We find perfect agreement in the overlap with all existing literature through sixth post-Newtonian order. The results presented here provide a key ingredient for isolating the local-in-time component of the conservative two-body dynamics of binary inspirals at 5PM order.

Generation of highly non-classical quantum states of light is essential for optical quantum information processing and quantum metrology. Given the lack of sufficiently strong nonlinear interactions between optical fields, the commonly employed optical quantum-state preparation schemes are conditional, based on nonlinearity induced by heralding photon number measurement on a part of a multimode squeezed Gaussian state. Development and optimization of such probabilistic quantum-state engineering schemes represents one of the central challenges in current quantum optics. As technology advances and experiments progress to detection of higher numbers of photons, the maximization of the heralding probability becomes essential to ensure sufficiently high state-preparation rates. Here, we show that for the conditional quantum state preparation schemes based on Gaussian states and photon number measurements the maximization of the heralding probability can be formulated as finding solution to a system of polynomial equations, which offers an efficient way to find the optimal configuration and allows us to apply techniques dedicated specifically to solving such systems of equations. Our approach can seamlessly incorporate bounds on the available single-mode quadrature squeezing, which is highly experimentally relevant. We mainly consider generation of finite superpositions of Fock states but show that the approach can be straightforwardly extended to generation of squeezed superpositions of Fock states. We focus on Gaussian states with vanishing coherent displacements, hence the conditionally generated states have well defined photon number parity. We illustrate our general methodology on examples of generation of single-mode and two-mode states with two heralding modes.

Continuous-variable quantum systems are promising candidates for quantum computing and quantum information processing. It is widely known that quadrature measurements on Gaussian continuous-variable systems can be described by a noncontextual hidden-variable model and cannot violate a Bell inequality. Here, we demonstrate that the observation fails when sequential measurements are involved. Our experiment is realized by mapping the spatial modes of a single photon, deterministically generated from an InAs/GaAs quantum emitter, to the logical operations in the Gottesman--Kitaev--Preskill code space. Employing a black-box-style approach, we observe a violation of the Bell-like noncontextual hidden-variable inequality by 380 standard deviations. Our results address the conceptual loopholes in previous works and open up new possibilities for studying fundamental quantum physics using photonic-encoded continuous-variable systems.

Quantum computing calibration depends on interpreting experimental data, and calibration plots provide the most universal human-readable representation for this task, yet no systematic evaluation exists of how well vision-language models (VLMs) interpret them. We introduce QCalEval, the first VLM benchmark for quantum calibration plots: 243 samples across 87 scenario types from 22 experiment families, spanning superconducting qubits and neutral atoms, evaluated on six question types in both zero-shot and in-context learning settings. The best general-purpose zero-shot model reaches a mean score of 72.3, and many open-weight models degrade under multi-image in-context learning, whereas frontier closed models improve substantially. A supervised fine-tuning ablation at the 9-billion-parameter scale shows that SFT improves zero-shot performance but cannot close the multimodal in-context learning gap. As a reference case study, we release NVIDIA Ising Calibration 1, an open-weight model based on Qwen3.5-35B-A3B that reaches 74.7 zero-shot average score.

There is widespread interest in many-body quantum systems that exhibit limit-cycle or time-crystalline behaviour. An ideal quantum limit cycle would be realized using fully coherent driving (to minimize noise) and also have a continuous internal symmetry (to ensure generation of monochromatic radiation). While these two requirements may seem incompatible, we introduce in this work a large class of multi-mode bosonic limit cycle models based on coherent parametric driving which possess an O(N) continuous symmetry. Surprisingly, the full quantum dissipative steady state of these models can be found exactly. They exhibit rich physics, including steady state entanglement, reduced phase diffusion and the possibility of realizing quantum limit tori. The basic mechanism we identify provides a unified way to understand how coherent parametric driving can yield symmetry-enriched limit cycles, and also helps us understand related models where the relevant symmetries are weakly broken. The models we study are compatible with a range of different experimental platforms, including quantum optical setups and superconducting quantum circuits.

Distributed Quantum Computing (DQC) and Quantum Error Correction (QEC) rely on dynamic circuits that include Mid-Circuit Measurements (MCMs) and classical feedback. These operations present a major bottleneck: MCMs suffer from high error rates that lead to real-time branching errors, while MCM and classical feedback latencies amplify decoherence errors. Current hardware controllers, qubit-state discriminators, and software error mitigation techniques fail to address these challenges holistically. We propose MCMit, a hardware-software co-design to mitigate branching and latency-induced errors. MCMit introduces a scalable, constant-latency multi-control branch instruction for faster classical feedback and two qubit-state discriminators, a transformer, and a CNN, with high accuracy even under short measurement durations. On the software side, static MCM elimination and stochastic branching complement the hardware by mitigating residual branching errors that persist despite hardware improvements. We implement MCMit on Qubic and evaluate it using experimentally extracted QPU readout traces. Our branch instruction reduces feedback latency by up to 70\%, improving circuit depths by up to $7\times$ over Qubic. Our CNN discriminator achieves 37-73\% higher accuracy for short measurement durations than the baselines, leading to up to 80\% lower logical error rates in QEC. Last, our software mitigation improves fidelity by 18--30\% over baseline methods.

The decomposition of complex quantum operations into experimentally feasible gate sets has been a central challenge since the early development of quantum computing. The multi-controlled Toffoli (MCT) gate is a key example, with applications across a wide range of quantum algorithms, whose decomposition into smaller gates, however, typically leads to deep circuits. In this work, we introduce a teleportation-based decomposition that implements an arbitrary MCT gate with unit Toffoli depth, independent of the number of controls, while maintaining a relatively low Toffoli count compared to existing approaches. This is achieved at the cost of a linear overhead in ancilla qubits and the ability to distribute entangled pairs across distant qubits, a capability already available in several quantum computing platforms. We further demonstrate the advantages of this implementation in circuits that rely on MCT gates, such as the adder operator, quantum read-only memory, quantum neurons, and quantum decision trees.

This paper investigates the interplay between the properties of quantum states on the Hilbert space \(\ell_2(κ)\) and the set-theoretic nature of the cardinal $κ$. We focus on the existence of singular $σ$-additive states~ -- functionals whose induced measures are $σ$-additive yet vanish on singletons. While the existence of such states is known to be equivalent to the Ulam measurability of $κ$, their structural and dynamical properties remain largely unexplored. We prove that any $σ$-additive state on the diagonal algebra is representable as a Pettis integral over a singular $σ$-additive measure, extending the classical representation theory to the non-normal sector. Furthermore, we construct a class of quantum channels using $σ$-complete ultrafilters that map normal states to singular $σ$-additive states, effectively <<archiving>> information into the singular part of the state space.

We construct an analytical black hole solution in the Einstein-Maxwell-Yang-Mills theory with a conformally coupled scalar field in four dimensions, which generalizes the charged Martínez-Troncoso-Zanelli (MTZ) black hole in the presence of self-gravitating non-Abelian gauge fields. The internal gauge group is determined by the horizon curvature, becoming $SU(N)$ in the case of positive curvature and $SU(N-1,1)$ when the curvature is negative. Moreover, this solution is employed as a conformal seed to obtain new meronic spacetimes dressed with all (super-)renormalizable contributions of the scalar field, which provides the generalization of the Anabalón-Cisterna (AC) solution when self-gravitating non-Abelian gauge fields are included. Finally, we consider the non-Noetherian extension of the conformal scalar fields, which still yields a second-order conformally invariant scalar equation, even though the action is not. In that case, we show that static black hole solutions can also be charged with Yang-Mills fields.

Partial differential equations (PDEs) are fundamental across numerous scientific fields. As these problems scale to high dimensions, classical numerical schemes introduce severe computational bottlenecks, known as the curse of dimensionality. Attempts to solve this problem typically rely on either classical sparsity and low-rank decompositions, or neural network surrogate models. On the other hand, Quantum Computing offers a promising alternative, as it allows us to operate in significantly larger spaces while demanding far fewer resources. In this work, we present a quantum subroutine to solve second-order linear PDEs by exploiting the structural properties of the filter in Fourier space using Quantum Block Encoding (QBE) with quantum reversible arithmetic. This approach serves as a specialized alternative to standard quantum matrix inversion, which typically relies solely on Quantum Singular Value Transformation (QSVT) without exploiting the inherent structural properties of the matrix. We validate the proposed methodology against its classical counterpart to prove its correctness. This framework provides a foundation for extending these methods toward quantum group Fourier transforms, wavelet-based analysis, and equivariant quantum neural networks (EQNNs), offering a promising path toward solving broader classes of problems, including nonlinear PDEs.

Universally robust dynamical decoupling (UR$n$) sequences were proposed to compensate pulse imperfections arising from arbitrary experimental parameters while achieving high-order error suppression with only a linear increase in the number of pulses. Although their performance was supported by analytical arguments, numerical simulations, and experiments, a complete mathematical proof of the claimed order of error compensation has been absent. In this work, we present a rigorous proof for UR$n$ DD sequences with even $n$. Using a series expansion of a quantity whose modulus is the fidelity $F$, we derive necessary and sufficient conditions for the cancellation of its coefficients up to, but not including, order $n$. The UR$n$ phase prescription satisfies these conditions, and therefore $1-F=O(ε^n)$. Our results establish the UR$n$ construction on firm analytical grounds and clarify the structure responsible for its high-order robustness.

The Nakajima-Zwanzig projected Liouvillian QLQ, the generator of the exact memory kernel in open quantum dynamics, is manifestly non-Hermitian yet has been reported to possess a purely real spectrum in the Jaynes-Cummings model -- an anomaly unexplained since observation. We resolve this anomaly by showing that QLQ is pseudo-Hermitian in the Mostafazadeh sense: a positive-definite metric eta>0 exists such that (QLQ)^dag eta = eta (QLQ), forcing the spectrum to be real. The pseudo-Hermiticity is genuine: the Delta N = 0 and Delta N = +/-2 sectors are individually non-Hermitian (residuals 1.70 and 5.06, respectively), yet the global spectrum is protected by eta. The metric survives the bath-truncation limit (N_max = 3--20, matrix dimension up to 1764 x 1764) with intertwining residual <10^{-11}. A continuous deformation to the full Rabi model reveals a re-entrant pseudo-Hermitian phase with two exceptional-point boundaries, in which the metric condition number diverges. The result supplies a structural reason for Hardy-space analyticity of the memory kernel in the canonical quantum-optical model.

The relation between uniformly accelerated laboratories and laboratories supported in a gravitational field lies at the conceptual core of the Equivalence Principle, yet its precise kinematical content beyond strictly local considerations remains subtle. In this work we develop a unified metric description of these configurations using the standard Arnowitt-Deser-Misner (ADM) formulation of General Relativity, which provides an explicit decomposition of spacetime into spatial hypersurfaces and their temporal evolution. In this setting the ADM shift vector is interpreted as a physical quantity encoding the kinematical relation between spatial slices and their temporal embedding associated with a chosen foliation. This interpretation allows uniformly accelerated laboratories and laboratories supported in a gravitational field to be described within a common structural framework, showing that configurations experiencing identical proper acceleration share an equivalent local shift structure. This viewpoint clarifies the apparent asymmetry between the spatial displacement and energetic cost associated with accelerated motion and their apparent absence in phenomenological descriptions of observers supported in a gravitational field. The formulation remains fully equivalent to standard General Relativity at the level of the field equations and constraints while making explicit kinematical features that are usually implicit in its geometric description. Consequences of this interpretation include a relational account of gravitational time dilation and the emergence of observer-dependent horizons.

In this study, by utilizing the constructed generalized free energy alongside the Mean First-Passage Time and the Kramers escape rate from stochastic dynamics, we have obtained a comprehensive landscape of the phase transitions for the Bardeen-AdS-class black hole. This black hole model admits two distinct categories of solutions. Type I black holes feature a regular black hole solution, and Type II black holes possess a vacuum state solution. In the phase transition between the small black hole and the large black hole for Type I, the process may pass through a stable, metastable, or unstable regular black hole as an intermediate state. In contrast, for Type II black holes, the phase transition occurs exclusively between the vacuum state and the small black hole, and the transition process does not involve any regular black hole intermediate states.

As emerging quantum architectures evolve into heterogeneous networks combining different physical substrates, such as qubits for logic and higher-dimensional qudits for robust communication, the traditional scalar metrics of quantum error correction become insufficient. To address this, we introduce a mathematical framework based on dimension multisets to characterize quantum error-correcting codes (QECC) and absolutely maximally entangled (AME) states in mixed-dimensional Hilbert spaces. By replacing scalar weights with multisets, we accurately capture the exact physical composition of error supports across these diverse systems. Our central result is the mixed-dimensional quantum MacWilliams identity, which establishes the formal algebraic relationship between Shor-Laflamme enumerators and unitary weight enumerators. From this foundation, we deduce the mixed-dimensional shadow identity and derive rigorous, generalized constraints on code parameters, explicitly formulating the mixed-dimensional quantum Hamming, Singleton and Scott bounds, and developing a linear program to systematically evaluate code viability. For the Singleton bound, a tighter bound that has no homogeneous analogue is derived for pure mixed-dimensional codes. Finally, we deploy this enumerator machinery to thoroughly analyze AME states, utilizing shadow inequalities to constrain their existence and introducing a combinatorial grid method for the explicit construction of mixed-dimensional tripartite AME states.

Quantum optimal control methods are widely used to design experimental control pulses such as laser amplitudes, phases, or detunings, that implement a target unitary evolution. In practice, what makes a pulse "good" depends not only on its fidelity, but also on the experimental setting and the relevant hardware constraints. Here, we introduce geometric quantum control with kernel optimisation (GECKO), a model-agnostic method for improving control pulses after a high-fidelity solution has been found. GECKO uses the Riemannian geometry of the special unitary group to identify directions in pulse space that leave the implemented unitary unchanged to first order, allowing one to traverse level sets of the control landscape while optimising a chosen differentiable pulse-quality function. We demonstrate GECKO on a transverse-field Ising Hamiltonian implementing CZ and CNOT gates, optimising pulse properties including spectral filtering, smoothness, robustness to parameter deviations, and pulse duration. In all cases, GECKO finds substantially improved pulse solutions.

The Quantum Approximate Optimization Algorithm (QAOA) follows a single, fixed evolution path, overlooking the potential computational advantage of coherently superposing multiple trajectories. Here we overcome this limitation with a hybrid quantum walk (HQW) ansatz that super poses multiple Hamiltonian-driven paths coherently within each circuit layer via a dynamical coin operator. QAOA emerges as a special case of this framework with a static Pauli-X coin. Using Pontryagin's minimum principle, we derive the optimal form of the coin operator, demonstrating that it generally differs from a constant gate. A dynamical Lie algebra analysis reveals that HQW generates a strictly larger Jordan-Lie algebra, providing an algebraic foundation for its enhanced expressivity. Especially, we reveal the connection between the unique Jordan product negativity in HQW's DLA and its performance advantages. Numerical experiments on Max-Cut and Maximum Independent Set problems show that HQW systematically outperforms QAOA in convergence speed, solution accuracy, and robustness. Our work establishes a path-superposition paradigm for quantum optimization, combining optimal control theory with algebraic structure to guide the design of advanced quantum algorithms.

Domain wall annihilation can leave a distinctive imprint on the induced gravitational wave spectrum. During annihilation, most of the domain wall energy transforms into the scalar field responsible for the initial $\mathbb{Z}_2$ symmetry breaking that created the walls, along with any coupled species. If the produced scalar is sufficiently long-lived, its delayed decay drives an early matter-dominated phase following domain wall annihilation, significantly amplifying induced gravitational waves from primordial perturbations. The subsequent transition to radiation domination dilutes the domain wall contribution through entropy injection while preserving the enhanced induced signal. This creates a gravitational wave spectrum with two distinct peaks detectable across complementary frequency bands. We explore the observable parameter space and demonstrate how multi-band detection can probe early universe symmetry breaking.

We find the ground-state energy of the Ising model using the Cascaded Variational Quantum Eigensolver (CVQE) algorithm with the Guided-Sampling Ansatz (GSA) using up to 63 qubits on a quantum computer. We study a heavy-hex lattice to match the qubit architecture, allowing us to perform calculations in the quantum utility regime. We study both a homogeneous and random-coupling model. We locate the boundary of acceptable quantum errors as a function of the number of qubits and coupling strength. An entropic analysis is performed giving insights into the quantum computing performance. A subspace analysis is performed that suggests that the Ising model is especially suited for near-term quantum computing.

We compare four polynomial-resource measurement strategies, (I) $Z$-basis-only, (II) nearest-neighbor $ZZ$ (NN), (III) multi-basis ($Z$, $X$, $Y$), and (IV) classical shadows, for classifying three quantum circuit families: IQP, Clifford, and Clifford$+T$. We find $Z$-only measurements outperform multi-basis and classical shadows across all qubit counts and all four classifiers evaluated, and the $O(\nqubits)$-feature NN strategy matches $Z$-only to within $0.02$ in Random Forest accuracy. The best result is a Random Forest accuracy of $0.91$ at 4--5 qubits under $Z$-only ($0.89$ for NN, $0.85$ for multi-basis, $0.67$ for shadows). All four strategies collapse to near-chance accuracy ($\approx 0.33$) above approximately 12 qubits under the quadratic shot budget $\shots = 16\nqubits^2$. These findings indicate that the discriminative signal between these circuit families is concentrated in local, nearest-neighbor $Z$-basis correlations, consistent with the diagonal gate structure of IQP circuits, and that additional Pauli correlator types or long-range correlations carry no compensating discriminative power for this task. We provide a formal theoretical framework showing that circuits with high diagonal fraction in a given basis concentrate their correlator structure in that basis, and that any deviation from the dominant basis incurs a provably higher estimator variance. These results establish that a quadratic shot budget is insufficient for reliable classification above approximately 12 qubits, but do not rule out the existence of a subquadratic or otherwise more efficient polynomial-resource strategy; whether any polynomial measurement protocol can classify these families at large qubit counts remains an open question.

We propose a general method to fully characterize a classical stochastic noise process causing qubit dephasing through repetitive Ramsey interferometry measurements (RIMs) on the qubit. Compared to filter-function-based spectroscopy, our method does not require complicated dynamical decoupling pulses and can directly detect arbitrary-order correlation functions of such noise processes. We show that each RIM with a short evolution time and suitably chosen control pulses can perform a direct sampling of the noise field and the $n$-point correlations of the RIM outcomes are proportional to the $n$-point correlation functions of the noise processes. Then we numerically demonstrate this method for characterizing two typical examples of classical noises, including the Ornstein-Uhlenbeck processes producing Gaussian noises and an ensemble of TLFs producing non-Gaussian noises. Our method is independent of qubit lifetime and robust against qubit decoherence and measurement errors, thus offering a universal and efficient protocol for qubit noise spectroscopy across diverse platforms.

Entropic uncertainty relations are universal quantifiers of fundamental uncertainties of quantum measurements and are widely discussed in the quantum metrology literature. Quantum memory is a phenomenon related to the specific type of quantum correlations that allows for reducing fundamental uncertainties of quantum measurements. In the present work, the modified concept of quantum memory for time-dependent problems is proposed. We compare the time-dependent formulation of quantum memory with the out-of-time-ordered correlator (OTOC). Quantum memory is a rigorous mathematical concept that requires demanding calculations. Thus, until now, quantum memory has been discussed mainly for simple model systems and stationary problems. In the present work, we demonstrate that quantum memory can also be studied for realistic and physically relevant systems, e.g., the atomic helical spin chain, as well as the emergence and propagation of quantum correlations in time. We found that quantum memory manifests faster oscillations in time than OTOC and does not equilibrate. Furthermore, an artificial neural network is trained and asked to predict results for OTOC and quantum memory. These results show that quantum memory is more sensitive than OTOC in terms of broken inversion symmetry and the nonreciprocal effect.

To meet the growing demand for nanoscale surface analysis, nitrogen-vacancy (NV) centers offer a high-sensitivity alternative by leveraging their ability to operate in immediate proximity to the sample. In this work, we propose a quantum control protocol designed to overcome the inherent challenges of solid-state environments, specifically by mitigating anisotropy and strong dipole-dipole interactions to enable the detection of isotropic chemical shifts at the nanoscale. To achieve this, our scheme synchronizes a slowly rotating magnetic field with tailored RF decoupling and MW control of the NV sensors. We provide an analytical mapping that explicitly links the measured spectrum to the control sequence features and the underlying system parameters, enabling a straightforward characterization of the sample.

Efficient quantum state preparation is a critical component in quantum algorithms that process large classical data, and it is fundamental to realizing quantum advantage in domains such as machine learning, quantum linear algebra, and quantum finance. Building on the framework of~\cite{berti2025efficient}, which integrates Bucket Brigade QRAM (BBQRAM) with a segment tree to achieve amplitude encoding in polylogarithmic query time, we present two improvements within the same architecture-aware framework. First, we remove the $U_{2\mathrm{CR}}$ subroutine by classically precomputing the rotation angles determined by the segment tree and storing these angles directly in the BBQRAM cells. The tradeoff is that the classically loaded QRAM stores precomputed fixed-point angles rather than raw subtree weights. Second, we extend the construction to complex-valued matrices $A \in \mathbb{C}^{M \times N}$ by storing a leaf phase alongside each precomputed rotation angle and using a two-step magnitude-then-phase procedure; the real signed case is naturally subsumed as a one-bit phase specialization. At unchanged $\mathcal{O}(\log_2^2(MN))$ BBQRAM query complexity, the QPU procedure reduces to BBQRAM retrievals and controlled-rotation cascades, with $\mathcal{O}(MN)$ memory cells per matrix and no reversible arithmetic on the QPU.

A key bottleneck in quantum machine learning is the computational cost of repeated quantum circuit evaluations during the inference phase. To address this, we present a framework for constructing fast, cheap, provably accurate classical tensor-train surrogates of fully trained quantum machine learning models within local patches of their input data space. The approach combines Taylor polynomial approximation with a tensor-train (TT) representation and embeds it in a statistical learning paradigm via empirical risk minimization. In our analysis, the Taylor-TT construction serves as a deterministic error certificate proving that the TT hypothesis class contains a good approximation; empirical risk minimization then provably recovers a surrogate with controlled generalization error and explicit bounds. This translates into three independently controllable error sources: (i) Taylor truncation error controlled by the patch radius $r$ and polynomial degree $p$, (ii) TT approximation error controlled by the bond dimension $χ$, and (iii) statistical estimation error. While the parameter count scales polynomially in the number of data dimensions $N$, i.e., $d_{\mathrm{eff}} = N(p+1)χ^2$ rather than the naive $(p+1)^N$, the worst-case constants inherit an exponential factor through the tensor-product feature norm during Taylor polynomial embedding onto TT. This cleanly separates representation complexity from feature-induced constants. Our risk bounds and sample complexity depend explicitly on the local patch radius $r$.

We investigate the thermodynamics, criticality, and selected radiative and optical properties of an Euler-Heisenberg AdS black hole surrounded by quintessence, perfect fluid dark matter, and a cloud of strings. Within the extended phase-space formalism, we derive the thermodynamic quantities, verify the modified first law and Smarr relation, and analyze the corresponding equation of state and critical behavior. We show that the Euler--Heisenberg coupling and the surrounding matter fields substantially modify the temperature profile, the stability structure, and the location of the critical point. We also examine the sparsity of Hawking radiation, together with the photon sphere, black hole shadow, and the associated geometric-optics emission rate.

The security of quantum key distribution (QKD) systems relies on the physical integrity of their components. While laser-damage attacks (LDAs) using high-power continuous-wave (cw) lasers have been well studied, the threat posed by pulsed lasers at alternative wavelengths remains underestimated. Here, we experimentally investigated the stability of four types of fiber-optic attenuators under exposure to sub-picosecond pulses at 1061 nm with average power reaching 1 W. Mechanical variable attenuators with blocking elements and fixed air-gap attenuators show resistance to this attack. MEMS-based variable attenuators exhibit increased attenuation or irreversible damage that causes a permanent reduction in attenuation of approximately 3.8 dB. For fixed attenuators with an absorption element, we demonstrate that initial pulsed irradiation significantly lowers the optical damage threshold of the components compared to direct cw attacks. The attenuation reduction achieved is up to 7 dB at a 1 W cw laser at 1550 nm. These results highlight the possibility of establishing a hidden side-channel for eavesdropping attacks and underscore the insufficiency of existing countermeasures against sophisticated LDA scenarios.

Scalable optical quantum technologies require interference between large numbers of indistinguishable single-photons emitted by independent sources. Semiconductor quantum dots are known to be excellent on-demand sources of single-photons. They show record efficiency when inserted into optical cavities to control their spontaneous emission and generate trains of near identical photons over microsecond timescales. However, generating perfectly identical photons from distant cavity-based sources has remained a long-standing challenge. It requires precise matching of the emission wavelengths and emission dynamics, while simultaneously minimizing spectral noise across all time scales for distant emitters in uncorrelated environments. Here, we report on the nanofabrication of a large number of quantum dot-cavity sources with ultra-low spectral noise and wavelength dispersion. The high source efficiency and the use of two tuning mechanisms enable precise optimization of the spectral overlap between distant sources. With this approach, we demonstrate a two-photon indistinguishability of $88\pm1$ % between photons emitted from two distant sources. Remarkably, this value reaches the upper bound set by the intrinsic indistinguishability of photons emitted successively by each source. These results represent a key milestone for scaling photon-based quantum technologies.

In this paper, we resolve an open question in the field of optimization algorithms for training parametrized quantum circuits: Does the popular Rotosolve algorithm converge? Until now, interpolation-based coordinate descent methods such as Rotosolve have mostly been treated as heuristics, lacking any formal convergence guarantees. We rigorously analyze Rotosolve, and show that it converges to $\varepsilon$-stationary points if the optimization landscape is non-convex and smooth; and to $\varepsilon$-suboptimal points if the objective function additionally obeys the Polyak-Lojasiewicz (PL) condition. Further, we derive explicit worst-case rates of convergence in the finite quantum measurement regime. These rates are contrasted against those from a similar coordinate-based method: Randomized Coordinate Descent (RCD). Although in the worst case their rates are, prima facie, equivalent, we present arguments for a more nuanced comparison between the two. We highlight that Rotosolve is hyperparameter-free, and implicitly uses first and second derivatives in its updates. Finally, we supplement our theoretical findings with numerical experiments from Quantum Machine Learning; and compare the performance of Rotosolve against RCD, Stochastic Gradient Descent, Simultaneous Perturbation Stochastic Approximation, and Randomized Stochastic Gradient Free methods.

Artificial intelligent language-model based coding agents have significantly changed the way we interact with computers in our day-to-day, as it is common to use them to create, improve, and run programming scripts only using natural language. Agent code updates can be better guided when such programs can be executed and scored automatically rather than judged by human preference. In quantum computing and classical quantum simulation settings, ground-state preparation has a parallel structure: candidate protocols can be ranked by estimated energies and other proxies indicating proper quantum-state convergence. In this work, we study how autoresearch, a code optimization strategy based on coding agents, can be used to optimize hyperparameter choices of different ground-state preparation and sampling protocols, including the variational quantum eigensolver (VQE), density matrix renormalization group (DMRG), and auxiliary-field quantum Monte Carlo (AFQMC). We validate the viability and capacity of this method on simple spin models and molecular Hamiltonians. Across all three settings, the agent mutates simple baselines into complex protocols with improved energy proxies while operating under constrained space-time computational budgets. We conclude with discussions of other quantum routines that support executable scalar scoring, enabling evolutionary coding agents to automate a substantial portion of the protocol-tuning work that would otherwise be required manually.

We explore the gravitational-wave phenomenology of equal-mass inspiralling boson-star binaries using numerical relativity simulations. In particular, we characterise the waveform differences between binary boson-star and black-hole systems across (i) the early inspiral, by matching our waveforms to post-Newtonian expressions, (ii) merger, and (iii) late ringdown, by extracting the quasi-normal mode frequencies of the remnants. We find that boson-star binaries exhibit the largest deviations from comparable binary black-hole systems during the late inspiral and merger phases. Remarkably, for a subset of these equal-mass boson-star binaries (with certain phase offsets in the scalar-field profiles) we identify the excitation of subdominant odd $m$-multipoles in the gravitational-wave emission, absent in equal-mass nonspinning black-hole binaries. Despite differences in the phenomenology of binary boson-star and black-hole signals, injections of some boson-star signals into detector noise exhibit degeneracy with current waveform approximants. Building on these results, we demonstrate how inspiral-merger-ringdown consistency tests can overcome these degeneracies.

We study algorithms inspired by quantum annealing that are suited for the NISQ era. First, we analyze approximate quantum annealing (AQA), which employs a discretized annealing ansatz in which the time step and the number of layers are allowed to deviate from a faithful implementation of quantum annealing. Parameter scans identify regimes that reproduce annealing-like behavior with reduced resources, making them more suitable for NISQ devices. The resulting parameters can then be used as an effective warm start for the quantum approximate optimization algorithm (QAOA), improving its performance compared to random initializations. We also introduce evolving Hamiltonian quantum optimization (EHQO), a multistep variational scheme that guides the optimization process through intermediate Hamiltonians derived from the standard annealing Hamiltonian. Numerical simulations on sets of hard 2-SAT instances suggest that quantum annealing-inspired algorithms provide practical strategies for enhancing variational quantum optimization.

The type IIB matrix model has been proposed as a nonperturbative formulation of superstring theory. While numerical simulations of this model are essential for probing nonperturbative effects, such as the emergence of time and an expanding 3--dimensional space, they are hindered by the sign problem. We address this using the Complex Langevin Method (CLM). Furthermore, to suppress spurious numerical artifacts that originate from large Lorentz boosts due to the Lorentz symmetry of the model, we nonperturbatively fix the Lorentz symmetry using the Faddeev--Popov procedure. We then study this model to investigate the impact of supersymmetry on the dynamical generation of (3+1)--dimensional spacetime.

In this work, we developed analytic asymptotic methods for computing the Fourier modes of gravitational waves from post-Newtonian binary systems in the quasi-Keplerian parametrization in the high eccentricity regime. We have also derived the large-eccentricity asymptotic expansion of the eccentricity enhancement function appearing in the tail contributions to the radiation. Furthermore, based on these results, we constructed an endpoint-constrained analytic approximation that significantly accelerate the computation of the Fourier modes at large eccentricity.The overall error of this analytic approximation is controlled within $10^{-3}$, and it remains valid for Fourier modes with $p\le200$. This approach provides an analytic building blocks for modeling frequency-domain gravitational wave from highly eccentric binaries.

We identify a missing local-refinement stage in the cotengra tensor-network contraction pipeline and show that its impact grows monotonically with bond dimension on the \emph{connectivity graph} of Sycamore-like topologies. Appending a nearest-neighbor interchange (NNI) search to the \cotengra{} output at matched 8-s wallclock yields a median \emph{predicted} cost-model gap $Δ\fT$ at $n{=}500$ that grows monotonically and approximately linearly in $χ$, from $\sim\!15$~bits at $χ{=}2$ to $\sim\!116$~bits at $χ{=}16$ (Fig.~\ref{fig:chi_sweep}), with the refiner winning on $25/25$ seeds at every tested $χ$. Two control families -- random $3$-regular and QAOA $p{=}2$ interaction graphs -- show median $|Δ\fT| \leq 0.71$~bits across both controls at every $χ$, with refiner win rate falling toward chance as $χ$ grows; the signal is topology-specific, not a generic refinement-budget effect. An ablation establishes that refinement itself, not the four-axis Pareto acceptance rule, drives the gain ($|Δ\fT| \lesssim 0.1$ bits between scalar and Pareto arms at $χ{=}2$). The Sycamore-circuit envelope (App.~\ref{em:sec:results:syccirc}) reports the corresponding refinement on actual random circuits at depths $m \in \{4, 6, 8, 10, 12\}$, where the refiner wins on $5/5$ instances at every depth. The advantage is therefore largest precisely in the bond-dimension regime relevant to physical contraction.

Defect-adaptive surface-code methods have substantially advanced the construction of valid logical patches on imperfect hardware, but fault-tolerant computation also requires executable logical oper ations on the resulting irregular geometries. We formulate the seam-boundary defect problem: how to perform a lattice-surgery merge when the intended seam intersects deformed boundaries, disabled checks, and gauge-inferred super-stabilizers. We introduce a defect-adaptive lattice-surgery method that reconstructs the target joint logical parity from the seam-related measurements available on the irregular merged patch, together with constraints inherited from the separated pre-merge code space. The reconstruction is expressed as a compact GF(2) binary-support synthesis problem. If the requested parity is realizable, the solution gives an executable parity-extraction rule over raw, schedule-tagged gauge outcomes; otherwise, it certifies a parity-synthesis failure rather than conflat ing it with patch invalidity. The framework accommodates boundary data-qubit defects, seam-check ancilla defects, and gauge-inferred seam super-checks within a single synthesis layer. Circuit-level samples of the synthesized merge operation show improved compile yield, preserved effective dis tance, and only modest success-conditioned logical-error overhead relative to the defect-free merge reference; an explicit ZZ-merge sampling check confirms the expected transposed-geometry behav ior under the same success-conditioned observable construction. More broadly, the results identify certified parity synthesis as a compilation layer between defect-adaptive patch construction and executable fault-tolerant logical operations on imperfect surface-code hardware.

Simon's algorithm is a polynomial period-finding algorithm that has been used to exploit the algebraic structure of specific symmetric ciphers, showing that exponential speedups in their cryptanalysis are theoretically possible. While the theoretical framework for an attack using Simon's algorithm on the Even-Mansour cipher is well-established, practical implementations on noisy intermediate-scale quantum (NISQ) hardware remain limited. This paper presents a proof of concept quantum cryptanalysis of the Even-Mansour cipher using Simon's period-finding algorithm on NISQ hardware. For N = 3 and N = 4, we successfully demonstrate secret key recovery for N-bit constructions on the ibm_miami processor. Our experiments also identify a scaling limitation in the classical pre-processing stage: The DORCIS circuit optimization tool encountered a memory bottleneck at N = 5, preventing the generation of optimized circuits for larger key lengths. Our results suggest firstly that Simon's algorithm is effective for the Even-Mansour cipher for short bit lengths on current quantum hardware. Secondly, while DORCIS is effective for the small-scale S-boxes for which it was designed, there remains a need for the investigation of more scalable and efficient synthesis tools capable of handling larger and more general permutations in the context of Even-Mansour ciphers.

Bent Boolean functions extremal objects that maximally resist affine approximation are notoriously hard to construct for large numbers of variables. We propose a hybrid quantum-classical genetic algorithm (GA) that uses a \emph{quantum circuit} to evaluate the Gowers $U_2$ norm as the evolutionary fitness function. Our central contribution is a complexity-theoretic separation: the quantum evaluation circuit requires only $3n$ qubits and $\bigO(n^2)$ two-qubit gates per function query, whereas the classical computation of the exact Gowers $U_2$ norm demands $\bigO(2^{2n})$ arithmetic operations an exponential overhead that renders it infeasible for $n \gtrsim 25$. We validate the framework on $n=6$ and $n=8$ variable systems. For $n=8$, our classical GA run extended to 1000 generations achieves best fitness $\Utwof = 0.250000$ \emph{exactly} the theoretical bent threshold $2^{-n/4}$ with average fitness $0.257267$, confirming that the Gowers $U_2$ norm is a superior fitness criterion over Walsh-Hadamard spectral flatness. Quantum-assisted evaluation faithfully reproduces the classical trajectory up to finite-sampling noise, and our complexity analysis demonstrates that for $n > 25$ the quantum evaluator provides a decisive computational advantage on fault-tolerant hardware.

We study finite-size adiabatic state preparation on Boolean hypercubes using graph-local drivers built from sector/path coordinates related to monotone Gray-code representatives. The construction is not presented as a new all-$n$ Gray-code existence theorem; rather, it provides finite representatives, explicitly checked through the cases used in the numerical experiments, for testing problem-dependent graph-local drivers. For ordinary diagonal-cost transverse-field annealing, the ordering does not yield a robust advantage, and we include this negative result as a baseline. For non-diagonal target Hamiltonians whose geometry is expressed in the same sector/path coordinates, hybrid drivers combining sector, path-window, and small transverse-field components can substantially improve the final ground-state fidelity in centered barrier instances. Reproduction runs from the accompanying code confirm a representative centered original-window barrier value of approximately \(0.9799\) for the fixed-control hybrid parameters \((w,α,ε)=(8,0.50,0.15)\), while also showing that the improvement is target-class dependent. Randomized and ablation controls indicate that the dominant contribution is the sector-preserving skeleton, with strict one-bit completion acting as a secondary refinement. We provide code, finite certificates, CSV files, validation logs, and reproduction scripts to make the finite-size claims traceable.

We study the quasinormal spectrum of a massive test scalar field in the exact asymptotically flat regular black-hole geometry supported by a phantom Dirac--Born--Infeld scalar. Using high-order WKB approximation improved by Padé resummation, together with characteristic time-domain integration and Prony extraction, we compute the fundamental mode and the first two overtones for representative values of the regularity parameter and the field mass. We show that increasing the field mass raises the oscillation frequency and reduces the damping rate, while increasing the regularity scale generally makes the ringing softer and longer lived. The time-domain profiles are in very good agreement with the WKB--Padé results and confirm the robustness of the spectrum. For sufficiently large field mass, the damping tends to zero, indicating the onset of quasiresonant behavior, although in the time domain these modes are eventually masked by oscillatory late-time tails. Our results show that massive scalar ringing provides a sensitive probe of this DBI-supported regular geometry.

Minor embedding is a required compilation step for quantum annealing, mapping logical problem graphs onto sparse hardware topologies. Despite its central role in determining solution quality, no standardized benchmark exists for comparing embedding algorithms: prior studies use incompatible graph libraries, inconsistent metrics, and non-reproducible experimental setups, making cross-algorithm comparisons unreliable. We present Ember (Embedding Minor Benchmark for Evaluative Reproducibility), an open-source benchmarking framework addressing this gap. Ember provides a standardized algorithm interface with seeded, reproducible execution infrastructure; a diverse graph library of 24,016 instances spanning structured, random, and physics-motivated problem types not previously used in embedding benchmarks; and a unified analysis pipeline supporting all three current D-Wave hardware topologies (Chimera, Pegasus, Zephyr). We evaluate five algorithms across the full library on Chimera and find that no algorithm dominates universally: rankings vary systematically with graph structure, and the best algorithm depends on the family being embedded. We also examine the effects of hardware topology (including Pegasus and Zephyr), qubit error rates, and evaluate a reinforcement-learning approach (CHARME) within a narrower test set. Ember is available at https://github.com/zachmacsmith/ember and is installable via pip install ember-qc.

We formulate a bounded distance decoding strategy applicable to all stabilizer codes including both CSS and non-CSS code-families. The framework emerges out of the local Clifford equivalence between arbitrary stabilizer states and graph states. Using the graphical representation of the stabilizers and the syndromes, we constitute the bounded distance decoding as an adaptable generalization of maximum likelihood decoding, ensuring correction of all errors with weights upper bounded by a target weight. We show that strategic pruning associated with a feed-forward network structure of the graph can reduce the search space and subsequently the runtime of the designed decoder. We demonstrate satisfactory performance of the bounded distance decoder in the case of the optimal non-CSS codes up to distance $d=11$ subjected to the depolarizing error on all qubits, and near-optimal decoding for the color and the surface codes, both belonging to the CSS family, under the bit-flip errors on the qubits. We also develop an open-source library, QGDecoder, enabling the graph-aware bounded distance decoding of arbitrary stabilizer codes.

In this study, we explore the cosmological evolution of the Universe in the framework of covariant $f(Q)$ gravity, with a coupling function that evolves dynamically in proportion to the Hubble parameter. Two specific forms of the function are examined: a power-law model and a logarithmic model. By rewriting the cosmological field equations as an autonomous dynamical system, we determine and classify the corresponding critical points and analyze their stability. Our results show that both models are able to reproduce the sequence of cosmic evolution, including radiation, matter, and dark energy-dominated eras, along with the transitions between them. The physical properties at each critical point are described using key cosmological quantities such as the total EoS parameter, density parameters, and the deceleration parameter. The stability of the non-hyperbolic critical point is analyzed through center manifold theory. In addition, we present phase space trajectories along with the stability behavior of each critical point. The evolution plots for the density parameters of radiation, matter, and dark energy, along with the EoS parameter for the total, are illustrated for further analysis. Overall, the analysis suggests that the $f(Q)$ models considered here, within the context of covariant formulation, provide a consistent description of cosmic evolution and offer a promising approach to explaining the late-time acceleration of the Universe.

We develop a systematic sign-embedding framework of operator-output quantum algorithms for matrix equations and matrix functions. Differing from the contour-integral treatment, we start with the matrix-sign embedding route: an augmented matrix $M$ whose half-plane matrix sign compresses the target operator either as a block of $\text{sign}(M)$ or, in projector form, through $(I-\text{sign}(M))/2$; we then construct a logarithmic-sinc approximation for the half-plane sign operator and combine it with structure-aware scaled multiplexing and nodewise rebalancing of shifted inverse families. For ordinary Sylvester equations, we offer an explicit block-encoding of the target matrix solution with query complexity linear in the inverse-conditioning parameters and logarithmic in the target error tolerance, under non-normal and non-diagonalizable settings given a field-of-values (FoV) gap or strip-resolvent hypotheses. These algorithms propagate the same overlap-based normalization bookkeeping to ordinary and generalized Sylvester equations, generalized Lyapunov equations, principal square roots and inverse square roots, matrix geometric means, and continuous-time algebraic Riccati equations (CARE). These results identify matrix-sign embeddings and nodewise rebalancing as reusable design principles for structured operator-output quantum linear algebra.

We present a criterion that serves as the basis for a polynomial-time algorithm to decide whether a finite set of qudit gates exponentiated by some Hamiltonians is universal. Our approach formulates universality in Lie algebraic terms and applies Borel--de Siebenthal theory with a diagonal generator having incommensurate spectrum. In this framework, nonuniversality is detected by invariant subspaces, equivalently by a graph-connectivity obstruction, while universality is repaired by adding generators that couple disconnected components. We further prove that two generators are sufficient for universal control. Our work reveals a profound link between qudit universality and irreducibility of Lie algebra representations.

We investigate the role of continuous measurement and postselection in the dynamics and entanglement of a transmon-cavity-transmon coupled system. In the dispersive regime, characterized by a large detuning between the transmons and the cavity, the two transmons interact via virtual excitation of the cavity, giving rise to an effective transmon-transmon coupling. In addition to this coherent interaction, each transmon undergoes spontaneous emission, which is continuously monitored through independent detection channels. By incorporating realistic detector inefficiencies, we analyze both efficient and imperfect monitoring scenarios and demonstrate that postselection significantly slows down the decay of entanglement compared to the unmonitored case. We formulate the stochastic master equation for the coupled system, derive the corresponding postselected master equation, and investigate the dynamics through the Liouvillian superoperator spectrum. In the interaction frame, we identify the emergence of an exceptional point and characterize the associated broken and unbroken PT-symmetric phases. We show how these phases influence the system dynamics and the corresponding entanglement behavior. Our results provide insight into how continuous measurement and postselection affect entanglement in dissipative quantum systems, with potential applications in quantum information processing.

Einstein Telescope (ET) is a third-generation gravitational wave (GW) detector with tenfold better sensitivity compared to the advanced LIGO detectors. It will be capable of observing copious stellar mass binary black hole mergers up to a redshift of 10 which will make it especially useful for cosmography. We generate a mock gravitational wave event catalog for the Einstein Telescope and show the recoverability of either the Hubble constant ($H_0$) or the matter density parameter ($Ω_{\rm m}$). We present a simple, effective and fast technique for inferring $H_0$ (or $Ω_{\rm m}$) using the intrinsic chirp mass spectrum of black hole binaries, and investigate the efficacy of the method assuming the standard model of cosmology. If only $H_0$ has to be constrained, we find that at least one year of ET's observation will be required to achieve 1% uncertainty. With the same amount of observation, $Ω_{\rm m}$ can be constrained to within 4% uncertainty. With ET operating as a standalone instrument, we show that the GW spectral sirens detected by it can constrain the Hubble constant.

Scaling superconducting quantum processors is increasingly constrained by the wiring, heat load, and calibration overhead associated with delivering high-resolution analog signals from room temperature to qubits at millikelvin temperature. Here we demonstrate a superconducting digital-to-analog converter (DAC) integrated with high-coherence fluxonium qubits in a multi-chip module architecture. The DACs generate persistent analog flux signals for tuning qubit parameters and are programmed deterministically using single-flux-quantum (SFQ) pulses, providing a digital interface compatible with established SFQ routing and demultiplexing technologies. Operating at millikelvin temperature, the DACs enable in-situ tuning of fluxonium qubits without measurable degradation of qubit coherence. The presented device provides a static control primitive for flux-tunable qubits, enabling parameter homogenization and eliminating the need for individual room-temperature DC bias lines. These results establish SFQ-programmable millikelvin DACs as a building block for digitally controlled superconducting quantum processors.

Discrete time crystals are non-equilibrium phases of matter in periodically driven systems, characterized by robust subharmonic oscillations and broken discrete time-translation symmetry. Their long-lived coherent dynamics and resilience to imperfections make them promising resources for quantum sensing. A disorder-free discrete-time crystal probe can provide the quantum-enhanced estimation of the coupling parameter. Here, we extend this sensing mechanism to nonlinear interactions and show that this nonlinear profile strongly enhances the sensing precision by increasing the system-size scaling exponent of the quantum Fisher information. Our analytical discussion separates a rigorous seminorm upper bound from the physically relevant scaling realized by product-state probes in the time crystal regime. Numerically, we find that the quantum Fisher information retains its quadratic long-time growth with the number of Floquet cycles, while its system-size exponent increases approximately linearly with the nonlinearity exponent, identifying nonlinearity as a resource for quantum-enhanced sensitivity. We further show that stronger nonlinearities shrink the time crystal stability window, making the probe more sensitive to small deviations from the resonant condition. We also analyze the effect of imperfect pulses and show that such imperfections can enhance, rather than suppress, the information encoded in the evolved state. Finally, we discuss a digital implementation of the nonlinear DTC sensing protocol using superconducting qubits.

We study parameter transferability for the Quantum Approximate Optimization Algorithm (QAOA) across multiple combinatorial optimization problem classes from a parameter generation perspective. Specifically, a meta-optimizer is trained on one problem class and deployed on another during test time. Prior work employs a Long Short-Term Memory network to emulate QAOA optimization trajectories, but the learned dynamics usually collapse to near-identical paths, limiting cross-problem transfer efficiency. In this paper, we present a problem-aware graph-conditioned meta-optimizer for QAOA that learns to generate parameter trajectories over a fixed horizon, providing strong initializations with only a few steps. The optimizer is conditioned on compact graph embeddings and trained end-to-end using differentiable feedback from the QAOA objective, avoiding the need for ground-truth angles. We evaluate across multiple graph problem classes, including MaxCut, Maximum Independent Set, Maximum Clique, and Minimum Vertex Cover. We report both solution quality and feasibility-aware metrics where constraints apply. Results across a comprehensive empirical study consisting of 64 settings show that the learned optimizer can reduce optimization effort and improve performance over standard initialization, while exhibiting transferable behavior across graph families and problem types.

We investigate the dynamics of massive test particles around a static black hole in nonlocal gravity and examine the corresponding properties of HF QPOs, constraining the nonlocal parameter to $α/M \leq 0.452$. The nonlocal parameter $α$ enhances the effective potential $V_{eff}$ and leads to a systematic reduction in the energy E and angular momentum L of circular orbits. Consequently, the ISCO radius, along with the associated energy and angular momentum, decreases monotonically with $α$, while the radiative efficiency increases, reaching a maximum of approximately $8.9\%$. Due to the spherical symmetry of the spacetime, the Keplerian frequency $Ω_φ$ and the vertical epicyclic frequency $Ω_θ$ coincide and are suppressed by $α$, whereas the radial epicyclic frequency $Ω_{r}$ is enhanced. The impact of $α$ on several twin-peak HF QPO models is examined, revealing that $α$ increases both the lower and upper bounds of the predicted QPO frequency ranges. By imposing the $2ν_U = 3ν_L$ resonance condition, we analyze the resonant radius, upper QPO frequency, maximum allowed black hole mass, and the time delay between the shadow and QPO signals. We find that the resonant radius decreases with $α$, while the upper QPO frequency increases, spanning the range $ν_U \sim(673/M-4360/M)$Hz. When the TOV imit is imposed, the upper frequency is further constrained to $ν_U \lesssim 1450$Hz. Combining astronomical observations for the classification of QPOs, where $ν_U \geq 100$Hz, the black hole mass in the nonlocal gravity should satisfy $M \lesssim 43.6M_\odot$. Although the radial separation between the resonant radius and the photon sphere decreases with $α$, the associated gravitational time delay increases, remaining below $\sim 1.3$ms and thus negligible for current observational capabilities.

Transmitter-device-dependence is a longstanding but often implicit problem in quantum key distribution (QKD), as compared to measurement-device-dependence. One-sided device-independent (1sDI) scenario relaxes the security conditions of DI framework and offers an intuitive solution to transmitter-sided dependence. Here we show that, by constructing a composable transmitter in a 1sDI configuration, the transmitter-device-dependence can be largely eliminated. Our scheme integrates the entanglement source into the transmitter as an individual part, and the detection module is another individual part treated as black box. In a proof-of-principle experiment we obtain a secure key-rate of 1~kbps at an equivalent transmission fiber distance of 20~km, demonstrating the first discrete-variable 1sDI-QKD. By implementing semi-transmitter-device-independent security while maintaining strong loss tolerance, our approach bridges security and practicality for real-world 1sDI-QKD deployments.

Early-time evolution away from an unstable equilibrium in a nonlinear system is often expected to be governed by the associated linear instability. Combining full nonlinear evolution with first- and second-order quasinormal mode (QNM) calculations, we show that this expectation can fail during the unstable growth stage of a Q-ball hairy black hole in Einstein-Maxwell theory with a charged self-interacting scalar field. The linear unstable QNM has a much larger amplitude in one component of the scalar field than in the other: the more strongly responding component follows that mode, whereas the early growth of the more weakly responding component is dominated by a second-order QNM sourced by the linear unstable mode. This occurs while the evolution remains perturbative. Our results thus show that the early growth of an individual component need not be governed by its linear response.

A minimal method to fabricate Al/AlO$_x$/Al Josephson junctions (JJs) using photolithography and argon etching, before metallization and oxidation, is demonstrated. JJs with areas ranging from 1 to 6 $μ$m$^2$ can be fabricated and, with the appropriate oxidation conditions, the junction resistance can be varied by $\sim$2 orders of magnitude. Transmission electron microscopy reveals the successful fabrication of JJs with few grain boundaries suggesting reduced energy loss from two-level-systems. Superconducting QUantum Interference Devices (SQUIDs) fabricated from this methodology exhibit reduced resistance variation over multiple chips, compared with electron beam lithography, and the devices can sustain repeated thermal cycles to 10 mK with the excellent flux response remaining unchanged. The quantum applications of this technology are demonstrated by embedding a SQUID resonator into a 3D cavity and parametrically amplifying low photon numbers with gains of $\sim$40 dB. This work establishes the simplest approach to fabricating JJs to date, and it could prove pivotal to the widespread utilization of superconducting circuit-based quantum technologies.

Network tomography refers to the use of inference techniques for inferring internal network states from end-to-end probes. Quantum probes, implemented by sending blocks of $n$ coherent-state pulses augmented with continuous-variable (CV) squeezing ($n=1$) or weak temporal-mode entanglement ($n>1$) over a lossy channel to a receiver with homodyne detection capabilities, are known to carry information about the channel transmissivity. Assuming a subset of nodes in an optical network is capable of sending and receiving such probes through intermediate nodes with all-optical switching capabilities, we leverage these quantum probes to estimate link transmissivities. To determine how to route the probes in a network, we propose a probe construction algorithm that guarantees link identifiability, while maximizing the number of information orthogonal sets of transmissivities. A set of probes induces a Fisher information matrix (FIM). We then derive two metrics, the determinant of the FIM and the trace of its inverse, to evaluate the performance of the probes. In particular, our results can be used to characterize the quantum improvement in estimating link transmissivities in a general optical network.

Among the many solid-state emitters being explored for scalable quantum technologies, the silicon T centre is a leading candidate offering long-lived spin qubits, a telecommunications-band spin-photon interface, and integration with on-chip photonic circuits. However, nanophotonic integration broadens both the inhomogeneous spectral distribution and individual emitter linewidths. Here, we integrate single T centres into silicon nanophotonic cavities with p-i-n diodes for local electronic control. These devices enable Stark tuning up to 30 GHz, sufficient to bring 55(2)% of on-chip T centres into mutual resonance, and demonstrate tunable lifetime reduction across the cavity resonance. A model of the joint excitation probability shows an orders-of-magnitude increase in entanglement rate by tuning distinct emitters into mutual resonance. Luminescence modulation at high reverse biases reveals a transition to a dark charge state. Finally, bias-induced modulation of the optical transition splitting uncovers a potential mechanism for electrically driven excited-state spin mixing via spin-orbit coupling. Localized and individual spectral tuning increases the yield of performant silicon spin-photon interfaces and the number of devices per chip available for large-scale entanglement and quantum information technologies.

Demonstrating quantum advantage for combinatorial optimization requires more than standalone algorithmic results; it calls for end-to-end case studies that integrate problem modelling, quantum execution, and classical refinement into practical workflows. This paper presents a sandbox platform for experimenting with hybrid quantum-classical workflows in graph optimization, enabling the systematic study of end-to-end optimization pipelines. Using our platform, we investigate three classically intractable and mutually reducible graph problems -- Minimum Vertex Cover, Maximum Independent Set, and Maximum Clique -- by transforming them into an unconstrained problem and solving the resulting instances with QAOA on IBM platforms. Our workflow combines classical pre-processing to reduce instance size, quantum optimization on the reduced problem, and classical postprocessing to map quantum outputs to high-quality feasible solutions, thereby avoiding direct constraint encoding in the quantum circuit. We evaluate the approach on synthetic graphs, benchmark instances, and real-world networks, and report hardware experiments on IBM Quantum System One at PINQ2 in Bromont, Quebec, powered by IBM's 156-qubit Heron r2 processor on graphs up to 128 vertices, with circuits involving up to 128 qubits and 13,555 two-qubit gates. The results illustrate how sandbox-style end-to-end experimentation can expose bottlenecks, clarify the role of classical-quantum workload partitioning, and provide domain experts and practitioners with a practical guide for interpreting quantum optimization outputs and assessing quantum utility on the road to quantum advantage in combinatorial optimization.

In an extension of the Unconventional Noiseless Intermediate Quantum Emulator, this work introduces a classical emulation of the quantum Harrow-Hassidim-Lloyd algorithm for sampling from the solution space of linear systems. The emulated HHL algorithm scales exponentially with the number of qubits required to represent the linear system, which is an advantage over the state vector simulation of the HHL algorithm, which scales exponentially as a function of both the size of the linear system and the magnitude of its largest (scaled) eigenvalue. We benchmark our emulator by comparing it with the Intel Quantum Simulator and demonstrate a runtime advantage for small linear systems.

We present an alternative scheme to achieve nonreciprocal unconventional magnon blockade (NUMB) in a hybrid system formed by two microwave cavities and one yttrium iron garnet (YIG) sphere, where the pump and signal cavities interact nonlinearly with each other and the signal cavity is coupled to the YIG sphere. It is found that the nonlinear coupling occurs between the pump cavity and magnon modes due to the dispersive interactions among three bosonic modes. Meanwhile, the Kerr nonlinearity is present in the pump cavity. Based on these nonlinear effects, a nonreciprocal magnon blockade could be achieved with the help of weak parametric driving of the pump cavity. The present work provides an alternative method to prepare single magnon resource, which may be helpful for quantum information processing.

We propose a parallel protocol for implementing distributed nonlocal quantum gates between spatially separated stationary qubits encoded in dual-species quantum emitters (i.e., color-center and superconducting qubits). By utilizing entangled photon pairs with distinct frequencies as a quantum data bus, our approach connects spatially separated devices without requiring quantum frequency conversion or preshared entanglement, while maintaining an always-ready and resource-efficient property for distributed quantum computing and networks. Furthermore, we demonstrate the feasibility of implementing parallel distributed nonlocal quantum gates on multiple pairs of spatially separated qubits using a single high-dimensional entangled photon pair, which directly benefits from the enhanced quantum capacity provided by optical qudit encoding. Our protocol establishes a scalable and practically implementable framework for distributed quantum networks, potentially enabling the development of future large-scale quantum computing architectures.

To grow the quantum information science and technology workforce, opportunities for students to gain experiential learning and build a sense of belonging in the broader community are essential. The Undergraduate School on Experimental Quantum Information Processing (USEQIP) is a two-week summer school for undergraduate students that has been held since 2009 with the goal of introducing undergraduate students from around the world to the tools of quantum information research, paired with a summer internship program. Here we report on the structure, impact, and outlook of the program, including hands-on laboratory activities refined over many iterations of the program. We highlight the career trajectories of program alumni, many of whom have made significant contributions to the quantum field.

We present a covariant quantization of the first-order formulation of the Einstein-Hilbert theory using the path integral and BV formalisms. In this approach, the metric $g^{μν}$ and the connection $Γ^λ_{μν}$ are treated as independent, with the connection playing the role of an auxiliary field. We show that the gauge algebra is closed and irreducible. A novel trivial local symmetry arises, leading to structural identities that relate the Green's functions of the auxiliary field to its classical value. We revisit the quantum equivalence between the first- and second-order formulations of the Einstein-Hilbert theory. By employing a suitable trick, a manifestly covariant form of the Senjanović measure is derived. We also show that the two formulations are equivalent at the level of the effective action when the auxiliary field is on-shell.

Although asymptotically flat black holes generically lack thermodynamic phase transitions, we show that curvature-induced scalarization of electrically charged black holes in Einstein-Maxwell- Scalar-Gauss-Bonnet theory provides a natural setting for nontrivial thermodynamic behavior, without invoking external confining mechanisms or an extended thermodynamic formalism. Working within the canonical ensemble and employing the Euclidean approach, we identify the coexistence of locally stable scalarized and small Reissner-Nordström thermal states, which promotes free-energy crossings to bona fide phase transitions between equilibrium phases. For weak coupling, a second-order phase transition coincides with the second bifurcation point, at which the scalarized branch reconnects with the Reissner-Nordström branch and scalar hair is spontaneously shed. As the coupling strength increases, this transition becomes zeroth order, the scalarized branch shrinks, and a fish-like structure develops in its Helmholtz free energy, rendering locally stable thermal states partially metastable, and yielding up to three phase transitions. In the strong-coupling limit, the scalarized branch reduces to a Schwarzschild-like solution, and the Reissner-Nordström phase ultimately emerges as the sole thermodynamically preferred configuration

Near-term FTQC system designs are constrained by limited error budgets and largely sequential execution of non-Clifford gates. As a result, reducing the number of the most-error prone instructions becomes critical for successful program execution. In this work, we study the extractor architecture, a recently proposed FTQC design that enables universal quantum computation on spatially-efficient QEC codes such as the BB code family. In these architectures, over $90\%$ of the total program error arises from the synthillation process, which involves $\lvert T\rangle$-state preparation and injection to implement non-Clifford gates. We observe that standard Rz synthillation requires multiple sequential $\lvert T\rangle$-state injections, each incurring an inter-module measurements, the most expensive instruction in the architecture, which cumulatively dominate the overall error budget. To address this bottleneck, we propose INJEQT, a $2$-factory design that uses an auxiliary code capable of synthesizing $Rz(θ)$ states with lower error rates. These states are then injected into the extractor modules using only a constant number of inter-module measurements. This approach reduces overall error rates by up to $22\times$. We further reduce the time overhead by a pre-fetching strategy that prepares the Rz states and their correction states in parallel. This approach improves the wall-clock time by up to $13\times$ and reduces the space-time cost by up to $7.2\times$, for an optimal choice of the number of INJEQT factories for each metric. We evaluate INJEQT for multiple state preparation techniques such as distillation, cultivation and STAR, and model the execution times for both lattice surgery-based and transversal CNOT based injections. Our results demonstrate that INJEQT is robust across factory choices and device technologies, enabling more efficient architectural designs for FTQC.

Coherent quantum rollout for sequential decision problems requires a unitary simulator: randomness must live in explicit quantum registers, and basis-state selectors must be mapped to actions reversibly. With branch-dependent valid actions, this mapping is totalized coherent rank-select over an entangled $N$-bit validity mask: return the position of the $r$-th valid bit, or a sentinel if $r$ is out of range. We give the first reversible-circuit complexity analysis of this primitive. For selector width $w = \lceil \log_2(N+1) \rceil$, rank-select admits an $O(Nw)$-gate low-ancilla bounded-span scan, proved gate-optimal in its model, and an $O(N\log w)$-gate low-ancilla blocked construction when long-range gates are available; across all bounded-fan-in layouts, the unconditional gate lower bound is $Ω(N)$. Composing rank-select with reversible transition and predicate-evaluation circuits gives an explicit polynomial-size coherent rollout oracle for finite-horizon planning problems satisfying these primitive assumptions. The resulting oracle satisfies the access model of the best-arm pipeline of Wang et al., yielding $\widetilde{O}(\sqrt{k}/\varepsilon)$ coherent oracle calls against the standard classical $Ω(k/\varepsilon^2)$ arm-pull lower bound. We give a bounded-influence lifting theorem that extends this lower-bound construction from a base configuration to an exponential family of configurations. We instantiate the construction on SIR epidemic intervention, with a stochastic placement-game sanity check, and machine-check the main results in Lean 4. Code and proofs: https://github.com/BinRoot/b01t/tree/main/demos/rollout.

The performance of the Quantum Approximate Optimization Algorithm (QAOA) on noisy intermediate-scale quantum (NISQ) devices is strongly limited by sparse qubit connectivity. When interactions required by QAOA Hamiltonians are not aligned to the hardware topology, transpilation introduces SWAP gates, increasing circuit depth and noise. We propose a SWAP-free QAOA framework based on modifying the cost Hamiltonian so that it can be implemented natively on the hardware. We formulate this as a mixed-integer semidefinite program (MISDP) that selects a hardware-compatible approximation of the original cost matrix and optimizes the allocation of logical variables to physical qubits. We prove that the associated decision problem is NP-complete and derive theoretical guarantees relating the MISDP objective to the loss in the original optimization problem through the Lovász number of the hardware graph. Since solving MISDPs is practical only for small instances, we introduce heuristics based on spectral properties of the problem matrix and hardware graph. Our experiments on a cardinality-constrained quadratic optimization model for index tracking show competitive performance against a baseline representing ideal QAOA under SWAP-induced noise. These results indicate that, on sparse NISQ architectures, a hardware-aware approximation of the objective may be more effective than an exact but heavily transpiled Hamiltonian implementation.

To implement quantum algorithms on a quantum computer, we must overcome the twin problems of fault-tolerance -- how can we realize a relatively noiseless computation by cleverly combining noisy components? -- and compilation -- how can we realize an arbitrary quantum algorithm given the basic operations available on the quantum device at hand? We show how treating the former problem via error-correcting codes enables greater flexibility in resolving the latter. Specifically, we explicitly leverage the fact that error-correcting codes introduce redundancy which renders physically distinct operators logically indistinguishable. In terms of computation, it suffices to implement any operator logically equivalent to some target, yet from a compilation perspective, certain choices may be preferable to others. Our novel contribution is making this intuition precise in the general setting of the special unitary group. In particular, we describe how to reduce the problem of making a compilation-ideal choice to a least squares problem and provide a closed form solution thereof. Using our framework, it is possible to circumvent inserting costly swaps to adhere to hardware connectivity; instead, we could realize the logical target through a distinct physical Hamiltonian that is natively accessible. We elucidate our approach using the $[[4,2,2]]$ code. We discuss connections to compressed sensing that may pave the way to efficient compilation leveraging physical degrees of freedom.

Quantum entanglement between gamma-ray photons emitted following electron-positron annihilation is expected to be maximal and may be characterized via non-classical polarization correlations. However, this is difficult to verify experimentally because there are no established schemes that approach ideal projective-polarization measurements for high-energy photons. Hence, polarization entanglement between MeV-scale annihilation photons has not yet been conclusively demonstrated. We develop here a framework that models polarization measurements of high-energy photons via Compton polarimetry, employing the formalism of positive operator-valued measures (POVMs). We extend the POVM description to sequences of repeated interactions and show that the measurement converges toward an ideal projective measurement of linear polarization as the number of interactions increases. We demonstrate that this progressive improvement in measurement sharpness can enable the experimental violation of CHSH inequalities.

Quantum low-density parity-check (QLDPC) codes with good parameters are promising candidates for low-overhead fault-tolerant quantum computing, but their non-local stabilizers require long-range connectivity and frequent qubit movement, introducing practical challenges. Prior work has studied the networked implementation of topological codes, where each node only holds one or a few qubits of the entire code, and demonstrated competitive performance under practical constraints such as the quality of network-provided entanglement. However, since these codes are already geometrically local, such a networked setting might not be essential. In this work, we propose and study the networked implementation of better QLDPC codes, specifically bivariate bicycle codes due to their similarity to surface codes and the controlled amount of long-range connections in their stabilizers. We begin by recreating networked surface codes in Stim, with one code qubit per node, and provide additional insights into their circuit-level noise performance. We then extend this approach to bipartitions of bivariate bicycle codes, using balanced min-cut partitioning on their combined X-Z Tanner graph to identify optimal qubit splits. For stabilizers spanning nodes, we implement teleported CNOTs and vary the Bell pair fidelity enabling these gates. Through circuit-level noise simulations with BP-OSD decoding, we provide the first insights into networked realizations of these codes and compare their performance with monolithic implementations. We conclude by outlining advantages, limitations, and future directions.

This paper describes the design and implementation of a two-day quantum hackathon for underrepresented high school students in Nova Scotia, Canada. The first day of the hackathon is spent introducing students to quantum computing through hands-on activities, whereas the second day teaches students to apply this knowledge through guided challenges. Both days are informed by the theory of mastery learning and specification grading, with the full curriculum being crafted within the Integrated Course Design framework. This requires identifying situational factors unique to our target demographics, from which we develop learning outcomes, and then work backwards to a full curriculum with educative assessments. A novel aspect of our hackathon is that all circuit simulations are performed within Quirk: a decision based on best practices in computer science education. Based on feedback from students, we conclude that our hackathon successfully introduced students to the basics of quantum computing, and was able to reach most of our target demographics.

We investigate the strong-field phenomenology of a static and spherically symmetric regular black hole supported by an Einasto dark matter (DM) distribution. For the exponential Einasto profile, the geometry is controlled by a single dimensionless halo parameter $a$, and we restrict the analysis to the black hole branch $0<a\leq a_{\rm crit}\simeq0.388$. We study both timelike and null geodesics, including the effective potential, circular orbits, ISCO radius, orbital period, periapsis advance, photon sphere, shadow radius, effective photon force, and representative photon trajectories. We also construct image-plane intensity profiles and face-on thin-disk images in a static-emitter approximation. The analysis reveals a hierarchy of strong-field sensitivity. Timelike observables remain largely degenerate with the Schwarzschild limit along most of the black hole branch, while the photon sphere scale, shadow diameter, and residual optical structures provide the most sensitive response to the Einasto halo near the critical black hole regime. A comparison with the EHT shadow-scale measurements shows that the full branch is consistent with Sgr A* at the $1σ$ level, whereas M87* mildly disfavors values very close to criticality. These results indicate that the most promising signatures of the Einasto halo are not expected from ordinary timelike orbital quantities, but from near-critical photon propagation and its imprint on the optical appearance.

Recent theoretical studies propose that Hawking radiation may not emerge strictly at the event horizon but rather from the spatially extended region surrounding a black hole, commonly referred to as the quantum atmosphere. In this work, we explore how this concept influences nonlocal quantum correlations in a bosonic bipartite system located at certain distance from a Schwarzschild black hole. By employing the measurement-induced nonlocality (MIN), as a quantifier of quantum correlations, we analyze the response of bosonic fields to the thermal and geometric characteristics associated with the Hartle-Hawking vacuum. In this manner, we extend previous studies that primarily focused on the fermionic systems. Our results reveal that, when quantum atmosphere is taken into account, the behavior of MIN departs from its conventional near-horizon profile. In particular, bosonic nonlocal correlations are found to exhibit a pronounced degradation at a finite radial distance from the event horizon and to ultimately vanish as scaled distance increases further. To some extent this behavior contrasts with the previously considered fermionic case, indicating that bosonic fields provide potentially stronger response to the quantum atmosphere.

Sparse quantum state preparation is a common subroutine in quantum algorithms, where classical data with few nonzero entries must be loaded into a quantum state. In this work, we consider the Grover-Rudolph algorithm, which has recently been shown to efficiently prepare sparse states, and we propose two improvements. First, we extend an existing gate-merging procedure by allowing rotations to merge with virtual zero-angle gates on unreachable branches of the preparation tree, reducing the number of CNOTs and control qubits. Second, we introduce an approximate variant in which rotations with similar but not identical angles are merged at the cost of a small, controllable error in the prepared state. We derive a classically computable estimate of the resulting overlap with the target state, which is used to guide the merging decisions.

The maximum flow problem asks to find the largest possible flow from a source to a sink in a capacitated network. It arises frequently in scheduling, project selection, and as a core subroutine in broader optimisation tasks. Classically, it can be efficiently solved using Dinic's algorithm, which repeatedly performs breadth-first search (BFS) and blocking flow computations on the graph. As a potential candidate for quantum speedups, these BFS subroutines can be naturally replaced with quantum BFS (qBFS), an instantiation of Grover's search algorithm. In this paper, we evaluate the expected performance of qBFS on standard classical datasets. These instances are too large to be solved directly on current quantum hardware, so we adopt a hybrid benchmarking approach: (i) we run a classical implementation of Dinic's algorithm and isolate the runtime of its BFS subroutines; (ii) we analytically estimate the minimum number of quantum cycles required to implement qBFS, where we use the classically logged data. Our results indicate that achieving a practical quantum advantage for realistic problem sizes would translate to quantum gate operation times surpassing physical limitations.

We study the analytic properties and three equivalent representations of the Toller matrices $T^{(\pm)}$ which appear in the causal formulation of spinfoam transition amplitudes for 4d Lorentzian quantum gravity. These are polynomially bounded functions on the Lorentz group which satisfy the relation $T^{(+)}+T^{(-)}=D$, where the Wigner matrix $D$ provides a unitary irreducible representation of $SL(2,C)$. Ruhl's definition of $T^{(\pm)}$ in terms of analyticity and asymptotic properties is shown to be equivalent to the recently introduced Feynman $i\varepsilon$ prescription in spinfoams. We show that, equivalently, they can be represented as an integral over eigenvalues of the boost operator, which results in a sum over residues. The latter reproduces the Wick rotation relating Euclidean $Spin(4)$ to Lorentzian $SL(2,C)$ spinfoams studied by Dona, Gozzini and Nicotra. We provide explicit expressions in terms of hypergeometric functions and specialize them to the $γ$-simple representations relevant for spinfoams.

The purpose of this work is to discuss how matter fields are coupled to gravity within the framework of General Relativity. Our particular focus here is on the coupling of scalar field models. In a first step, we suggest a new method for coupling scalar fields to the Einstein equations \emph{without} the use of a variational principle or Lagrangian. We show that, under the appropriate assumptions, this new method (for coupling scalar fields to gravity) reproduces the minimally and $k$-essence scalar field couplings with a non-zero potential. We therefore interpret this formalism as describing a \emph{generic} method for coupling scalar fields to gravity. The approach described here allows for a number of free fields which we interpret as constitutive freedoms. In a second step, we choose these free fields in such a way that the resulting system is somehow ``near minimal''. In this setting we investigate Bianchi I type solutions. We establish conditions under which the solutions are asymptotically Kasner, near the initial singularity, and investigate their stability properties.

We provide a formulation of the initial boundary value problem for Friedrich's extended conformal Einstein field equations in which boundary data is prescribed on a timelike hypersurface located at a finite position in the spacetime. Our construction relies on a gauge based on the properties of conformal geodesics and requires the the boundary is ruled by timelike conformal geodesics. The consequences of this assumption on the timelike boundary are analysed and we identify a subset of maximally dissipative boundary conditions which are consistent with this assumption. For this class of consistent boundary conditions we establish the wellposedness of the initial boundary value problem and prove the propagation of the constraints.

Black hole spectral instability is counterintuitive and contradicts many plausible assumptions of the properties of black hole quasinormal modes. The present study aims to explore different types of instability phenomena. It is understood that the fundamental mode is surprisingly sensitive to small perturbations of the effective potential of the linearized wave equation. Such perturbations can be produced by small-scale modifications of the spacetime metrics. From both the analytical and numerical perspectives, we elaborate on a few qualitatively different examples illustrating the strong sensitivity and diversity in black hole spectral instability caused by such effective potential perturbations. It turns out that the qualitative way in which the fundamental mode becomes perturbed depends on many factors such as the shape of the potential and how the perturbation changes as it moves away from the central black hole.

Quantum simulation offers a promising framework for quantum field theory calculations. Obtaining reliable results, however, requires careful characterization of systematic uncertainties. One important source is the boson truncation error, which arises from representing infinite-dimensional local Hilbert spaces with finite-dimensional ones. Previous studies have examined this problem from several perspectives. In particular, Jordan, Lee, and Prekill (arXiv:1111.3633) derived an energy-based bound applicable to generic low-energy states across a broad class of field theories. However, this approach often yields overly conservative bounds, especially at large volumes. In this work, we introduce a new methodology that significantly tightens the energy-based boson truncation bound through two complementary advances: an improved analytic derivation and a Monte Carlo-based numerical procedure. We demonstrate the method in (1+1)-dimensional scalar field theory and (2+1)-dimensional U(1) gauge theory in the dual formalism. Our approach substantially mitigates the volume dependence of the required truncation cutoff, achieving reductions nearly proportional to the volume in some cases and to the square root of the volume in others.

Dual-species Rydberg atom arrays extend single-species platforms by introducing competing interaction scales and enhanced quantum fluctuations, enabling phenomena beyond homogeneous settings. In this work, we study the ground-state phase diagram of a one-dimensional dual-species Rydberg atom ladder using large-scale density-matrix renormalization group calculations. We identify disordered phases, multiple ordered phases with $\mathbb{Z}_2$, $\mathbb{Z}_3$, and $\mathbb{Z}_4$ symmetry, as well as floating phases characterized by incommensurate wave vectors and algebraically decaying correlations. Importantly, we observe a smooth crossover between distinct $\mathbb{Z}_2$-ordered regimes, reflecting a reorganization of low-energy degrees of freedom rather than a true phase transition, which is absent in single-species Rydberg arrays. We further uncover a multi-critical point at the boundary between the $\mathbb{Z}_2 \otimes \mathbb{Z}_2$ and $\mathbb{Z}_3 \otimes \mathbb{Z}_3$ ordered phases, where Ising, chiral, and first-order transition lines intersect. Our results demonstrate that dual-species Rydberg atom arrays provide a unique platform for realizing crossover physics and multi-critical behavior inaccessible in single-species architectures.

Quantum neural networks generalize classical artificial neural networks into the quantum domain. They are formulated as parameterized quantum circuits which are optimized by measuring and minimizing a suitably chosen loss function. The core challenge in understanding, implementing and ultimately using quantum neural networks is that they represent many-body systems with an exponentially large Hilbert space, in combination with a large parameter search space. Moreover, noise -- which is inherent to any quantum measurement -- sets practical limits for the estimation of training loss. Here, we study physics-informed large-scale quantum neural networks that are trained through a finite number of noisy loss function measurements. We show that this architecture permits the construction of nontrivial decision boundaries that enable the classification of quantum states through measuring an order parameter. Our approach can directly process quantum data that is output from quantum simulators and computers and is well suited for implementation on current hardware. Moreover, owed to a close link between the neural network dynamics and the evolution of Markovian open many-body quantum systems, one may expect a certain robustness to noise, which is ubiquitous in the current NISQ era.

This article discusses incorrect statements appearing in textbooks on quantum field theory (QFT); some of these mistakes also appear in the research literature. The focus is not on errors made by an individual author, but on conceptual muddledness that is widespread in introductory textbooks. We start from a bare-bones summary of QFT, meant to establish the notation. We then turn to our six paradigmatic themes, in each case quoting a specific example of the textbook mistake, a summary of material that is known to experts but is frequently mishandled in introductory works, pointers to authoritative references where the relevant concept is handled properly, as well as a concise correction that rectifies any issues. The goal of this work is to warn readers of the existence of several pitfalls and thereby stop these errors from further propagating in the literature on QFT.

Quantum random number generation (QRNG) relies on the inherent unpredictability of quantum mechanical phenomena to efficiently generate high-quality random numbers that can be used in a wide range of cryptography and simulation applications. Here we report the experimental demonstration of QRNG from the arrival times of photons emitted by nitrogen vacancy (NV) centres in fluorescent nanodiamonds. The generation rates achieved range from 0.173 Mbits/s for a region with a single NV centre to 4.77 Mbits/s for a region with just under 50 NV centres, where the latter demonstrates an order of magnitude improvement compared to the highest generation rate previously achieved with NV centres. For all the regions investigated, the generated bits passed the ENT and NIST Statistical Test Suites without post-processing. The results are consistent with our theoretical analysis, where we show that the min-entropy is very close to the ideal value of one per bit for all the regions investigated. This work opens up new possibilities for robust QRNG in highly compact on-chip settings.

Spontaneous scalarization phenomenon in scalar-tensor gravity is known to be a form of phase transition, and it was recently shown that the order of this transition changes depending on the parameters of the theory. There exists a phenomenological description of this result based on Landau theory, but the underlying mechanisms which determine the coefficients of the Landau expansion were unknown. In this study we calculate these coefficients starting from first principles. To this end, we start with an energy functional that describes the nonlinear behavior of the theory, and reduce it to an energy function. This allows us to explain the previously observed, but not well-understood, features of the scalarization phase transition, and enables us to predict which phase transition order will be present for which coupling function or in which regime of the parameter space. The details of the phase transition determine certain astrophysical observables such as signals sourced by transitions from metastable states in first-order scalarization. Thus, predicting these details is an important part of understanding scalarization itself.

Lunar gravitational-wave (GW) detectors relying on the measurement of the response of the Moon to GWs are susceptible to a seismic background, which might pose a fundamental sensitivity limitation. The Lunar Gravitational-wave Antenna (LGWA) was conceived as an array of accelerometers with the idea that data can be processed to distinguish between a GW signal and the seismic background. As a result, the seismic noise of the GW measurement would be mitigated. However, so far, no quantitative assessment of the mitigation of the seismic background has been provided. In this article, we derive the analytical expressions for the optimal squared signal-to-noise ratio considering two seismic stations in an isotropic, random, Gaussian seismic field. Our numerical analysis reveals that the capacity to mitigate the seismic noise critically depends on the distance between the two stations relative to the seismic-correlation length. We demonstrate that optimal placement of the two stations can yield significant improvements in the equivalent seismic noise amplitude spectrum density (ASD), approximately a factor of 2.3 at 0.3 Hz, compared to the measurement with a single station. The equivalent ASD of the seismic noise also exhibits distinct oscillatory and mitigation features arising from the Bessel-function structure of the noise correlation.

We investigate Hawking-like particle production in analogue gravity systems with superluminal modified dispersion relations. For a broad class of even, convex, and polynomially bounded dispersion relations, we show that the relevant outgoing modes are governed by an effective horizon induced by dispersive propagation. Extending the near-horizon S-matrix method beyond the purely sonic regime, we compute the Bogoliubov coefficients and demonstrate that, in the low-energy and adiabatic limits, they agree with the tunneling result obtained from the approximant ray. In both cases, the emission spectrum is controlled by an effective surface gravity associated to the effective horizon, leading to controlled deviations from exact thermality. Our results establish an analytical connection between the Bogoliubov and tunneling descriptions in dispersive settings and clarify the conditions under which Hawking radiation remains robust against ultraviolet modifications, with implications extending beyond analogue gravity.

We construct a Born-Infeld-type $f(R,{\cal G})$ modification of gravity, where ${\cal G}$ is the Gauss-Bonnet term, by embedding Born-Infeld electrodynamics in a five-dimensional pure modified gravity. This method leads to the correspondence between curvature scalars and electromagnetic field strength scalars -- $R\leftrightarrow F_{μν}F^{μν}$ and ${\cal G}\leftrightarrow (ε_{μνρσ}F^{μν}F^{ρσ})^2$ -- allowing us to replicate the structure of Born-Infeld electrodynamics in the gravitational sector. The resulting Born-Infeld-type gravity is a ghost-free $f(R,{\cal G})$ theory which reduces to Einstein gravity in the low energy limit. In this work we focus on bouncing cosmological solutions of such a theory, which require positive spatial curvature. By using both the Jordan and Einstein frame analyses, we find a vast space of bouncing solutions with different asymptotic behaviors, including solutions with multiple bounces grouped together. Observational consequences of such solutions will be investigated in the future.

We study a chain of periodically driven Rydberg atoms and identify a class of drive protocols for which the system exhibits emergent prethermal Bethe integrability at special drive frequencies. We provide a perturbative analytic expression of its Floquet Hamiltonian in the large drive amplitude regime. We demonstrate integrability of the leading term of this Floquet Hamiltonian at special drive frequencies, which we identify, by mapping it to the Hamiltonian of the paradigmatic spin-$1/2$ ${\rm XXZ}$ chain. We support our analytical results by exact diagonalization studies on finite chains. Our numerical results on level statistics, half-chain entanglement entropy, and longitudinal magnetization of the driven chain brings out its emergent integrable nature at the special drive frequencies which persists up to a large prethermal timescale.

We classify mobile Pauli stabilizer codes up to gapped interfaces and coarse-graining using the framework of algebraic $\mathrm{L}$-theory. We compare this classification with that of framed TQFTs, theories that arise naturally in the continuum, highlighting a close structural relationship between the two. Our approach is formulated in the category of perfect chain complexes equipped with quadratic functor over the Laurent polynomial ring $R = \mathbb{Z}/p[x_1^{\pm 1}, \ldots, x_n^{\pm 1}]$, within which the collection of topological operators of Pauli stabilizer codes arise naturally as objects. In particular, we establish a bulk-boundary correspondence for lattice theories: the equivalence class of a Pauli stabilizer code up to gapped interface is described by a Clifford QCA in one dimension higher. This is done using the universal target category for stabilizer codes, which is the categorical spectrum whose existence and universal properties are introduced in this work. We conclude by highlighting subtle differences between the classification of Pauli stabilizer codes and TQFTs, leading to qualitative distinctions between lattice and continuum theories.

A matter wave propagating through curved spacetime accumulates phase that encodes both geometry and gauge structure. We develop a semiclassical framework for charged spin-$1/2$ matter-wave interferometers based on a WKB expansion of the covariant Dirac equation, in which the phase decomposes into dynamical, spin, and electromagnetic Aharonov-Bohm (AB) contributions. In a freely falling detector frame, all three channels are governed by local tidal fields. In a weak gravitational-wave (GW) background, the dynamical and spin phases probe the gravitoelectric and gravitomagnetic sectors of curvature, while the AB phase arises from curvature-induced electromagnetic fields obtained from Maxwell's equations in curved spacetime. For a Mach-Zehnder interferometer (MZI), all three responses are determined by the same tidal scale, $\ddot{h}_A \sim Ω^2_{gw}h_0$, and filtered by a common geometric kernel, while entering through distinct physical couplings. In particular, the AB contribution depends not only on the enclosed flux but also on spatial variations of the induced fields and exhibits an intrinsic frequency dependence set by the traversal time. These results provide a unified description of matter-wave interferometric phases in time-dependent GW backgrounds and identify complementary dynamical, spin, and electromagnetic pathways through which spacetime curvature imprints itself on quantum interference.

Using the recent work of Brendle--Wang on the Riemannian positive mass theorem, we prove the spacetime positive mass theorem for asymptotically flat and asymptotically hyperboloidal initial data sets in arbitrary dimension $n$.

Locally constrained gauge theories underpin our understanding of fundamental interactions in particle physics and the emergent behaviour of quantum materials. In strongly correlated systems, they can give rise to quantum spin liquids that lack conventional order and are defined by coherent superpositions of an extensive number of many-body configurations. Realising and probing such exotic states experimentally is an outstanding challenge both in solid-state and synthetic quantum systems, not least due to the difficulty of detecting the fragile coherences between many-body states. Here, we report a large-scale (>3,000 sites) realisation of a two-dimensional U(1) lattice gauge theory with ultracold atoms in a square optical superlattice and demonstrate non-equilibrium preparation of extended regions of U(1) quantum spin liquids. We demonstrate Gauss's law validity in a quench experiment, enabled by a new microscopy technique for detecting doubly occupied sites. We observe characteristic real-space correlations and momentum-space pinch points, hallmarks of the emergent U(1) gauge structure. Using round-trip interferometric protocols, we directly observe large-scale coherence between many-body configurations, providing strong evidence for quantum spin liquid regions extending over ~100 lattice sites. Our results establish non-equilibrium quantum simulation protocols as a powerful route for accessing and probing exotic, highly-entangled states beyond those hosted by the engineered Hamiltonian in thermal equilibrium.

This paper presents an analysis of the applicability of a quantum amplitude redistribution algorithm to the data filtering problem and the results of modeling the algorithm's operation in comparison with a median filter.

The emergence of classicality is fundamentally driven by the interaction between a quantum system and its environment. Foundational open-system approaches, notably the Caldeira-Leggett model, successfully captured how these interactions lead to macroscopic effects like quantum dissipation and decoherence. However, these approaches often leave the precise definitions of classicality and quantumness ambiguous. In quantum information theory, this boundary is a heavily scrutinized question, and Kochen-Specker contextuality emerges as a hallmark of nonclassicality. It is therefore natural to investigate whether decoherence can actually suppress this property. Taking this path creates an apparent conundrum, once there exist two distinct manifestations of quantum contextuality: state-dependent and state-independent ones. While state-dependent contextuality naturally vanishes under state degradation, state-independent contextuality could persist for any quantum state, since it shows up even for the maximally mixed state! In this paper, we resolve this apparent paradox by analyzing sequential measurement implementations of the paradigmatic Klyachko, Can, Binicioğlu, and Shumovsky (KCBS) and Peres-Mermin prepare-and-measure scenarios under the influence of depolarizing channels. By introducing noise both prior to and in between measurements, and by analyzing the resulting sequential correlators in both the Schrödinger and Heisenberg pictures, we show how open-system dynamics suppress the correlations required to witness contextuality, leading to classicalization.

The Varenna school is a hub where generations of physicists, including numerous Nobel laureates, have shaped the field, often through collaborative exchanges across political and cultural boundaries. We examine the scientific legacy of Enrico Fermi and its influence on modern atomic, molecular, and optical physics. Beginning with Fermi's 1954 lectures at the Varenna school, key developments are traced from high-energy physics to laser spectroscopy, precision metrology, and the control of ultracold atoms. Milestones such as Doppler-free spectroscopy, optical frequency combs, Bose-Einstein condensation, and degenerate Fermi gases are highlighted as turning points leading to quantum simulation and quantum computation. Fermi's early advocacy for building a computer, rather than buying it, can be viewed as a precursor to today's efforts in quantum science and technologies. This historical trajectory and legacy continues to inform current research in quantum matter and information science.

Cooperative emission is well understood for idealized symmetric systems, but its limits in spatially extended, free-space ensembles remain an open question. Here, we derive a universal law for the scaling of the maximum photon emission rate with system size that unifies both ordered arrays and disordered atomic clouds in arbitrary dimensions at fixed density. We demonstrate that, for a fixed atomic density, the maximum emission rate scales universally as the product of the atom number and the system's optical depth, with the latter encoding the dimensional scaling across all regimes from independent emission to the Dicke limit. Furthermore, we establish a scaling law for directional detection, revealing that the observed rate depends on the detector's numerical aperture: small apertures yield Dicke-like quadratic scaling, whereas large apertures recover our integrated universal bound. Our results establish optical depth as the parameter governing many-body cooperative emission in both ordered and disordered ensembles, and reveal that directional and total-emission scalings must be carefully distinguished in experimental settings.

We obtained quasinormal modes and greybody factors of the quantum Oppenheimer-Snyder (QOS) black hole with quintessential dark energy and string clouds. We compute the effective potentials of the scalar and the vector perturbation fields and analyze their graphical behavior. For this, we compute quasinormal mode frequencies of the QOS black hole, that is surrounded by quintessence dark energy and a cloud of strings and this is done by 6th-order Wentzel-Kramers-Brillouin (WKB) approximation method. We observe the quasinormal frequencies by analyzing both the quintessence parameters and the string cloud parameters. We also examine the GFs by analyzing the role of string cloud ($\tilde{a}_{e}$), quantum deformation ($\tildeσ_{e}$), and quintessence ($\tilde{h}_{e}$) parameters. It is found that GFs are significantly influenced by the parameters of $\tilde{a}_{e}$, $\tildeσ_{e}$, and $\tilde{h}_{e}$. In the case of scalar perturbation, we also give strict limits of the GFs and check the influence of the parameters of $\tilde{a}_{e}$, $\tildeσ_{e}$, and $\tilde{h}_{e}$.

We investigate whether retarded scalar propagators on causal sets can be expressed in terms of the link matrix $\mathbf{L}$. For Poisson sprinklings into $1+1$ dimensional Minkowski spacetime, we show by asymptotic analysis and supporting numerical simulations that the averaged massless retarded propagator is naturally associated with a normalized exponential exp$(\mathbf{L})$. We then extend the construction to the massive case via the usual mass-scattering series and obtain good agreement with the continuum propagator after averaging. Finally, we discuss the inverse kernel exp$(-\mathbf{L})$ as a possible candidate for a discrete d'Alembertian.

We place several known indicators of Kochen--Specker contextuality -- the KCBS correlator $χ$, the contextual fraction $\CF$, the Shannon-entropic $n$-cycle inequality of Chaves and Fritz, and the operational commutator witness $D$ of Paper~I -- into a single projector-geometric framework organized around the overlap matrix $\Tcal_{ij} = d^{-1}\tr[(\hat P_i \hat Q_j)^2]$, where $\hat P_i$ and $\hat Q_j$ are the joint-eigenspace projectors of the two compatible observable pairs within a measurement context. The state-independent scalar content of $\Tcal$ is carried by two independent contractions: the mutual information energy $E = \sum_{ij}\Tcal_{ij}$ of Paper~I (equivalently, its logarithmic form $S_2 = -\log_2 E$), and the Maassen--Uffink extremal overlap $c_\MU = \max_{i,j}|\langle a_i,b_i | c_j,b_j\rangle|$. We prove that $S_2$ is non-increasing under coarse-graining, that $S_2(\Gcal) > 0$ is a necessary configuration-level condition for observable contextuality, and that the additive composition $S_2(\Gcal) = \sum_αS_2(\Gcal_α)$ is exact for the KCBS pentagon. We further show that in the spin-$1$ realization of the KCBS pentagon, a shared $m_s=0$ eigenstate in each context forces $c_\MU = 1$, rendering every Maassen--Uffink-type bound trivial -- a structural mechanism that makes explicit why outcome-entropic uncertainty relations based on $c_\MU$ are silent on KCBS contextuality, while $S_2 \approx 2.7266$~bits throughout. Applied to KCBS and CHSH, the framework identifies regimes in which every state-dependent witness considered here is silent yet $S_2(\Gcal) > 0$ by an amount set by the projector geometry alone.

We study a renormalization-group (RG) improved Schwarzschild-like black hole (BH) whose lapse interpolates between a classical Schwarzschild exterior and a quantum-smoothed interior governed by a cutoff scale $ξ$ and an interpolation parameter $γ$. We work out the horizon structure together with the photon sphere and shadow radius $R_{\mathrm{sh}}$, set up the scalar, electromagnetic, and Dirac Regge-Wheeler-Zerilli problems in a unified treatment, and compute fundamental and overtone quasinormal modes via sixth-order WKB cross-checked against time-domain ringdown. We examine Strong Cosmic Censorship (SCC) at the inner Cauchy horizon -- a horizon the improved geometry generates without charge or rotation -- and find the ratio $β=|\mathrm{Im}\,ω|/κ_-$ to be multipole-independent at the $6\%$ level, following $β\simeqλ_{L}/κ_{-}$. A thermodynamic analysis uncovers a Davies-type phase transition of the outer horizon and a nontrivial Weinhold-Ruppeiner geometry on the $(S,γ)$ slice; the Schwarzschild $T_{H}\propto 1/r_{+}$ decay is replaced by a bell-curve profile peaking at $T_{H}^{\max}\simeq 0.062$. A scan of the $(ξ,γ)$ plane summarizes the joint behavior of the shadow, the scalar barrier, the SCC ratio, and $T_H$, and exhibits a thin crescent adjacent to the extremal boundary where Christodoulou-SCC marginally fails. A comparison against Bardeen, Hayward, and Bonanno-Reuter BHs places the improved Schwarzschild BH as the most Schwarzschild-like of the regular-BH family at matched perturbation scale, with shadow-degenerate imaging against Hayward and Bonanno-Reuter at the $1\%$ level. We close with the sparsity of the Hawking flux and the energy-emission rate, both depending cleanly on $(ξ,γ)$ through a single auxiliary function tied to the outer-horizon surface gravity.

We present a novel architecture-aware transpilation method for exact general unitary gate synthesis on superconducting quantum hardware. Our approach is tightly integrated with the optimized block-ZXZ decomposition, exploiting its recursive structure to make hardware-aware decisions at each level of the recursion rather than treating transpilation as an independent post-processing step. The method introduces three key techniques: a greedy qubit mapping strategy that minimizes pairwise distances between physical qubits, an adaptive Gray code selection combined with qubit swapping that optimizes the construction of uniformly controlled Rz gates for the target topology, and a heuristic for reducing CNOT gates by exploiting the structure of long-range CNOT ladders. We benchmark our method against TKet, Qiskit, and Pennylane on the 20-qubit IQM Garnet (square lattice) and the 156-qubit IBM Marrakesh (heavy-hex) architectures with qubit counts ranging from 3 to 11. Our method achieves CNOT count reductions of up to 36 percent on the IQM Garnet and up to 34 percent on the IBM Marrakesh compared to the best competing transpiler, while simultaneously achieving transpilation speedups of up to 553x. Furthermore, our method is the only one capable of transpiling circuits beyond 10 qubits within a 30-minute time limit across both architectures.

This work introduces a novel quantum algorithm for gradient-based edge detection that operates entirely within the quantum circuit model. Grayscale images are encoded using the Novel Enhanced Quantum Representation (NEQR), allowing exact arithmetic on pixel intensities. Directional gradients are computed by generating superpositions of neighboring pixels via cyclic shift operations and performing subtraction with an exact quantum arithmetic circuit. To refine accuracy, we introduce a direction-aware shifting mechanism that aligns edges with the darker side of intensity transitions. Our novel Quantum Partitioning Algorithm enables efficient in-place thresholding of edge candidates. This work exhibits polynomial-time improvements and optimizes the ancilla count compared to previous NEQR-based quantum edge detection algorithms. These results demonstrate a resource-efficient and fully quantum approach to edge detection, highlighting a practical quantum advantage in image processing.

We investigate whether Scalar-Tensor-Vector Gravity in its weak-field modified gravity form can account for the cluster-scale inverse-square force law inferred from recent kinematic Sunyaev-Zeldovich measurements of cluster pairwise motions. The starting point is the X-COP cluster fit of STVG-MOG, for which a representative baryonic cluster mass $M\sim 10^{15}M_\odot$ together with parameters $α\sim 9.11$ and $μ\sim 0.196~{\rm Mpc}^{-1}$ provides a successful description of cluster dynamics without particle dark matter. We extrapolate this fit to the separation range $30$ to $230~{\rm Mpc}$, relevant for the pairwise kSZ analysis. Since the Yukawa transition length $μ^{-1}\simeq 5.1~{\rm Mpc}$ is much smaller than these separations, the STVG-MOG acceleration law reduces to an effective inverse-square form. This explains why the theory can satisfy the observed Newtonian behavior while remaining distinct from MOND-like long-distance modifications. We derive the corresponding pairwise velocity curve and show that, after fitting a single overall kSZ amplitude, the extrapolated STVG-MOG prediction reproduces the measured trend of the pairwise kSZ data. The analysis shows that the X-COP cluster fit and the cosmological-scale kSZ force-law result are mutually consistent within STVG-MOG.

We develop a full-covariance formalism for pulsar timing array(PTA) -- astrometry verlap reduction function (ORF) estimators and use it to forecast graviton-mass constraints from a nanohertz stochastic gravitational-wave background (SGWB). Analytic covariance expressions are derived for auto- and cross-channel ORF estimators, including signal-signal, noise-noise, and signal-noise contributions, and are validated against numerical simulations. For an observational configuration with sensitivities comparable to NANOGrav and Gaia, we obtain an expected joint 90\% upper limit of $m_g<4.41\times10^{-24}\,\mathrm{eV}/c^2$, which remains PTA-dominated and lies at the same order of magnitude as the existing NANOGrav 15-year PTA-only bound. For a future-like configuration with sensitivities comparable to the SKA and Theia/Gaia-NIR, the astrometric channels contribute significantly to the constraining power, and the joint limit improves to $m_g<0.48 \times 10^{-24} \, \mathrm{eV}/c^2$. These forecasts indicate that PTA -- astrometry multichannel inference provides a viable avenue for improving graviton-mass constraints under next-generation observational conditions.

We study quantum gravimetry when the interrogation time carries intrinsic uncertainty, motivated by a fundamental limit on temporal resolution associated with the energy--time uncertainty relation. For linearly gravity-coupled gravimeters, we obtain the effective gravity information by profiling the interrogation time from the two-parameter quantum Fisher information (QFI) matrix. In this class, the time-information block is quadratic in the gravitational parameter, and for quadratic background dynamics, the gravity--time cross term becomes affine in $g$. These properties yield a normalized expression for the fraction of standard single-parameter gravity QFI that remains once interrogation time is treated as a nuisance parameter, with an affine numerator and a Lorentzian denominator. We work out these results in three benchmark models: a freely falling Gaussian wavepacket, the Kasevich--Chu light-pulse atom interferometer, and an idealized closed-unitary optomechanical model. The Gaussian free-fall benchmark yields an exact closed-form expression for the effective gravity information and shows explicitly how nuisance-time profiling suppresses the momentum-spread-dependent part of the standard single-parameter gravity QFI. In the Kasevich--Chu interferometer, internal state population readout gives a rank-deficient measured two-parameter geometry unless independent timing information is supplied, whereas full access to the final motional and internal states restores a full-rank geometry with retention controlled by the competition between initial velocity spread and gravitationally accumulated motion. In atom-interferometric benchmarks, the framework yields explicit conditions for minimizing nuisance-time information loss, together with corresponding constraints on momentum spread, spatial localization, and long-interrogation-time operation.

We propose a mathematically rigorous unified framework for hybrid quantum mechanics that systematically combines algebraic deformation and spatial non-locality within a single operator formalism. By constructing a self-adjoint hybrid kinetic operator through spectral calculus, we derive exact generalized uncertainty relations that interpolate between $q$-deformed and fractional quantum mechanical bounds. Furthermore, we establish a rigorous quantum speed limit theorem for the hybrid Hamiltonian, revealing how deformation parameters, fractional orders, and external potentials tune the fundamental evolution rate of quantum states. We prove that algebraic deformation accelerates coherent dynamics through discrete momentum quantization, while fractional non-locality induces spectral broadening that suppresses evolution speed. The framework recovers standard quantum mechanics, $q$-quantum mechanics, and fractional quantum mechanics as limiting cases, and provides explicit phenomenological signatures for experimental discrimination in trapped-ion, superconducting, and cold-atom platforms.

This paper introduces a conceptual framework of technology-dependent ternary quantum gates that could be implemented and fabricated into future superconducting and photonic quantum systems for operating 3-valued quantum bits (qutrits). The "technology-dependent" means that such ternary quantum gates are on-purpose designed analogy to the contemporary binary quantum gates. Conceptually, the final built technology-dependent one-, two-, and three-qutrit gates are Chrestenson, Z3, 01, 02, 12, +1, +2 (including their corresponding inverse and controlled gates), a non-phase relative SWAP gate, and a cost-effective Toffoli gate, which is a generic ternary Galois Field (GF3) multiplication and addition circuit.

We consider a holographic Einstein-Maxwell model in five dimensions with pure gauge and mixed gauge-gravitational Chern-Simons terms to study anomaly-induced transport in the presence of explicit symmetry breaking. We include the full backreaction of the scalar field and gauge fields on the metric and compute the anomalous transport coefficients using Kubo formulae involving charge and energy current correlators. Our findings reveal that, in the presence of explicit symmetry breaking, anomaly-induced transport phenomena can extend beyond anomalous currents and affect the non-anomalous sector as well. The transport coefficients exhibit a clear dependence on the symmetry-breaking mass parameter, highlighting the interplay between quantum anomalies and explicit symmetry breaking in holographic systems.

The Standard Model (SM) predicts the universality of lepton couplings with the electroweak gauge bosons. Semileptonic decays of $b$-hadrons provide a powerful framework for testing the SM and probing possible New Physics effects. In particular, the processes mediated by charged-current interactions benefit from a relatively large branching fractions and theoretically well-controlled hadronic matrix elements. This contribution presents three recent results from the LHCb experiment: the first measurement of the ratio of branching fractions $\mathcal{R}(D^{**})$ using $B^{-} \to D^{**0} τ^{-} \barν_τ$ decays, the determination of the branching fraction for $Λ\to p μ^{-} \barν_μ$ and the extraction of form-factor parameters from $B^0 \to D^{*-} μ^{+} ν_μ$ decays.

The nanohertz gravitational waves (GW) observed by pulsar timing arrays may originate from a cosmological first-order phase transition (PT) at $\sim$ 100 MeV. Taking this possibility seriously motivates the question: why 100 MeV? We point out that a PT at exactly those scales is predicted by the generation of the baryon asymmetry from a dark asymmetry via resonant neutron-dark matter oscillations, and we prove that this PT can induce an observable GW signal compatibly with all experimental constraints. This proposal predicts dark matter self-interactions close to their observational upper limits and lowers the maximal expected mass of neutron stars. Independently of GW, this baryogenesis mechanism is tested by searches for missing-energy at the LHC and for neutron decays. We keep the model consistent with big-bang nucleosynthesis by adding heavy neutral leptons below 100 MeV, which generate neutrino masses and can induce further experimental tests.

We discuss some new results on the construction of supersymmetric solutions of Type IIB supergravity of the form ${\rm WAdS}_3\times{\rm WS}^3\times T^4$, ${\rm WAdS}_3$ and ${\rm WS}^3$ denoting \emph{warped} anti-de Sitter spacetime and sphere, respectively. The distinctive feature of these backgrounds is that, in spite of them being supersymmetric, the warpings of the two factors are described by independent parameters. We illustrate how some of these geometries, characterised by a lightlike warping of the anti-de Sitter factor, arise in the near-horizon limit of a regular, asymptotically locally flat configuration of D-branes and fluxes. Central to the construction of the latter solutions is the use of two independent TsT transformations. We also give a new class of supersymmetric solutions of the general form ${\rm WAdS}_3\times{\rm WS}^3\times T^4$, which has not been published yet. They feature warpings of the anti-de Sitter factor of the lightlike, spacelike and timelike types. We discuss their properties.

We develop a precision framework for doubly charmed baryon decays based on symmetry-protected observables and effective-field-theory diagnostics. In nonleptonic $Ξ_{cc}$ decays, we construct a null combination of widths that vanishes in the heavy-diquark factorization limit, providing a direct probe of nonfactorizable QCD dynamics. For semileptonic decays, we identify the light-lepton universality ratio $R_{Ξ_c}^{μe}$ as an observable in which leading hadronic normalization cancels at the amplitude level, yielding direct sensitivity to short-distance charged-current interactions. Percent-level precision probes $|C_{V_L}^μ|\sim \mathcal{O}(10^{-2})$, whereas $\mathcal{O}(10^{-1})$ deformations induce order-one deviations. Scalar contributions remain parametrically suppressed. Combining baryonic and mesonic inputs, we show that $Ξ_{cc}$ decays constrain the same short-distance interaction with complementary scaling, lifting degeneracies inherent to meson-only analyses. Mapping to a charged-vector benchmark demonstrates sensitivity to multi-TeVnew-physics scales. These results establish doubly charmed baryons as an independent probe of charged-current interactions beyond the Standard Model.

We identify a novel next-to-leading order renormalization effect in the dimension-six SMEFT with direct phenomenological impact. The Higgs-Yukawa operator that modifies the top-Higgs coupling $κ_t$ induces a shift in the $ W $ mass at two-loop order through a large anomalous dimension, rendering electroweak precision observables a powerful indirect probe of $κ_t$. We show that this effect is essential for the consistent interpretation of data from future Tera-$Z$ and Giga-$W$ factories such as FCC-ee. The effect is realized in a simple renormalizable two-Higgs doublet model.

Accurate modeling of the absorption of tens-of-MeV $ν_e$ on $^{40}$Ar is needed to enable measurements of astrophysical neutrinos using large liquid argon time projection chamber (LArTPC) detectors, such as those planned for the Deep Underground Neutrino Experiment (DUNE). We revisit the MARLEY neutrino interaction model used in present estimates of DUNE sensitivity to supernova and solar neutrino signals. Multiple theoretical refinements are pursued, especially in the unbound continuum region of nuclear excitation energy. Inclusive charged-current neutrino-argon cross sections are calculated using a hybrid strategy. Nuclear transitions to unbound states are treated using a Hartree-Fock Continuum Random Phase Approximation (HF-CRPA) model, including forbidden contributions. Allowed transitions to low-lying discrete levels are also included using indirect measurements and approximate corrections for the momentum transfer dependence. Exclusive predictions are obtained by coupling these calculations with a statistical nuclear de-excitation model. The impact on observables of interest for DUNE and similar experiments is examined in terms of both total and differential cross sections. Our refined calculations predict a lower allowed portion of the cross section relative to the prior MARLEY model. At neutrino energies appreciably below 100 MeV, the inclusion of forbidden transitions does not fully compensate for the loss of allowed strength. For a representative neutrino burst from a galactic core-collapse supernova, our results suggest that MARLEY 1.2.0 overestimates the event yield in a DUNE-like detector by approximately 20%. However, because this overestimation is more severe at backwards angles, use of the charged-current $ν_e$-$^{40}$Ar reaction for supernova pointing may be more feasible than previously expected.

We present GLoop, a Fortran90 computational framework that allows one to compute by Monte Carlo a certain class of higher-loop integrals in terms of lower-loop building blocks. This is based on a recently introduced method that enables the numerical computation of integrals defined by i epsilon deformations acting on single pole singularities without the need for an explicit analytic contour deformation. We provide detailed, worked-out examples and routines to show how our strategy works. These can be used as a starting point for the reader to develop her/his own calculations.

The impact of the sub - nucleon fluctuations on the Deeply Virtual Compton Scattering (DVCS) process at the Electron - Ion Collider (EIC) is investigated considering proton and nuclear targets. Assuming the hot - spot model, we estimate the energy dependence of the coherent and incoherent cross - sections for different values of the photon virtuality and atomic number. Predictions for the $t$ - distributions are also presented. We demonstrate that the sub - nucleon fluctuations in the proton, as described by the hot - spot model, implies a turn - over in the energy dependence of the incoherent cross - section, with the position of the maximum being dependent of the photon virtuality. Our results indicate that the ratio between the coherent and incoherent cross - sections increases with energy, atomic numbers and for smaller values of $Q^2$. Moreover, we predict a maximum in the $t$ - distribution of the nuclear incoherent cross - section at a fixed center - of - mass energy, which is dependent on the atomic number and $Q^2$.

We introduce equivariant localization as a method for computing the action of probe branes in supergravity backgrounds. We apply it to supersymmetric probe D3-branes in type IIB supersymmetric spacetimes obtained by uplifting the Kerr-Newman-AdS$_5$ black hole on a toric Sasaki-Einstein space. Depending on the cycles they wrap, such branes represent non-perturbative corrections to or defect operator insertions in the superconformal index of a large family of four-dimensional $\mathcal{N}=1$ quiver superconformal field theories. The resulting action reduces to equivariant integrals and can be evaluated entirely from toric data.

The first evidence for the decay $B^+\toπ^+ e^+ e^-$ is reported using proton-proton collision data recorded by the LHCb experiment at centre-of-mass energies of 7, 8 and 13 TeV, corresponding to an integrated luminosity of 9 fb$^{-1}$. A signal excess with a significance of 3.2$σ$ is observed and the branching fraction is measured to be $\cal{BR}(B^+\toπ^+ e^+ e^-) = (2.4\,{}^{+0.9}_{-0.8} \,{}^{+0.4}_{-0.2}) \times 10^{-8}$, where the first set of uncertainties is statistical and the second is systematic. The result is consistent with the Standard Model expectation.

Multi-messenger astronomy requires real-time systems capable of rapidly responding to external alerts and sharing significant detections with partner observatories. KM3NeT, a deep-sea Cherenkov neutrino telescope in the Mediterranean Sea, is actively contributing to these efforts through a dedicated real-time analysis framework. It comprises two detectors - ARCA, optimised for TeV-PeV neutrinos, and ORCA, for GeV-TeV neutrinos - both also sensitive to MeV neutrinos from core-collapse supernovae, providing a wide field of view and an almost continuous duty cycle. The framework performs low-latency event reconstruction and classification, follows up external alerts from the multi-messenger community, monitors for core-collapse supernova neutrino bursts, and autonomously identifies and distribute cosmic neutrino alerts. Now in advanced commissioning, the KM3NeT real-time alert system represents a major step toward rapid, coordinated multi-messenger observations.

The electron ionization predicted by the Migdal effect in dark matter-nucleus scattering enhances experimental sensitivity to sub-GeV dark matter. In this work, we demonstrate that lower-energy electron excitations provide a novel and promising pathway, enabling the detection of even lighter dark matter particles previously considered inaccessible for direct searches. Direct detection experiments employing a superfluid $^4$He target can exploit this channel by observing electronic excitations via UV-photon emission. We calculate the resulting event rates and find that electron excitations induced by the Migdal effect make it possible to probe dark matter-nucleus scattering for dark matter masses as small as a few MeV.

We propose the first examples of scale-separated vacua with extended supersymmetry. They arise as circle compactifications of four-dimensional vacua of massive type IIA supergravity with scale separation, upon introducing additional fluxes and sources. We provide both the ten-dimensional solutions and the three-dimensional effective descriptions in terms of Kähler potential and superpotential. The conformal dimensions of the putative dual two-dimensional field theory appear not to be integers. The superpotential for the additional fluxes of one of our models was guessed by ChatGPT and, to the best of our knowledge, it does not appear in existing literature. Should these vacua be solutions of string theory, they would allow to address the open problem of scale separation from the vantage point of extended supersymmetry.

The transverse momentum spectrum of charged particles in ultra-relativistic heavy-ion collisions fluctuates event-by-event, encoding signatures of underlying collective dynamics. Such fluctuations originate from a combined effect of thermal and geometric fluctuations in the initial state. We present a direct decomposition of these spectral fluctuations through principal component analysis performed on the joint covariance structure of normalized spectrum, mean transverse momentum and elliptic flow squared. The first two leading modes explain 99.5\% of the total variance, and are orthogonally rotated by imposing physical constraints motivated by the initial state thermal and geometric response. The resulting thermal and geometric modes bear direct analogy with the vibrational normal modes of a linear triatomic molecule. The thermal mode entirely drives the experimentally measured $v_0(p_T)$, while the geometric mode contributes substantially to $v_{02}(p_T)$ in non-central collisions, providing a transparent explanation of its characteristic low-$p_T$ sign change. The study establishes the first physically motivated interpretation of principal component modes in the field of heavy-ion collisions and provides an experimental window into the thermo-geometric structure of the QGP initial state.

This chapter, to appear in the section on QCD under extreme conditions within the Encyclopedia of Nuclear Physics, aims to provide a pedagogical introduction to the physics of quarks and gluons in the presence of high temperature, nonzero (isospin) density and strong background electromagnetic fields. Extreme conditions of these types are relevant for the description of high-energy heavy-ion collisions, neutron stars and their mergers, as well as the evolution of the early Universe in its first microsecond. Most of the existing results on this topic have been obtained by means of first-principles simulations of the discretized theory of the strong interactions, lattice Quantum Chromodynamics (QCD). This lays the focus of this review chapter, although various calculations within effective theories of QCD -- most notably chiral perturbation theory -- are also discussed. Furthermore, we provide an outlook concerning open questions and yet uncharted parameter regions within this fascinating system.

We present a unified description of heavy hybrid hadrons based on a constituent-gluon picture embedded in the Born-Oppenheimer (BO) framework. In this approach, the gluonic excitation is treated as a dynamical quasiparticle with a mass generated by instanton-induced interactions. We propose a simple variational derivation of the BO potentials. The main focus of the paper is the derivation of light-front wave functions for hybrid systems, specifically for the $ccg$ and $qqqg$ cases. We employ both variational methods and numerical solutions of the Schrödinger equation in momentum representation. Using the resulting wave functions, we compute the gluon PDFs for these systems.

We study two-point correlation functions of heavy scalar fields in the Nariai geometry. Utilizing the heat kernel formalism, we obtain this result from a geodesic approximation to the two-point function on a product of spheres. By analytically continuing one of the spheres, we obtain the correlation function in the Nariai geometry. This result involves a sum over complex geodesics, extending previous results in pure de Sitter space. We emphasize the important role of the phase of each geodesic contribution, which needs to be taken into account to avoid spurious singularities in the correlator.

We investigate the spreading of entanglement, and entanglement memory effects, in two dimensional conformal field theory (CFT) propagating on evaporating black hole backgrounds. Memory effects leading to late-time spikes in mutual information for widely separated intervals are well known in CFTs admitting a quasiparticle description. In this work we examine the effect of black hole scrambling on late time mutual information spikes for disjoint intervals in free fermion CFT prepared in a thermofield double state. Late-time entanglement revival is driven by island-induced purification of modes in the union of the intervals. We show across two distinct 2d gravity models, Jackiw-Teitelboim (JT) gravity and the Russo-Susskind-Thorlacius (RST) model, that parametrically dialing up black hole scrambling time smooths out and suppresses entanglement spikes until they disappear at a critical scale, interpolating between free quasiparticle and maximal scrambling pictures. At the critical point, the interval lengths are exponential in black hole scrambling time. We further find a very closely related effect manifest as an entanglement dip for a single interval in a single-sided evaporating RST black hole.

In the background of a Kerr-Newman black hole, the motion of a scalar particle is integrable by virtue of an extra conserved charge known as Carter charge. When the particle is endowed with spin, it is known that another conserved charge, the Rüdiger charge, maintains the integrability at least at low orders in the spin magnitude. We explore the extent of this integrability in a simpler model where both the source and the probe are root-Kerr particles, the non-gravitating limit of the Kerr-Newman black hole. At the leading order in the probe charge, the integrability holds to all orders in the spin magnitude if the interaction vertices of the probe are dictated by the Newman-Janis shift. At the second order in the probe charge, the integrability can be extended to the spin-squared order but begins to fail at the spin-cubic order. An argument based on asymptotic conservation suggests that it is impossible to restore the conservation at the spin-cubic order by a further deformation of the probe action. We compare our results with related observations for Kerr black holes with gravitational interactions.

We point out that constraints on $ΔN_\mathrm{eff}$ reported by the ACT collaboration in their DR6 data release are surprisingly sensitive to the assumptions made about the initial power spectrum from inflation. The ACT collaboration reports no evidence of new light degrees of freedom alongside a low value of the expansion rate, thus confirming the Hubble tension. However, as we show here, when considering self-interacting dark radiation and including running, $α_s$, and running of the running, $β_s$, of the spectral index $n_s$, the picture changes significantly. Confronting this extended model with Planck, ACT DR6, DESI DR2, and uncalibrated Pantheon+ data, we find the significantly relaxed bound $ΔN_\text{eff}< 0.58$ at 95$\%$ CL, together with a $2.9 σ$ ($2.6 σ$) preference for $α_s>0$ ($β_s>0$), while the Hubble tension is reduced to $2.2 σ$ with only three more parameters compared to $Λ$CDM. If the dark radiation fluid is initially coupled to dark matter, and undergoes dark radiation-matter decoupling (DRMD) around matter-radiation equality, predicting dark acoustic oscillations with drag horizon $r_{d,\mathrm{DAO}} \approx 60 \,\mathrm{Mpc}/h$, the bound is further relaxed to $ΔN_\text{eff}< 0.68$ at 95$\%$ CL, reducing the Hubble tension below $2σ$. We also discuss how $α_s$ and $β_s$ could naturally appear in inflationary scenarios, possibly connected to the end of a first act of inflation. In this case dark radiation is mostly probed by scales covered by Planck and DESI, while smaller scales carry information on inflationary dynamics.

We study the dynamics of circular cosmic string loops in a spatially flat Friedmann Lemaître Robertson Walker universe within a fractional Polyakov framework that incorporates nonlocal memory effects. Allowing both the loop radius and polar angle to evolve, we obtain a coupled non-autonomous system governed by string tension, cosmological expansion, and an emergent centrifugal contribution. We show that angular dynamics plays a crucial role in determining the loop evolution. In contrast to standard scenarios where loops collapse, we identify a class of solutions exhibiting sustained expansion driven by dynamically generated angular motion. The system also displays nonlinear behavior with signatures of chaos, with the onset of chaotic dynamics closely correlated with expanding solutions. Our results demonstrate that fractional memory effects and angular degrees of freedom qualitatively modify cosmic string loop dynamics, providing new mechanisms for stability in cosmological backgrounds.

The internal structure of hadrons is characterized by form factors which correspond to matrix elements of currents. Among those, the stress-energy-momentum tensor is a universally conserved quantity providing the gravitational form factors, from which mechanical properties may be derived via the response to the space-time fluctuations. They have received much attention because of their role as moments of the Generalized Parton Distributions, where the stress-energy-momentum tensor couples to two photons, and more recently, due to the explicit lattice QCD determination for the pion and nucleon. In these lectures we attempt a pedagogical review of the topic from a purely hadronic point of view, based on the notion of dispersion relations, meson dominance, and parton-hadron duality. We show that despite the overwhelming simplicity of the approach, a rather successful description of the lattice QCD data is achieved.

We investigate the consequences of a transient phase of enhanced parity violation during inflation. Modeling this phase through a time-localized Chern--Simons-like coupling, we show that it amplifies primordial gravitational waves at small scales, producing a robust spectral shape with a blue growth of effective slope $n_T \simeq 2$, largely insensitive to microscopic details. This prediction lies in the range explored by recent pulsar timing array (PTA) analyses under cosmological power-law interpretations, while differing from the canonical supermassive black hole binary expectation. Our framework thus provides a predictive cosmological template to benchmark astrophysical versus primordial origins of the signal, consistent with cosmic microwave background bounds. The signal also exhibits large linear polarization and non-trivial Stokes correlations, corresponding to an almost maximally phase-coherent helicity state. Such features are difficult to realize in classical stochastic backgrounds, and their detection would provide circumstantial evidence for a primordial, coherently generated origin of the gravitational-wave background.

We investigate the structure of logarithmic modes in critical topologically massive gravity (CTMG) at the chiral point $μ\ell=1$ from the perspective of analytic continuation and monodromy. Starting from the degeneration of massive and left-moving graviton modes, we construct the logarithmic mode as a derivative in parameter space and show that it acquires a natural multivalued structure upon complexification of the radial coordinate. We demonstrate that this multivaluedness induces a nontrivial monodromy action on the space of linearized solutions, under which the left-moving and logarithmic modes form an indecomposable (Jordan block) representation. This monodromy is unipotent and provides a bulk realization of the logarithmic structure typically associated with logarithmic conformal field theories. We further show that the monodromy representation alone is sufficiently constraining to determine the characteristic logarithmic form and mixing structure of two-point functions, up to normalization, without assuming logarithmic conformal field theory data a priori. These results suggest a geometric interpretation in which logarithmic modes act as sources of branchlike behavior in the bulk, analogous to twist fields that generate monodromy. While this perspective is compatible with proposed connections to branched coverings, Hurwitz theory, and integrable hierarchies, establishing a precise correspondence is left for future work.

We study supersymmetric quarter-indices for corner configurations in 4d $\mathcal{N}=4$ super Yang-Mills theory with orthogonal and symplectic gauge groups. For the basic Y-junctions, we obtain exact closed-form expressions for the indices by making use of the Gustafson type integral formula and the Higgsing method. We demonstrate the equality of the quarter-indices between dual configurations, providing evidence for S-duality of the corner configurations. In the special fugacity limit, the indices admit an interpretation in terms of the vacuum characters of the W-algebras of type BCD, and the Lie superalgebra $\mathfrak{osp}(1|2N)$ as the corner vertex operator algebras.

The semileptonic decays of $B_q$ mesons ($q=u,d,s$) to $J^P=2^+$ charmed mesons are investigated. The form factors parametrising the hadronic matrix elements of the weak vector and axial-vector quark currents are evaluated using light-cone QCD sum rules with external $B_q$ state, including the 3-particle contributions. The form factors of pseudoscalar and tensor quark currents occurring in extensions of the Standard Model are also determined. The relations expected in the heavy quark limit are tested and the size of finite heavy quark mass corrections are determined for the various form factors. Results are presented for the semileptonic $B_q$ to heavy tensor meson decay rates in the Standard Model, and for the ratios testing Lepton Flavour Universality.

We present absolute frequency measurements of 2S_{1/2}-nS_{1/2} two-photon transitions with n = 8, 9, and 10 in a cryogenic beam of atomic hydrogen. Each transition has been measured with a fractional uncertainty of approximately 2.6*10^(-12). Combining the results from this work and the 1S_{1/2}-2S_{1/2} transition frequency, we extract a root-mean-square proton radius of r_p = 0.8433(31) fm and a Rydberg frequency of cR_{\infty} = 3,289,841,960,252.9(9.7) kHz. These are in good agreement with the CODATA 2022 recommended values.

At the High Luminosity Large Hadron Collider (HL-LHC), silicon pixel detectors will be exposed to radiation fluences about 5 to 10 times larger than those experienced by the current innermost pixel layers up to today. Signal loss will be the main limitation to tracking and vertexing performance due to radiation damage in hybrid pixel detectors, with the increase in leakage current and depletion voltage posing severe constraints on operating conditions. It is important to have reliable predictions for all observables - such as charge collection performance, leakage current level and breakdown voltage - after irradiation, in order to estimate operational voltage values and to test the robustness of tracking algorithms. In this paper, the predictions of Silvaco and Synopsys TCAD device simulations are compared when the surface and bulk defects and traps of the ``Perugia radiation damage model'' are included. The results are quite promising regarding leakage current, depletion voltage, electric field and trap statistics, at two distinct reference temperatures and fluences.

The measurement of the time-dependent CP-asymmetry of $B^0 \to K_S^0 π^+ π^- γ$ decays is sensitive to contributions from physics beyond the Standard Model. To translate this measurement into a constraint on new physics, it is necessary to distinguish between decay modes that contribute to the final state via a kaonic resonance that is a CP eigenstate, and those that proceed through non-CP eigenstates. This requires an amplitude analysis of the $B \to K_{res} γ\to K π^+ π^- γ$ decay, that is discussed in this article.

We present a light-front spectator model that describes the proton as an active sea antiquark paired with a composite scalar-vector spectator. Using a spatial profile based on soft-wall AdS/QCD, we fit our initial parameters to CT18NNLO data and Bacchetta-Radici extractions at an initial scale of $μ_0^2=1.0~\text{GeV}^2$. By allowing these parameters to evolve dynamically, we extend our distributions up to $μ^2=100~\text{GeV}^2$. We specifically calculate the sea-quark asymmetry at the SeaQuest scale, predicting a sustained $\bar{d}$ enhancement at high $x$ that is in excellent agreement with recent E906 measurements. Additionally, we calculate the leading chiral-even generalized parton distributions (GPDs) and evaluate the total angular momentum carried by the sea quarks, comparing our findings with established results.

The BESIII Collaboration has recently reported two measurements of the two-photon decay width of the $η_c$ meson. The 2024 result is significantly larger than most theoretical and empirical expectations, while a subsequent measurement published in early 2026 shows better agreement with the world average and conventional theoretical estimates. In this work, we study the $η_c\toγγ$ and $J/ψ\toγη_c$ processes within a contact interaction model that incorporates valence-quark anomalous magnetic moment effects, which are absent in standard treatments. Besides achieving agreement with modern lattice QCD estimates for these observables, we find that the 2024 central value for $η_c\toγγ$ lies above the range that could be accommodated by the present framework, whereas the 2026 result is naturally consistent with it.

N=2 supyersymmetric Yang-Mills theories are described in terms of a Hitchin system over a Riemann surface C. Focusing on strongly coupled Argyres-Douglas theories, we show that the corresponding flat bundle over C can be quantized such that the resulting quantum flat connection is integrable. For $sl_2$, the quantum connection takes values in $gl_2$(A) where A is an associative algebra which we explicitly describe for the cases of Painlevé I, II and IV. Moreover, we find that the quantum connection is equivalent to irregular versions of Knizhnik-Zamolodchikov (KZ) connections. Utilizing a suitable gauge transformation, one can show that the corresponding KZ equations give rise to BPZ equations.

We present recent developments in precision high energy collider physics based on the IR-improvement of unintegrable singularities in the infrared regime via amplitude-based resummation in $QED\times QCD \subset SU(2)_L \times U_1 \times SU(3)^c$. We focus on specific applications relevant to precision observables in LHC/FCC, LC, CLIC, CEPC, and CPPC physics, for which we present new results and some new issues.

We consider high energy scattering in the effective theory of the Color Glass Condensate. The most convenient degrees of freedom are Wilson lines encoding multiple gluon exchanges, whose evolution with energy follows the JIMWLK equation. Instead of using the latter, very often one resorts to a Gaussian Approximation (GA), which is known to be remarkably accurate in describing a wide class of multi-gluon correlators whose expansion in the weak scattering limit starts with an exchange of only two gluons. Here we demonstrate, both analytically and numerically, that such an approximation is not valid for correlators which start with an exchange of four gluons. As a main example, we focus on incoherent diffraction in photon-nucleus collisions and we show that the discrepancy between the JIMWLK and the GA results is driven by weak scattering and further persists in the regime where unitarity corrections begin to become important. The JIMWLK calculation leads to cross sections which are systematically larger in all kinematic regimes of interest.

Dual-readout calorimeters achieve superior energy resolution by simultaneously measuring Cherenkov and scintillation signals for event-by-event electromagnetic fraction correction, making them attractive for next-generation Higgs factories. However, if a full waveform readout is required for time-based analysis to separate Cherenkov and scintillation signals, high off-detector data rates might present challenges. These challenges can be mitigated by real-time signal processing in front-end electronics. We present a systematic comparison of machine learning (ML) and template fitting approaches for the separation of scintillation and Cherenkov light components in homogeneous dual-readout calorimeters across three representative crystal types. ML models achieve comparable signal extraction performance at lower sampling rates than template fitting. A single model trained over a range of incident particle energies demonstrates robust performance, and FPGA-compatible compression achieves latencies suitable for real-time application. This work establishes both baseline template fitting performance and ML-enhanced alternatives for crystal-based dual-readout calorimeters, offering practical pathways towards front-end feature extraction in future detector design.

The super Virasoro minimal string is defined by coupling spacelike and timelike super Liouville theories on the worldsheet. There are four different theories 0A$^\pm$ and 0B$^\pm$ depending on discrete choices on the worldsheet. We show that these theories arise naturally from quantization of 3d supergravity, and the amplitudes compute the dimension ($+$) or superdimension ($-$) of the space of $\mathcal{N}=1$ superconformal blocks modulo crossing symmetry. Both 0A$^+$ and 0B$^+$ are perturbatively dual to the same matrix integral as the bosonic Virasoro minimal string, while 0B$^-$ is dual to a matrix integral with an inverse square root singularity. We show that all non-trivial perturbative amplitudes of the 0A$^-$ theory vanish.

We introduce the kinematic lensing ratio (KiLeR), a geometric dark-energy probe from weak lensing. Combining shear ratios with intrinsic galaxy shapes inferred from kinematics, KiLeR naturally mitigates most first-order lensing systematics, including redshift errors, intrinsic alignments, and baryonic effects. The forecast on Roman shows a 192% improvement in the state-of-the-art dark energy constraints from adding KiLeR, providing independent tests of the evolving dark energy hint in the DESI DR2 BAO+CMB+SN analysis. We quantify the science requirements on systematic and statistical error control and discuss the pathways towards KiLeR observation.

A millicharged neutrino propagating through magnetic fields experiences a small Lorentz-force deflection, which induces a geometric time delay. In the ultra-relativistic regime relevant for supernova neutrinos, this delay scales as $q_ν^2 E_ν^{-2}$, where $q_ν$ and $E_ν$ denote the neutrino millicharge and energy, respectively, and thus shares the same leading energy dependence as the standard time-of-flight delay induced by neutrino mass. Motivated by this similarity, we propose a framework to reinterpret supernova time-of-flight limits on neutrino mass as constraints on neutrino millicharge. We express both effects in terms of a common $E_ν^{-2}$ dispersion coefficient and compute the millicharge-induced contribution using a line-of-sight-dependent magnetic delay kernel, extending the original SN1987A uniform-field estimate. Applying this translation to existing SN1987A limits and to projected sensitivities for future Galactic core-collapse supernova observations, we obtain bounds ranging from the $\sim 10^{-17}\, e$ level for SN1987A to the low-$10^{-19}\, e$ regime for next-generation Galactic bursts, with optimistic combinations of detector sensitivity and Galactic sightline approaching $\sim 10^{-20}\, e$. We compare these results with other bounds in the literature and discuss how nonzero neutrino mass affects the interpretation.

We present a novel framework of the post-inflationary composite axion to address the strong CP problem without the cosmological domain wall problem. Conventional composite axion models lead to the domain wall number greater than one, producing stable axion domain walls that overclose the Universe. We show that by considering a special embedding of the confining gauge group responsible for the composite axion as well as QCD into a larger product gauge group, the domain wall number is essentially set to unity in the ultraviolet (UV) theory. In this setup, small instanton effects associated with the UV gauge dynamics induce a controlled explicit breaking of the residual discrete symmetry, providing a bias term in the axion potential. As a result, the domain walls become unstable and decay sufficiently quickly, while the axion solution to the strong CP problem remains intact. We construct an explicit realization of this framework, identify a viable parameter region and analyze the axion dark matter abundance. Decays of exotic hadrons from the composite dynamics are also investigated. Our special-embedding UV completion renders the domain wall problem in composite axion models cosmologically harmless.

We address the long-standing puzzle of false vacuum decay catalyzed by black holes. Naively, small black holes with large Hawking temperatures can generate highly-boosted true vacuum bubbles in the early universe and trigger vacuum decay without any exponential suppression. Working in the thin-wall regime of the $φ^4$ and sine-Gordon models, we show that radiative losses play a crucial role in decelerating these bubbles and preventing runaway vacuum decay. We find that while the production rate is enhanced compared with vacuum tunneling in some parts of the parameter space, it is always exponentially suppressed.

We study the formation of asteroid-mass Primordial Black Holes (PBHs) as a dark matter candidate in supersymmetric extensions of the Standard Model. We show that the presence of heavy particles predicted in the Minimal Supersymmetric Standard Model (MSSM) can lead to a transient softening of the equation of state of the Universe during their non-relativistic transition, enhancing PBH formation. We compute the effective equation of state for different realizations of the MSSM mass spectrum, parametrized by three characteristic mass scales. Assuming a broad and approximately scale-invariant primordial curvature power spectrum, we evaluate the resulting PBH mass functions and compare them with current observational constraints. We find that, for supersymmetric masses above $\sim 10^5\,\mathrm{GeV}$, the PBH mass function is significantly enhanced in the asteroid-mass window, allowing PBHs to account for the total dark matter abundance without violating existing bounds. In contrast, within the Standard Model the same configurations lead to PBH mass functions that are observationally excluded. For lighter supersymmetric mass spectra, PBH production is shifted toward masses above $\sim 10^{22}\,\mathrm{g}$, which are strongly constrained by microlensing searches, thereby reducing their allowed contribution to the dark matter density.

The "Turner window" corresponds to axions with masses $\gtrsim$ 1 eV that have sufficiently strong couplings to matter to evade limits from the cooling of SN1987A. This window, through which the trajectories for the KSVZ and DFSZ QCD axions run, has been thought to be largely closed because of (1) the floor established by SN1987A cooling, (2) the absence of SN1987A-associated photons in the Kamioka II detector, and (3) the limit on neutrons produced by solar axions in the Sudbury Neutrino Observatory. We show that a more complete treatment of the axion opacity in SN1987A, significantly weakens (2). Consequently, for axion or axion-like particles with hadronic couplings, $g_{ann}$ and $g_{app}$, significant regions within the Turner window now become viable. We describe a new opportunity to constrain such hadronically coupled axions via their resonant absorption in NaI detectors. The source is the Milky Way's carbon-burning stars -- the progenitors of ONeMg white dwarfs as well as electron-capture and core-collapse supernovae -- which synthesize significant quantities of $^{23}$Na, keeping it at temperatures $\sim 10^9$K for periods up to tens of thousands of years. $^{23}$Na acts as a thermal pump to convert stellar energy into axions, which arrive at the Earth as a thermally broadened line at 440 keV. These axions can be detected via resonant absorption in NaI, with the needed detector arrays already in place, developed by DAMA/LIBRA and other collaborations to search for the elastic scattering of light WIMPs. In axion detection, NaI serves as both the target, producing $γ$'s following resonant absorption, and the detector for those $γ$'s. With current array masses and backgrounds, we find that the coupling range $|g_{app}| \sim 10^{-6.5}$--$10^{-2}$ can be covered after two years of data, including QCD axions with $m_a \gtrsim 10$ eV.

Anomalies in the non-leptonic $\bar{B}^0\rightarrow D^{(*)+}K^{(*)-}$ and $\bar{B}^0_s\rightarrow D^{(*)+}_sπ^-$ decays may be an indication of physics beyond the Standard Model, but the large deviations require strongly coupled new physics that should be visible at colliders. We explore three new directions that could lead to viable new physics models, performing a detailed collider study to examine the possible weakening of previously known constraints on additional $SU(2)_L$ doublets. Our results show that, despite the difficulty of probing $t\bar{t}$ final states, increasing the branching ratio to this decay mode does not significantly weaken the bounds on weak doublet scalars, as additionally existing charged Higgs searches are equally strong. Beyond this, we analyse a potentially large breakdown of QCD factorisation by including large-power corrections to $B$ decays, and the effect of diluting collider searches with multi-scalar extensions. We find that these typical model-building routes for constructing a viable scenario remain constrained by collider measurements, indicating that these non-leptonic anomalies remain among the most puzzling discrepancies from the SM.

We compute the probability distribution for collisional energy loss of an ultrarelativistic parton crossing a quark-gluon plasma. This collisional quenching weight has not been determined previously, unlike the average collisional loss per unit distance, although it should be a more accurate quantity to use in jet-quenching phenomenology. The quenching weight is obtained from a well-known kinetic equation which resums an arbitrary number of elastic scatterings of the energetic parton with the medium, providing a complete description of the stochastic energy exchange, including the possibility of energy gain from thermal fluctuations. The formulation also naturally extends the standard treatment of collisional energy loss to finite path lengths, which could be relevant not only for heavy-ion collisions, but also for light-ion, and possibly proton-nucleus and proton-proton collisions. We predict the quenching weight in a setup where individual elastic scatterings are described using the hard thermal loop approximation for soft exchanges, with a smooth matching to the hard domain.

We develop a systematic effective field theory framework for studying $|ΔB|=2$ interactions across energy scales. Using chiral symmetry, we construct the complete and non-redundant set of operators governing these interactions at low energies and establish their connection to the corresponding operators in the Standard Model effective field theory, as well as to their realizations in baryon chiral perturbation theory. The framework is then applied to the phenomenology of baryon-antibaryon oscillations and dinucleon decay. While oscillations probe only a limited subset of operator structures, dinucleon decay is sensitive to a significantly broader class, including transitions that are otherwise inaccessible. In addition, we identify previously unexplored dinucleon decay channels, which can probe these unconstrained regions of parameter space. More generally, this formalism makes explicit the complementarity of different probes and provides a consistent way to trace baryon-number-violating effects from their ultraviolet origin to low-energy hadronic observables, thereby providing a basis for systematic studies of ultraviolet models generating such interactions.

Quantum corrections can severely modify or even remove classical infinite distance limits in four-dimensional gravity theories with minimal N=1 supersymmetry. In this note we study this effect for infinite distance directions in the classical Kähler moduli sector of Type IIA orientifolds at fixed four-dimensional dilaton. We present several independent arguments why such infinite distance directions are absent at the quantum level. These involve the worldsheet theory of EFT strings and the putative asymptotically massless tower of particles. Key insights are provided by the uplift to M-theory on G2 manifolds, which allows for a unified treatment of quantum obstructions of seemingly different origin in the dual Type IIB/F-theory frame. Our results apply also to the Type IIA dual of infinite distance limits for which no quantum obstruction could be detected in the Type IIB frame in previous work. In conclusion, infinite distance limits in the part of the orientifold moduli space descending from a 4d N=2 vector multiplet sector are only possible at the quantum level if also some of the remaining moduli are taken to infinity.

This paper corroborates a statement that perturbative string theory does not admit a solution whose spacetime metric is de Sitter times a closed manifold, to all orders in the $α'$ and $g_s$ expansions, under the assumption that the logarithm of the sphere partition function of Euclidean quantum gravity receives a nonzero contribution proportional to $\frac{1}{G_N}$ in a saddle-point approximation. This assumption is related to the Gibbons-Hawking proposal that the entropy of the cosmological horizon of the static patch is $\frac{A}{4G_N}$. Evidence for the statement comes from independent approaches to the effective action of string theory, all of which agree that the tree-level action vanishes for closed Euclidean target-space solutions. One possible implication is that the state of the Universe will depart from an asymptotically de Sitter spacetime.

These notes are written for a memorial Session dedicated to George Lazarides. They revisit a joint work on the cosmology of a gauged axion and place it in a broader line of ideas connecting anomalous gauge symmetries, orientifold effective actions, Stueckelberg fields and dark matter. In models with an anomalous extra $U(1)$ symmetry, the Stueckelberg pseudoscalar participates in the restoration of gauge invariance through Wess-Zumino counterterms and, after electroweak symmetry breaking, may leave a physical axion-like state. Its cosmological history differs from that of an ordinary Peccei-Quinn axion: the physical field appears only after Higgs-Stueckelberg mixing, is subject to sequential electroweak and QCD misalignment, and can give an appreciable dark-matter relic abundance only when the Stueckelberg scale is sufficiently large. This perspective connects naturally with George's earlier insight that the vacuum structure of axion models must be understood together with the gauge structure in which it is embedded. I dedicate these notes to his memory, with gratitude for the collaboration and for the clarity with which he connected particle physics to the early universe.

The singly Cabibbo suppressed (SCS) decays $D\to πa_0$, with $a_0\equiv a_0(980)$, have been measured with the branching-fraction ratios $r^{+/-}_{\rm ex}\equiv {\cal B}(D^0\toπ^- a_0^+)/{\cal B}(D^0\toπ^+ a_0^-)=7.5^{+2.5}_{-0.8}\pm 1.7$ and $r^{+/0}_{\rm ex}\equiv {\cal B}(D^+\toπ^0 a_0^+)/{\cal B}(D^+\toπ^+ a_0^0)=2.6\pm 0.6\pm 0.3$, deviating significantly from the short-distance expectations $(r^{+/-},r^{+/0})\simeq (0.07,0.2)$. This discrepancy indicates the necessity of long-distance rescattering effects. In particular, the process $D\to K^*K\to a_0π$ generates ${\cal M}_s$ comparable in magnitude to ${\cal M}_d$ in the amplitude ${\cal M}=λ_d{\cal M}_d+λ_s{\cal M}_s$, with $λ_q\equiv V_{cq}^*V_{uq}$, accompanied by nontrivial strong phases essential for $CP$ violation. Consequently, the direct $CP$ asymmetries naturally arise at the ${\cal O}(10^{-3})$ level, for example, ${\cal A}_{CP}(D^0\toπ^- a_0^+)=(-0.7\pm 0.1\pm 0.1\pm 0.1)\times 10^{-3}$, ${\cal A}_{CP}(D^+\toπ^0 a_0^+)=(-1.4\pm 0.1\pm 0.1\pm 0.1)\times 10^{-3}$, and ${\cal A}_{CP}(D^0\toπ^+ a_0^-)=(-2.1\pm 0.9\pm 1.1\pm 0.4)\times 10^{-3}$. These results establish SCS $D\to πa_0$ decays as a new avenue for probing $CP$ violation.

Reconstructing the shape of the Higgs potential realised in Nature is one of the most pressing tasks of current and future colliders. This will offer deep insights into the dynamics of the electroweak phase transition and provide a unique opportunity to probe physics beyond the Standard Model (BSM). In this context, it is essential to have precise theory predictions for trilinear scalar couplings, which control the form of the potential, and for Higgs pair production processes, which are the observables that allow accessing the trilinear couplings. I summarise in these proceedings recent progress in precision calculations of both trilinear scalar couplings and Higgs pair production at the (HL-)LHC in BSM models with extended scalar sectors.

The measurement of polarization serves as an important probe to investigate the production mechanism of quarkonia, the bound state of heavy quark anti-quark (charm or bottom) pairs, in hadronic collisions. In experimental invesigations, the polarization is usually measured by analyzing the anisotropies in the angular distribution of the muons originating from the decay of the quarkonium state. In the present article, we study the charmonia ($J/ψ$) and bottomonia ($Υ(1S)$) polarization at $\sqrt{s} =7 $ and 13 TeV in proton-proton(p-p) collisions at LHC using Monte Carlo (MC) event generator model PYTHIA8, which is based on perturbative QCD. The transverse momentum ($p_{T}$) differential distribution has been calculated at forward rapidity ($2.5 < y_{μμ} < 4.0$) and the polarization parameters are estimated in Helicity and Collins-Sooper reference frames. In addition, to mimic realistic experimental conditions, we have incorporated, in PYTHIA simulations, effects like detector inefficiencies and muon momentum smearing. These contributions alter the polarization parameters, introducing an artificial degree of polarization, if not properly corrected for. The simulation results have been compared with the recent ALICE measurements for quarkonia polarization in p-p collisions at LHC energy regime.

The Jiangmen Underground Neutrino Observatory (JUNO) is a 20-kton liquid scintillator-based, low-radioactivity, multi-purpose neutrino detector located 693 meters (1800 m.w.e.) underground in the Guangdong province, China. To detect scintillation light produced in the target, the detector is equipped with 17,612 20-inch photomultipliers (PMTs), forming the Large PMT system (LPMT). In addition, 25,600 3-inch photomultipliers (the Small Photomultiplier System or SPMT) are deployed in the gaps between the LPMTs. This paper presents the design and performance of the underwater front-end electronics developed for the SPMT system. It details the individual electronics boards and their key components, the inter-board interfaces, the system-level design, and the firmware architecture that supports data acquisition and control. It also outlines mechanical and thermal integration, board validation procedures, and system performance metrics. The readout chain includes digitization of 128 PMT channels per unit, synchronized time-stamping, charge measurement, event packaging, and bandwidth management. Comprehensive validation confirms the system's readiness to meet JUNO's stringent physics goals. The underwater electronics achieve noise levels as low as 0.04 photoelectrons with minimal crosstalk (below 0.4%) and a bandwidth of 57 MB/s, ensuring reliable single photo-electron detection and operation under high-rate conditions. The SPMT system has now been fully integrated and installed in JUNO. Its commissioning and physics performance will be reported in a future publication.

We consider the anomalous dimension $γ_{gg}^{(3)}(N)$ of the twist-two gluon operator of arbitrary Lorentz spin $N$ in the quark flavor singlet sector of a general gauge theory at four loops and construct its contribution proportional to $ζ(3)$ in analytic form by applying the Lenstra-Lenstra-Lovász algorithm to the available low-$N$ moments. We exploit generalized Gribov-Liptov reciprocity, establish new self-tuning relations for the anomalous dimension matrix of the singlet sector, and inject information from $\mathcal{N}=1,4$ supersymmetric Yang-Mills theories. We also present the contribution to the rational part of $γ_{gg}^{(3)}(N)$ with color factor $C_F^2n_f^2$. Exact contributions to the four-loop splitting function $P_{gg}^{(3)}(x)$ hence resulting via inverse Mellin transformation help us to reduce theoretical uncertainties in scaling violations of parton distribution functions in QCD.

We propose that the symmetry category associated to a 2D quantum field theory with 0-form $G$-symmetry with 't Hooft anomaly $k\in H^4(BG,\mathbb{Z})$ for a large class of Lie groups $G$ is the category of twisted measurable fields of Hilbert spaces over $G$ denoted by $\mathrm{Hilb}^k(G)$, which is equivalent to the category of unitary representations of $C_0(G)$ with convolution product twisted by a multiplicative bundle gerbe labeled by $k$ denoted by $\textbf{Rep}^k(C_0(G))$. We find that the Drinfeld center of the symmetry category $\mathcal{Z}(\mathrm{Hilb}^{k}(G))$ equivalent to the category of unitary representations of the groupoid $C^*$-algebra of the Fell line bundle $Σ_k$ over the conjugation action groupoid $G//_{\rm Ad} G$, denoted by $\textbf{Rep}(C^*(G//_{\rm Ad}G,Σ_k))$, where the twist is characterized by the transgression $τ(k)\in H^2(G//_{\rm Ad}G,U(1))$. To the full generality, our framework applies to a Lie group $G$ that is a direct product of a compact connected Lie group and a number of $\mathbb{R}$ or $GL(1,\mathbb{C})$ factors. We compute the braiding of anyon lines in the bulk 3D SymTFT from this formalism. Finally we provide physical examples for abelian and non-abelian $G$, and discuss the physical consequences of flat gauging continuous global symmetries.

Different gaugings of the global symmetry of a quantum field theory are closely related to its various phases. In this work, we study candidate gaugeable symmetries by analyzing candidate Lagrangian algebra data in the Drinfeld center of a symmetry category $\mathscr{C}^k(G)$ associated to a QFT with continuous global $G$-symmetry and possible 't Hooft anomaly labeled by an integer $k$. We use the combination of the $BF$ theory and the level-$k$ Chern-Simons theory with gauge group $G$ as a semiclassical kernel-theoretic model for the corresponding SymTFT. Under two explicit assumptions, namely that this $BF{+}k$CS theory provides the relevant SymTFT model and that the common $+1$ eigenspaces of the resulting modular kernels detect candidate Lagrangian algebra data in the continuous setting, we derive candidate modular $S$- and $T$-kernels from Hopf-link and framing correlators in $S^3$ semi-classically. We then use these kernels to obtain candidate modular invariants and candidate gaugings. The resulting formulas recover the established cases and suggest a possible extension of this kernel-theoretic picture to compact Lie groups.

We give an explicit description (in component fields) of a holomorphic theory associated to a general supersymmetric background of $\mathcal N=1$ supergravity in ten dimensions. Conjecturally, this provides a sought-for holomorphic realisation of the supergravity twist in such backgrounds, generalising the minimal type I BCOV theory for Calabi-Yau manifolds. Our theory unpacks the recently introduced Courant contact model associated to a holomorphic Courant algebroid. We also show that a coisotropic reduction of this model reproduces the recent model of ref [1], which is formulated in terms of constrained fields.

Effective field theories have established themselves as key pillars of modern nuclear physics. They enable a quantitative understanding of the strong nuclear force, provided low-energy constants that parametrize short-distance physics can be determined from experimental data. This, however, often becomes prohibitively expensive due to a significant computational cost of solving the A-body problem. The computational challenge is particularly severe for three-body forces, which are at the frontier of nuclear and atomic physics and play an important role in the equation of state of neutron stars. Here we prove that for a parametric low-rank update of a Hamiltonian, the A-body problem at a fixed energy exactly reduces to a low-dimensional matrix equation regardless of the size of the Hilbert space. As a proof-of-principle, we present exact and computationally cheap snapshot-based emulators for few-body scattering and bound states. Unlike alternatives, our emulators can be used far away from the snapshot region without loss of precision and yield accurate results for parameter values not accessible using conventional solution techniques. Our approach is not restricted by the interaction type, number of particles, and methods for generating snapshots and can be applied to mitigate the computational burden of the A-body problem to a broad class of problems in nuclear, atomic, and molecular physics.

We study the asymptotic distribution of level crossings for random matrix pencils A_n+λB_n in several ensembles, including complex and real i.i.d. matrices and Gaussian/Hermitian settings. We derive a representation of the normalized log-discriminant in terms of pairwise eigenvalue interactions and formulate conditions under which its limit is governed by a deterministic potential. Under assumptions combining a uniform circular law, logarithmic tail control, and small-spacing (repulsion) estimates, we prove convergence of the empirical measure of level crossings to an explicit deterministic limit. In the complex Gaussian case these assumptions are verified (modulo a uniformity step), while in the general i.i.d. setting the results are conditional and motivated by universality theory. We further analyze the real case, showing that any limiting measure does not concentrate on the real projective line under suitable hypotheses, and discuss analogous phenomena for elliptic/Hermitian ensembles. Our results highlight the role of logarithmic energy and universality in governing spectral degeneracies of random matrix pencils.

We present a systematic technique for constructing the Lorentz-covariant structures of hadronic matrix elements of local operators. The spinor Young tableaux of the Lorentz group is employed to construct all possible structures for the matrix elements of arbitrary operators, using the relativistic wave functions and momenta of the initial and final state particles of arbitrary spin as building blocks. We obtain the form factor bases for the scalar, vector, and rank-2 tensor operators for particles with spin-$\frac{1}{2}$, $1$, $\frac{3}{2}$, and $2$, among which the general $P$ and $T$ form factors for spin-$\frac{3}{2}$ and spin-$2$ particles are presented for the first time. The independent form factor structures are also cross-checked by the non-relativistic counting and Hilbert Series method and we find there is redundant $P$ and $T$ conserved structure for spin-$2$ particles in literature. As an application, the matrix elements of general nonlocal operators can be expanded by towers of the matrix elements of local operators, and thus can be decomposed by the constructed form factor bases above.

The CONUS+ experiment has recently reported the first detection of coherent elastic neutrino-nucleus scattering (CEvNS) of reactor antineutrinos on germanium nuclei and is now entering a precision phase. The dominant uncertainty in the first measurement was the energy scale, which contributed 14% to the uncertainty of the prediction of the combined signal. We present a dedicated neutron activation campaign in which one of the new 2.4 kg CONUS+ germanium detectors was irradiated with a strong 241AmBe source, demonstrating that a contribution below 4% to the uncertainty of signal prediction is achievable. For the first time, the 71Ge M-shell X-ray line was clearly resolved at (158.7+-1.4) eVee, validating the CONUS+ energy reconstruction down to the detection threshold. This validation includes the understanding of the energy scale, the energy resolution, the trigger efficiency, and the correct separation of physical from noise events. These results establish the foundation for a future activation campaign at the Kernkraftwerk Leibstadt reactor site, strengthening the CONUS+ energy calibration and extending its sensitivity to precision CEvNS and beyond Standard Model physics measurements.

Graphical functions have emerged as a powerful framework for evaluating multi-loop Feynman integrals in perturbative quantum field theory. Defined as massless three-point position-space integrals, they reveal rich analytic structures and have enabled major advances, including the highest-loop results currently known in several quantum field theories. Their role extends to conformal field theory, and recent algorithmic developments now allow many graphical functions to be computed automatically. This review, based on graduate-level lectures held by O.S. in 2025/26 at the University of Hamburg, introduces the central ideas behind graphical functions, covering periods, Feynman residues, and the treatment of regular and singular cases in both integer and non-integer dimensions. It also discusses connections to momentum space and self-duality, and provides guidance for further study, offering a coherent entry point into a topic not addressed in standard textbooks.

The Neganov-Trofimov-Luke (NTL) effect is used by experiments based on cryogenic detectors to boost the sensitivity of light-sensitive devices down to a few optical photons. In this work we introduce a silicon light-detector technology that implements NTL amplification at millikelvin temperatures using transparent indium-tin-oxide (ITO) electrodes. The ITO electrodes enable an electric field perpendicular to the wafer surface, mitigating surface charge recombination, and thanks to their optical properties, simultaneously serve as an anti-reflective coating. By combining these two functions in a single element, the fabrication process is simplified, yielding more robust and cost-effective devices. We report on the production and characterization of the first batch of these detectors. We performed a room-temperature characterization of the ITO electrodes, verifying the structural and optical characteristics of the deposited electrodes. We then operated 2 of these devices as cryogenic calorimeters at millikelvin temperatures. Finally, we develop a consistent analytical model for the NTL gain for both ionizing particles and optical photons, successfully describing the gain dependence on the NTL bias and explicitly accounting for the partial electrode coverage of the device surface.

A measurement of the eight angular polarization coefficients, $A_0$ to $A_7$, in the cross section for the Drell$-$Yan production of two muons is presented. The analysis is based on proton-proton (pp) collision data recorded with the CMS detector at the LHC at a center-of-mass energy of $\sqrt{s}$ = 13 TeV, corresponding to an integrated luminosity of 140 fb$^{-1}$. The coefficients are determined double differentially in eight intervals of transverse momentum and two intervals of rapidity of the muon pair $μ^+μ^-$. The results are presented for the $μ^+μ^-$ invariant mass range 81$-$101 GeV and are compared with theoretical predictions calculated at next-to-next-to-leading order in perturbative quantum chromodynamics. The measurement provides relevant information about the underlying partonic dynamics and the Z boson production mechanisms.

Coherent hard exclusive reactions on light nuclei provide access to their quark and gluon structure and enable three-dimensional tomography of these complex systems. We study deeply virtual Compton scattering on a helium-4 target, including both kinematic twist-3 and twist-4 corrections, as well as next-to-leading-order corrections to the twist-2 amplitude in the strong coupling $α_s$. We show that these contributions are crucial for achieving a precise description of the data and, as a result, obtain the first tomographic image of the helium-4 nucleus at the quark-gluon level.

In this work we explore the semileptonic $τ$ decays into the axion-like particle ($a$)-meson final states within chiral effective field theory. The next-to-leading-order mixing matrix for the $π^0$-$η$-$η'$-$a$ system with the linear isospin-breaking effects, is exploited and then implemented to calculate the hadronic form factors relevant to the $τ$ decays. The resonance parameters entering the form factors are determined from fits to the experimental spectra of $τ^- \to π^- π^0 ν_τ$, $τ^- \to K_S π^- ν_τ$, and $τ^- \to K^- ην_τ$. We then focus on the predictions to the branching ratios, invariant-mass distributions, and forward-backward asymmetries from the $τ^-\to P a ν_τ$ processes, with $P=π^-$ and $K^-$. Our results provide a quantitative basis for future searches of the axion-like particle signals in semileptonic $τ$ decays.

We investigate the femtoscopic correlation functions of systems associated with the $Ω(2012)$, $Ω(2380)$, and $Ω_c(3120)$ resonances, with the aim of elucidating their internal structures. By employing effective potential models that incorporate both $s$-wave and $d$-wave interactions, we calculate the correlation functions for the relevant coupled channels. Our numerical results reveal pronounced enhancement structures in the $Ξ^0K^-$ and $Ξ_c^0K^-$ correlation functions, which provide direct evidences for the dynamically generated $Ω(2012)$ and $Ω_c(3120)$ states. Furthermore, significant low-momentum enhancements are observed in the $Ξ^{*0}K^-$ and $Ξ^{*0}K^{*-}$ channel, which are attributed to the $Ω(2012)$ and $Ω(2380)$ resonances. These theoretical calculations provide crucial insights for future high-precision measurements at the LHC and RHIC, offering a novel and independent approach to determine the dynamically generated hadronic molecular nature of these $Ω$ excited states.

One of the most important applications of topological recursion concerns spectral curves for which the functions $(x,y)$ defining the spectral curve are allowed to have logarithmic singularities. This occurs for instance for Seiberg-Witten curves and mirror curves computing Gromov--Witten invariants of toric Calabi--Yau threefolds. A recently introduced extension of topological recursion, the so-called logarithmic topological recursion, exhibits the correct behavior under certain limits of those spectral curves. In this article, we derive the dilaton equations in the setting of logarithmic topological recursion, as well as variational formulas, and provide a definition of the free energies in situations where standard topological recursion was known to fail. We present examples in which the new definition of the free energies \textit{directly} (without any computation) reproduces the full perturbative part of the Nekrasov--Shatashvili partition function of 4d $\mathcal{N}=2$ pure supersymmetric gauge theory, as well as the all-genus free energies of mirror curves of strip geometries, including in particular the topological vertex and the resolved conifold.

Radon is one of the most important sources of background in rare event search experiments, such as those searching for Dark Matter and neutrinos, due to its unavoidable production from natural uranium. In low-background experiments, radon emanation from detector materials and components accounts for a major portion of contamination. To investigate this, a radon detection system was developed at the Carleton nOble Liquid Detector Laboratory (COLD Lab). The setup consists of a stainless steel emanation chamber, a low-background ZnS(Ag) cell, and an assembly for radon transfer and collection. This setup was used to study radon emanation from materials under vacuum conditions. Additionally, a charcoal trap made of activated charcoal and equipped with a flow meter was constructed to study radon levels in nitrogen gas and the residual radon in the gas filter used in the DEAP-3600 processing system. The radon concentration in the glove box, where critical DEAP-3600 internal detector components were completed, was also calculated based on these measurements. Now calibrated and in-use, the COLD lab radon emanation counter is an essential diagnostic tool for reducing backgrounds in future rate-event search experiments.

We present a unified analysis of the $Z_c(3900)$ and $Z_{cs}(3985)$ states considering both experimental and lattice data. The study simultaneously includes the processes $e^+e^- \rightarrow J/ψπ^+π^-, J/ψK^+ K^-, D^0 D^{\ast-} π^+, (D^{\ast 0} D_s^{-}+D^0 D_s^{\ast -}) K^+$, together with finite-volume energy levels from recent lattice QCD simulations. Open-charm meson loops with triangle singularities, the $J/ψπ(J/ψ\bar{K})$-$\bar{D}D^*(\bar{D}D^*_s)$ coupled-channel interactions, and the $ππ$-$K\bar K$ final-state interaction are all taken into account. We find that pole contributions associated with the $Z_c(3900)$ and $Z_{cs}(3985)$ are indispensable for describing the data. The successful joint description of the experimental and lattice data supports the interpretation that the $Z_c(3900)$ and $Z_{cs}(3985)$ are SU(3) flavor partners within the same octet multiplet and indicates that both are resonance states. The extracted pole masses and half-widths of the $Z_c(3900)$ and the $Z_{cs}(3985)$ are $(3879.6 \pm 4.8)$ MeV and $(32.2 \pm 4.7)$ MeV, and $(3976.9 \pm 5.1)$ MeV and $(28.8 \pm 5.9)$ MeV, respectively. The ratios of the $Z_c(Z_{cs})$ couplings to the $D\bar D^*(D_s\bar{D}^\ast+D\bar{D}_s^\ast)$ and $J/ψπ(J/ψK)$ channels are also determined. A compositeness analysis indicates that, although the $D\bar D^* (D_s\bar{D}^\ast+D\bar{D}_s^\ast)$ component in the $Z_c(3900) (Z_{cs}(3985))$ state is sizable, additional components are still needed to form these exotic states.

A search for supersymmetric electroweakinos is performed using events with a low-momentum (soft) isolated track and large missing transverse momentum, targeting nearly mass-degenerate higgsino-like charginos and neutralinos. For mass splittings of 0.3$-$1 GeV, the chargino decays to the lightest neutralino and a low-momentum pion, which can produce a soft, potentially displaced track. A parameterized neural network separates signal from background using kinematic and impact parameter information. The analysis uses 138 fb$^{-1}$ of proton$-$proton collision data at a center-of-mass energy of 13 TeV recorded with the CMS detector. No significant excess above the standard model expectation is observed. At 95% confidence level, the considered higgsino model is excluded for mass splittings in the range 0.28$-$1.15 GeV and for chargino masses up to 185 GeV, setting stringent constraints on natural supersymmetry scenarios.

We investigate a pre-inflationary dynamical compactification scenario in which a higher-dimensional expanding universe evolves into an effectively four-dimensional one through curvature-assisted modulus trapping. To obtain a calculable semiclassical realization, we consider a simple five-dimensional $S^1$ compactification in a time-dependent open FRW background. Bulk quantum fields generate both the late-time Casimir contribution that stabilizes the radion and the early-time Kaluza-Klein thermal contributions relevant for the cosmological evolution. We show that negative curvature can sustain tracker-like radion evolution, allowing the radion to be trapped in a compactified vacuum before a subsequent four-dimensional inflationary phase dilutes the curvature remnant. While our analysis is performed in a 5D toy model, it illustrates a broader mechanism by which dynamical compactification can arise in a pre-inflationary higher-dimensional cosmology.

Macroscopic spin ensembles in solids are powerful platforms for quantum sensing and precision metrology. A key challenge is controlling the nuclear spin population relaxation time $T_1$, which can become prohibitively long at cryogenic temperatures due to phonon freeze-out. We demonstrate optical control of the $T_1$ relaxation time of the $^{207}$Pb nuclear spin ensemble in lead-containing ferroelectric crystals PbTiO$_3$ (PT) and (PbMg$_{1/3}$Nb$_{2/3}$O$_3$)$_{2/3}$-(PbTiO$_3$)$_{1/3}$ (PMN-PT). Using X-band electron paramagnetic resonance (EPR) spectroscopy at 10 K, we characterize light-induced paramagnetic centers created by 405 nm laser illumination. In PT, we observe paramagnetic Pb$^{3+}$ centers and their hyperfine interaction with nearby nuclear spins. In PMN-PT, we identify two populations: isotropic Pb$^{3+}$ centers and anisotropic Ti$^{3+}$ centers occupying $d$-orbitals, with spin number densities of $(2.5 \pm 1.0) \times 10^{17}$ cm$^{-3}$ and $(4.1 \pm 1.7) \times 10^{17}$ cm$^{-3}$, respectively. Power-dependent EPR measurements enable extraction of spin relaxation times. We investigate the ionization and recombination dynamics of these transient paramagnetic centers. Using saturation-recovery nuclear magnetic resonance, we demonstrate that laser illumination reduces the $^{207}$Pb nuclear $T_1$ by approximately a factor of two, from $(17 \pm 2)$ s to $(7 \pm 1)$ s at 4.6 MHz, and from $(1550 \pm 40)$ s to $(850 \pm 70)$ s at 40 MHz. We develop a model relating the nuclear relaxation rate to the density of photoinduced paramagnetic centers. This optical control of nuclear spin relaxation provides a pathway toward accelerated thermal polarization and dynamic nuclear polarization in solid-state NMR-based precision measurements, including searches for axion-like dark matter.

We revisit the time-symmetric interpretation of antiparticles and explore whether it can be connected, as a speculative benchmark, to the distinct spectral structures observed in the AMS-02 electron and positron data. Rather than modifying the local radiative loss law, we adopt the standard high-energy energy-loss form for Galactic leptons, $b(E)=b_0E^2$, and introduce the time-symmetric hypothesis at the level of the effective propagation response. In this framework, the electron sector is treated as purely retarded, while the positron sector is modeled as an effective admixture of retarded and advanced-associated components. We perform a systematic 2D scan of the parameter space governing the macroscopic admixture fraction and the exposure reduction, and demonstrate that while the characteristic peak position is degenerate, the spectral morphology, especially the breadth of the turnover region, helps identify physically preferred benchmark regions. Based on this, we adopt a representative, order-of-magnitude benchmark, in which the advanced-associated component dominates the effective admixture but experiences a ten-fold reduction in effective radiative exposure, to dissect the underlying spectral components and verify the results against AMS-02 data. While omitting complex spatial diffusion and solar modulation, this minimal phenomenological framework naturally captures the widely separated energy scales of the lepton spectra, explicitly isolating the reduced effective radiative exposure of the advanced branch as a postulated ingredient requiring deeper theoretical justification.

A simultaneous measurement of 25 substructure observables is presented using large-radius jets with high transverse momentum from proton-proton collisions at $\sqrt{s}$ = 13 TeV. The measurement is carried out on dijet events and $\mathrm{t\bar{t}}$ events enriched in Lorentz-boosted W bosons and top quarks decaying hadronically. The three data samples consist of jets with one, two, or three prongs from the showering and hadronization of a gluon or light-flavour quark, two quarks, or three quarks, respectively. The data correspond to an integrated luminosity of 138 fb$^{-1}$, recorded by the CMS experiment in 2016$-$2018. A detailed characterization of the jet substructure is provided using a 6-body basis of $N$-subjettiness observables that overconstrains the phase space of the resolved emissions in the jet. The measurements are unfolded to the level of stable particles, and an estimate of the particle-level correlations between observables is provided, ensuring that the results can be used to systematically assess and refine the modelling of radiation in jets.

We study the trajectories of massive particles in spherically symmetric black holes in arbitrary dimensions, and find certain universal features based on the topological classification of the fixed points. If the system admits a center, we find two possible outcomes: regardless of the value of the angular momentum, the center always survives, which is realized in global AdS spacetimes or, the center disappears below a critical value of angular momentum, which happens for various spherically symmetric black holes. For the latter case, we find that irrespective of the details of the black hole, there must always be a saddle point. Topological arguments show that there exists a certain critical value of energy, angular momentum and the angular velocity, where the center and the saddle coalesce. This happens at a special point in the parameter space where the trajectories are the limiting innermost stable circular orbits (ISCOs). At the critical point, conserved quantities show universal, van der Waals-like mean-field scaling typical of a second-order phase transition. The anomalous dimensions $γ$ of the double-twist operators in the CFT are found, both using AdS/CFT and through the the heavy-heavy-light-light four point correlators, giving negative and positive values for the center and saddle, respectively, including the emergence of certain non-analytic behaviour at the ISCO. For the center, we also find subleading corrections in $\frac{1}{Δ_H}$ to $γ$ in the dual CFT, and dsicuss the implications of our results.

Many extensions of the Standard Model (SM) generically introduce new sources of CP violation, which can induce observable $P$-odd and $T$-odd permanent electric dipole moments (EDMs) of nuclei, atoms, and molecules. A future observation of nonvanishing EDMs would therefore provide a sensitive probe of physics beyond the SM, while also posing a nontrivial inverse problem: identifying their underlying ultraviolet origin. In this work, we identify six representative classes of CP-violating effective operators near the QCD scale, including the QCD $θ$-term, that are particularly relevant for low-energy EDMs and can arise in a broad range of SM extensions. We show that these operator classes lead to distinct EDM patterns across different systems, thereby enabling discrimination among them through experimentally measured EDMs. We further emphasize that EDM measurements can shed light on the origin of the vacuum expectation value of the QCD axion. In particular, they may help distinguish whether a nonzero axion vacuum expectation value is predominantly induced by high-scale Peccei--Quinn symmetry-breaking effects, such as those associated with quantum gravity, or by the interplay between beyond-the-SM CP violation and the QCD anomaly.

We probe the thermodynamic phase structure of four dimensional Horava Lifshitz black holes by Lyapunov exponent analysis. For both massless and massive test particles, the Lyapunov exponent exhibits a multivalued dependence on temperature in regimes with a first-order phase transition, with distinct branches corresponding to small, intermediate, and large black hole phases, and this behaviour disappears at the critical point. The discontinuity in the Lyapunov exponent acts as an effective order parameter with critical exponent $δ=1/2$, consistent with mean-field universality. We also find that the chaos bound is generically violated below a threshold horizon radius, with the violation occurring within the thermodynamically stable phase and persisting even in the absence of a phase transition. These results establish the robustness and universality of Lyapunov exponents as probes of black hole thermodynamics in alternative theories of gravity.

Jet production at the LHC is a powerful probe of QCD, making it ideal for precision tests and determinations of QCD parameters such as parton distribution functions and the strong coupling constant. To make the most of the abundant jet production data collected at the LHC, precise calculations are required. While state-of-the-art calculations reach next-to-next-to-leading order (NNLO) QCD accuracy, a critical assessment of the remaining uncertainties arising from non-perturbative effects and missing higher orders remains crucial for correctly interpreting comparisons between theory and data. Scale variation is nearly always used to determine effects from missing higher orders. In this article, we reassess this method in the context of inclusive jet production by performing NNLO QCD calculations supplemented by small-jet-radius resummation through next-to-next-to-leading-logarithmic accuracy (NNLL). We find that NNLL resummation can have an appreciable impact on the scale uncertainty for inclusive jet cross sections, and, for some scale choices, can lead to sizeable shifts of the central cross section. We conclude that scale variations in fixed-order and resummed calculations can drastically underestimate the impact of higher orders for commonly used jet radius parameters, and that missing higher-order estimates obtained via scale variations should be considered unreliable. Our findings add further evidence to the importance of going beyond scale variations in jet and jet substructure calculations.

In this paper, we investigate the role of spectral networks in quantum Liouville theory, with particular emphasis on spectral networks of Fenchel-Nielsen-type. In the first part, we construct q-parallel transport for Fenchel-Nielsen networks through q-nonabelianisation, and compare with quantum parallel transport computed using the Moore-Seiberg formalism. This motivates a proposal for a quantum version of the NRS proposal. In the second part, we reproduce Liouville conformal blocks through the standard free-field formalism with Fenchel-Nielsen-type integration contours. However, we observe that this approach is not complete with respect to wall-crossing. We therefore develop an extension of the free-field formalism to smooth spectral coverings, with the Maulik-Okounkov R-matrix playing a central role. We conjecture that this new formalism generates the full spectrum of Liouville conformal blocks, and provides a first-principle definition for Goncharov-Shen conformal blocks.

We investigate the modification of the $J/ψ$ meson mass in asymmetric nuclear matter at zero and finite temperatures employing an effective Lagrangian approach that considers the contributions of $DD$ and $DD^*$ meson loops. The medium dependence of $D$ meson masses is determined using the hadronic chiral SU(3) model, where scalar condensates are calculated and subsequently utilized in the QCD sum rules approach. Our findings indicate that an increase in baryonic density results in a negative mass shift of the $J/ψ$ meson. This suggests that the $J/ψ$ meson is attracted to nuclear mean fields indicating the possibility of the formation of meson-nucleus bound states. Moreover, we have also determined the binding energy and absorption decay width of the $J/ψ$ meson for both the ground and excited states of $\text{O}^{16}$, $\text{Ca}^{40}$, $\text{Zr}^{90}$, and $\text{Pb}^{208}$ nuclei. These results are expected to contribute to the understanding of experimental data from upcoming studies at Jefferson lab and the facility for antiproton and ion research, where low-momentum charmed mesons can be produced and examined within nuclei.

Destructive interference between signal and background processes poses a fundamental challenge in searches for top-philic scalar resonances, significantly reducing experimental sensitivity to well-motivated extensions of the Higgs sector. Traditional bump-hunting strategies fail in this instance because interference effects invalidate the narrow-width approximation across large regions of the BSM parameter space. As a result, experimental analyses typically rely on detailed simulations to accurately model these effects throughout the full analysis chain. In this work, we consider the inverse problem in a proof-of-principle study: given an observed pattern in a discriminating distribution, what is the likelihood that it originates from a BSM scalar? To address this, we employ parametric neural networks to learn the likelihood ratio as a function of both background and key BSM parameters, based on a ratio-of-signed-mixtures framework. We perform inference by testing the compatibility of observed data with a scan over the parameter space of a minimal scalar extension of the Standard Model. While BSM parameter extraction remains inherently model-dependent, our approach provides a robust diagnostic in perturbative regimes and motivates a complementary strategy of `dip-hunting'. This strategy extends traditional bump-hunts and could point the way as we navigate towards future discoveries.

In this paper, we elaborate on the connection between leading singularities and canonical bases of Feynman integrals beyond polylogarithms. We start by discussing a notion of leading singularities in dimensional regularization, which can be generalized from the Riemann sphere to more complex geometries, and use it to demonstrate how selecting Feynman integrals with unit leading singularities necessitates introducing new transcendental functions related to the periods of the underlying geometries. Integrals with unit leading singularities in this generalized sense, satisfy $ε$-factorized differential equations, and the new transcendental functions are in direct correspondence to the new differential forms appearing in their Gauss-Manin connection. We argue that this construction is mathematically equivalent to the splitting of the period matrix into semi-simple and unipotent parts plus a clean-up step, and demonstrate its use with examples of increasing complexity that require the interplay of multiple geometries.

We analyze the nonabelian surface holonomy on a bipartite hypercubic lattice following a proposal in arXiv:1002.4636 [hep-th]. The bipartite structure of the lattice enables us to introduce spike string configurations. These spikes play a crucial role for the time evolution of the string when the total number of color indices changes.

The anisotropic emission of direct photons from Au+Au collisions at $\sqrt{s_{NN}}$=200 GeV was calculated using a (3+1)-dimensional viscous hydrodynamic model with the impact parameter Glasma initial condition. The transverse momentum spectra of direct photons in different centrality bins (0-20\%, 20-40\%, and 40-60\%) are in good agreement with experimental data measured at RHIC. For the elliptic flow $v_{2}$ and triangular flow $v_{3}$, the agreement is centrality dependent, showing good correspondence for the 20-40\% bin but underprediction for the 0-20\% bin and overprediction for the 40-60\% bin. After carefully accounting for the contribution from prompt photons, the measured $v_{2}$ of direct photons is no longer too large to explain. An overestimation of the prompt photon yield can suppress the $v_2$ and $v_3$ of direct photons to values lower than those observed in the experimental data.

We present a minimal basis for non-derivative baryon-number-violating operators in the Standard Model Effective Field Theory up to mass dimension 11, as well as for the $(ΔB,ΔL) = (2,2)$ and $(2,-2)$ operators at dimension 12. Compared to existing results, our bases generally contain fewer terms and simpler contractions, although we also highlight select cases where a minimal basis is incompatible with simple structures.

We study an extension of the singlet-scalar Higgs portal featuring a dark vector $V_μ$ and a real scalar $φ$. The vector is a dark matter (DM) candidate, while $φ$ is long-lived and decays via higher-dimensional operators. We explore the DM production via freeze-in at low and high reheating temperatures. At colliders, the decay $φ\to Z+V$ yields distinctive long-lived particle (LLP) signatures. We explore the interplay between cosmological constraints and LLP searches at the LHC and FCC-hh, showing that far detectors can probe otherwise inaccessible parameter space and place novel bounds on the reheating temperature.

We use the Chapman-Enskog method to investigate the shear viscosity of the quark-gluon plasma with a focus on its relation to parton cross sections. We use the recently obtained analytical expression for the shear viscosity $η$ of a massless quark-gluon gas at chemical equilibrium with Boltzmann statistics and all $2\leftrightarrow 2$ scatterings with arbitrary cross sections. Here we apply this general expression to cross sections at finite temperature that are based on perturbative-QCD and screened with scaled thermal masses $\sqrtκ\,m_D$ and $\sqrtκ\,m_F$. We find that the Chapman-Enskog results on $η\, g^4/T^3$ versus $m_D/T$ at $κ=1$ are qualitatively similar to but higher than the corresponding leading-order results from the AMY framework. We then find that using $κ=0.4$ allows the Chapman-Enskog results to match well the corresponding AMY results as it includes the effect of using thermal masses (instead of self-energies) to screen the cross sections. In addition, we show that the shear viscosity-to-entropy density ratio $η/s$ is very sensitive to the choice of momentum scale $Q$ in the strong coupling, where the choice of $Q=3T$ leads to $η/s \sim 0.15$ for $N_f=0$ or 3 at the QCD phase transition temperature $T_c$. These results lay the foundation for mapping parton cross sections to given shear viscosity in parton transport models and QCD effective kinetic theory.

The IKKT matrix model, which is proposed as a non-perturbative formulation of superstring theory, has an issue typical of zero-dimensional theory -- ambiguity in the definition of its path integral. To tackle this issue, we revisit the path-integral formulation of perturbative string theory. In this article, we review recent progress in the string world-sheet path-integral formulation, especially in the Minkowski signature. We first derive the Minkowskian path integral of the Nambu-Goto type equivalent to Polyakov's Euclidean path integral for critical closed string theory, showing equivalences among the Nambu-Goto-, Schild- and Polyakov-type formulations both in the Minkowskian and Euclidean signatures. We also show that ``stringy causality'' is realised in the path-integral formulation at the level of string perturbation theory. We then obtain the matrix model with a property like the stringy causality, which turns out to be a Minkowskian version of the NBI-type IKKT matrix model, by matrix regularisation of the path integral for perturbative type IIB string theory.

We analyse the laws of thermodynamics governing the behaviour of cosmological horizons in de Sitter space and their map to a holographic description at future infinity, $\mathcal{I}^+$. In this case, the boundary can receive signals from two cosmological horizons. We find a universal form for the first law of thermodynamics, valid in general circumstances, when matter-energy crosses both horizons and impinges on the boundary. This universal form leads to a well defined notion of entropy in the holographic dual. It is specified on a co-dimension one surface of the boundary, and can be expressed as a function of two boundary charges, pressure and angular momentum, both of which are derived from the Brown-York stress tensor. Additional comments on the second law, confusing factors of $i$ which arise, and comments pertaining to JT gravity, are included towards the end.

We present the first version of the new scanning tool NMSSMScanner that allows to perform efficient scans in the complex multi-parameter space of the Next-to-Minimal Supersymmetric extension of the Standard Model (NMSSM) while taking into account all relevant constraints. As a proof of concept we apply it to the search for NMSSM parameter configurations that maximize Higgs boson pair production from resonant scalar or pseudoscalar production in various final states.

We present a general and systematic framework to quantify uncertainties arising from imperfectly known systematic correlations in data combinations. Formulated at the level of the combined data, the method enables controlled variation of the correlation structure, leading to the construction of covariance matrices directly on the resulting combination and thus providing a robust and systematic estimate of correlation-induced uncertainties. We apply the method to $e^+e^- \to \mathrm{hadrons}$ cross section data, with the resulting covariance matrices propagated to derived observables, including dispersive determinations of the hadronic vacuum polarization (HVP) contribution to the muon anomalous magnetic moment, $a_μ^\mathrm{HVP}$. We find that uncertainties from systematic correlation assumptions are generally subdominant but non-negligible, and do not fully account for differences between existing $e^+e^- \to \mathrm{hadrons}$ data combinations. The framework is broadly applicable to correlated data combinations in precision measurements and constitutes a new component of the upcoming KNTW data combination for $a_μ^\mathrm{HVP}$.

This paper presents the first neutrino cross-section measurement from the ICARUS detector at Fermilab, using NuMI (Neutrinos at the Main Injector) beam data collected from two beam operation periods corresponding to $2.5\times10^{20}$ protons-on-target in neutrino beam mode. The signal is defined by events with no pions produced in the final state, a topology dominated by charged-current quasi-elastic-like (CCQE-like) signatures. The measurement is reported as flux-averaged differential cross sections as functions of kinematic variables that provide sensitivity to the complex nuclear effects which often dominate the systematic uncertainty budgets of neutrino oscillation measurements. Specifically, this work reports cross sections in two angular variables -- the angle of the outgoing lepton and the opening angle between the lepton and leading proton -- and two variables characterizing the kinematic imbalance between the muon and proton in the plane transverse to the incoming neutrino. These results are compared against predictions from a variety of neutrino event generators, with $p$-values calculated between the extracted cross sections and each prediction. Overall, the predictions agree with the data; however, the current budget of uncertainties does not yet provide sufficient discriminating power to favor a specific model.

Recent measurements of the integrated CP asymmetry in $τ\to Kπν_τ$ decays by the BaBar collaboration exhibit a $2.8σ$ deviation from the Standard Model (SM) prediction. In this work, we investigate CP-violating effects in $τ\to Kπν_τ$ within the model of the Left--Right Inverse Seesaw (LRIS) model. We show that, although the integrated asymmetry remains too small to account for the BaBar result, the model nevertheless predicts a pronounced signal in the \emph{differential} forward--backward CP asymmetry, $A_{\rm CP}^{\rm FB}(s)$. We derive the effective $|ΔS| = 1$ Hamiltonian relevant for these decays and identify a dominant non-decoupling scalar operator, $g_S$, generated by a top-quark flavor-changing neutral current box diagram involving heavy neutrinos and scalar exchange. Our numerical analysis demonstrates that, while this contribution largely cancels in the integrated $A_{\rm CP}$, it significantly enhances $A_{\rm CP}^{\rm FB}(s)$ through interference with the SM vector current, leading to distinctive kinematic features near the $K^*(892)$ and $K_0^*(1430)$ resonances. These angular and differential observables provide a sensitive probe of the LRIS scalar sector at current and future flavor experiments, in particular Belle~II.

We present a global analysis of the current $B$-meson flavor anomalies and extend it to the baryonic sector through the decay $Λ_b^0 \to Λ_c^+(\to Λ^0 π^+) τ^-(\to π^- ν_τ)\barν_τ$. The lepton flavor universality ratios $R_{τ/(μ,e)}(D^{(*)})$, measured by BaBar, Belle, and LHCb, exhibit a combined $3.8σ$ deviation from Standard Model (SM) predictions. Using the latest HFLAV averages and $B_c$-lifetime constraints, $\mathcal{B}(B_c \to τν) < 60\%, 30\%, 10\%$, found new physics (NP) solutions to the cascade decay $Λ_b^0 \to Λ_c^+(\to Λ^0 π^+) τ^-(\to π^- ν_τ)\barν_τ$. The scenario $(C_{V_L},C_{S_R})$ emerges as the most favored NP solution, largest pull from the SM and insensitive to branching-ratio constraints; followed by $C_{V_L}$ case. We study the impact of NP operators on a complete set of angular observables on the five-fold $Λ_b$ decay using Lattice-QCD form factors and find that the scenarios $(\Re[C_{S_L}=4C_T],\Im[C_{S_L}=4C_T])$ and $(C_{S_L},C_{S_R})$ generate the largest deviations from the SM predictions. In particular, the observables $\mathcal{K}_{1c}$, $\mathcal{K}_{2ss}$, $\mathcal{K}_{2cc}$, and $\mathcal{K}_{4s}$ show the highest sensitivity to NP effects. The correlation analysis reveals the $(\Re[C_{S_L}=4C_T],\Im[C_{S_L}=4C_T])$ scenario exhibits inverse correlations among $\mathcal{K}_{1c}$ and $\mathcal{K}_{2ss,2cc,4s}$ and direct correlations between $\mathcal{K}_{2ss}$ and $\mathcal{K}_{2cc,4s}$, pointing to a possible CP-violating phase, while the $(C_{S_L},C_{S_R})$ scenario displays complementary behavior consistent with CP-conserving dynamics. These results establish baryonic semileptonic decays as a powerful and independent probe of the $R_{τ/(μ,e)}(D^{(*)})$ anomalies, with future measurements providing critical tests of the underlying NP structure.

Self-dual theories are powerful toy models of their completions. It was shown recently that there are infinitely many SD-theories once massless higher-spin fields are allowed. The maximal SD-theory is chiral higher-spin gravity. Following the recent [arxiv:2602.12176] we show that all SD-theories, including those with massless higher-spin fields, have nontrivial tree-level amplitudes in Kleinian signature or complex Minkowski kinematics. Within celestial holography, the nontriviality of amplitudes in chiral higher-spin gravity provides the missing ingredient needed to complete the celestial analogue of the vector-model/higher-spin AdS/CFT duality.

The $(1+1)$-dimensional chiral anomaly is a paradigmatic exact result in quantum field theory, traditionally formulated for zero-temperature pure states where it arises from spectral flow induced by external gauge fields and captures universal ground-state properties. In mixed states, however, the participation of many states and charge exchange with the environment invalidate this mechanism. Naive extensions yield model-dependent anomaly coefficients, calling its universality into question. Here, we resolve this problem for Abelian symmetries by deriving the anomaly from an algebraic relation between the symmetry and its flux-insertion operator. We obtain symmetry-charge flow, a mixed-state generalization of spectral flow, in which an applied field redistributes statistical weight across symmetry-resolved charge sectors. Fixed solely by symmetry, the anomaly restores universality and applies to both pure and mixed states in fermionic and bosonic systems. We substantiate these results in tight-binding fermionic models with continuous symmetry and in spin models with discrete symmetries.

We propose a general framework for integrable field theories in arbitrary spacetime dimension $d+1$ which is based on $d$-term $L_\infty$-algebras. Specifically, we introduce cyclic $L_\infty$-algebras describing topological-holomorphic higher Chern-Simons theories on $M \times \mathbb{C}P^1$ with suitable singularity structures and boundary conditions, controlled by a meromorphic $1$-form on $\mathbb{C}P^1$. Using homological perturbation theory and homotopy transfer, we construct weakly equivalent models describing $(d+1)$-dimensional field theories on $M$. Their integrability is witnessed by a natural map to an $L_\infty$-algebra describing higher Lax connections, yielding conserved charges associated with higher-dimensional cycles in $M$. The resulting theories admit natural action functionals and recover the Costello-Yamazaki construction in $2$ dimensions.

We present a solution to the strong CP problem in which a simple chiral U(1) gauge symmetry gives rise to an accidental Peccei-Quinn symmetry that is both explicitly and spontaneously broken by mirror QCD dynamics, yielding a framework without massless fermions or a light QCD axion. The model contains no stable domain walls or colored relics, and it accommodates a sufficiently high reheating temperature to account for the baryon asymmetry of the Universe via leptogenesis. Metastable domain walls and first-order mirror QCD phase transition generate a stochastic background of primordial gravitational waves. Additionally, one of the pseudo-Nambu-Goldstone bosons serves as a viable WIMP dark matter candidate. The gauge boson associated with the chiral U(1) gauge symmetry, which kinetically mixes with the Standard Model hypercharge gauge boson, provides a vector portal connecting dark matter to the Standard Model and plays a central role in the dark-matter phenomenology. Colored pseudo-Nambu-Goldstone bosons can be probed at the LHC through searches for dijet resonances, jets plus missing energy, multijet events with leptons, and displaced vertices.

We study three-gluon scattering amplitudes in global de Sitter spacetime, in the angular momentum basis of SO(1,4) symmetry representations. At the tree level, they are determined by the intertwiner integrals of harmonic one-forms on the three-sphere. We derive a general formula valid for all helicity configurations of incoming and outgoing gluons and express the amplitudes in terms of Wigner 3j symbols.

Motivated by the Swampland Distance Conjecture, we study distances in field space using the framework of Optimal Transport. The associated optimisation problem naturally leads to a notion of distance in terms of a (generalised) Wasserstein distance between probability distributions over field space. In the absence of dynamical gravity, we relate the transport problem to Hamilton-Jacobi and continuity equations arising from a WKB expansion of a Schrödinger equation associated with the physical configuration. We then formulate an extension in the presence of dynamical gravity. Using the ADM formalism, we establish the corresponding transport problem through the Wheeler-DeWitt equation, giving rise to different possible choices of cost functions. The resulting notions of distances are naturally defined on the full configuration space, while an interpretation in terms of a genuine scalar field distance requires additional modifications. We further discuss several applications and examples, and indicate possible implications for different themes within the Swampland program.

The three-dimensional cubic conformal field theory governs the critical behaviour of Heisenberg magnets with cubic anisotropy. Studying this theory non-perturbatively is challenging, because its most easily accessible observables are numerically very close to those of the more symmetric $O(3)$ model. In this work, we overcome this difficulty using the fuzzy sphere regularisation method. By adding a cubic-invariant two-body interaction to the quantum rotor Hamiltonian used for the $O(3)$ model, we break the continuous rotational symmetry by construction and unambiguously isolate the cubic critical point. Using exact diagonalisation and the density matrix renormalisation group, we calculate the scaling dimensions of several key operators, including the leading scalar singlets, and resolve the splitting of the $O(3)$ rank-two traceless symmetric tensor into the $E_g$ and $T_{2g}$ representations of the cubic group. Our results are consistent with existing Monte Carlo, conformal perturbation theory, and $\varepsilon$ expansion benchmarks, demonstrating the power of the fuzzy sphere in resolving closely spaced universality classes.

The total decay widths of heavy mesons can be systematically calculated in terms of an expansion in the two parameters $1/m_Q$ and $α_s(m_Q)$, where $Q=c,b$ denotes the heavy quark. The dominant contributions to meson lifetime splittings stem from terms which are suppressed by $1/m_Q^3$ with respect to the leading universal contribution to the total decay width. We calculate three-loop contributions of order $α_s^2/m_q^3$ to the lifetime ratios $τ(B^+)/τ(B_d^0)$, $τ(D^+)/τ(D^0)$, and $τ(D_s^+)/τ(D^0)$ in the limit of exact isospin and V-spin symmetry, respectively. Furthermore, we present new $α_s/m_q^3$ corrections to the Cabibbo-suppressed terms in $τ(B^+)/τ(B_d^0)$. Combining our perturbative coefficients with hadronic matrix elements calculated from Heavy Quark Effective Theory sum rules, we find $τ(B^+)/τ(B_d^0)= {1.072 \pm 0.024 }$. Using hadronic matrix elements from a recent lattice QCD calculation we find $τ(D^+)/τ(D^0)={2.344 \pm 0.170} $ and $τ(D_s^+)/τ(D^0)={1.289 \pm 0.042}$. We find good agreement of our predictions with experimental data, which constitutes a successful probe of the calculations of hadronic matrix elements and permits estimates of the unknown $1/m_Q^4$ contributions as well as the V-spin breaking terms in $τ(D_s^+)/τ(D^0)$.

Despite the archetypal status of the BTZ background in quantifying quantum aspects of black holes, several features at low temperatures remain imprecise and incomplete. Here, we systematically investigate the behaviour of the Euclidean path integral at low temperatures in the context of AdS3 supergravity, including an analysis of quantum fluctuations in both the near-horizon and asymptotic regions. We clarify and rectify aspects of the bosonic fluctuations, highlighting the role of boundary conditions in AdS3, and show in particular that the gravitational path integral in the near-horizon region is inequivalent to that around BTZ at low temperature. We further account in detail for the contributions of Chern-Simons fields and spin-3/2 modes, thereby refining the disparities between the near-extremal and near-BPS limits at low temperature. Altogether, our analysis sharpens the distinction between near- and far-region dynamics and demonstrates a disagreement in the gravitational path integral at the quantum level.

The method of extremal flows has presented an alluring alternative approach to numerically solving bootstrap constraints. Here I present the development and adaptation of that approach to a more general class of flows with apparent discontinuities. I focus on upgrading solutions of gap maximization for the spinning modular bootstrap from low to high numerical order, though the methodology is generic to a broader class of bootstrap constraints and flows. This methodology presents various nontrivialities and nuances which reflect a richness of the space of bootstrap solutions. The result is a prototype which successfully upgrades solutions in a simple test case at small scale.

This work investigates the role of the $U(N) \times U(\tilde{N})$ global symmetry in tree-level scattering amplitudes of the bi-adjoint $φ^3$ theory from three perspectives: combinatorics, correlation functions, and a massive extension of the CHY formalism. We derive a planar scattering potential whose extrema reproduce Dolan and Goddard's massive scattering equations, providing physical intuition of the construction. This potential enables the counting of kinematic invariants via maximally symmetric Ferrers shapes, and it is expressed in terms of conformally invariant cross-ratios. We find that the $U(1)$ decoupling identity provides a physical interpretation of two different Catalan recursion relations, and also reveals an interplay between Catalan and Narayana numbers in the $U(1)$ splitting. Finally, we construct correlation functions for a fixed particle ordering using the CHY formalism, offering new insights into the dynamics of such amplitude structures. We derive a closed form expression of the reduced number of solutions for this set-up, as well as an off-shell scattering potential.

In ultra-relativistic heavy ion collisions at the LHC, each nucleus acts as a source of high-energy quasi-real photons that can participate in scattering processes without causing either participating nucleus to break up and emit forward neutrons. This paper extends recent measurements of $γ+A\rightarrow\mathrm{jets}$ production in ultra-peripheral Pb+Pb collisions at $\sqrt{s_\mathrm{NN}} = 5.02$ TeV with forward neutron emission on exactly one side of the event. The data presented here was recorded by the ATLAS collaboration at the LHC in 2018, corresponding to a luminosity of $1.72$ nb$^{-1}$. These results examines $5.02$ TeV Pb+Pb collisions where neither nucleus breaks up ($0n0n$), providing a mixture of photon--pomeron ($γ+I\!\!P\rightarrow\mathrm{jets}$), photon--photon ($γ+γ\rightarrow\mathrm{jets}$), and peripheral photonuclear ($γ+A\rightarrow\mathrm{jets}$) events. The different processes are statistically separated via a template fit of the minimum rapidity gap distribution. The kinematics of the hard processes are determined from $R = 0.4$ jets reconstructed using the anti-$k_t$ algorithm. The statistical separation of the different processes then allows for the first measurement of $γ+I\!\!P\rightarrow\mathrm{jets}$ cross-sections in nuclear collisions at the LHC. The rate for electromagnetic dissociation of $0n0n$ $γ+A\rightarrow\mathrm{jets}$ events is also measured and compared to the analogous result from collisions with single-sided neutron emission. These comparisons support the hypothesis that $γ+A\rightarrow\mathrm{jets}$ events without forward neutron emission select a more peripheral class of $γ+A$ collisions.

We report on an experimental demonstration of efficient neutron generation based on direct laser acceleration in microwire-array targets irradiated by ultrashort (tens of femtoseconds) laser pulses. The optimal array period was identified, at which the maximum proton energy and the number of protons with energies exceeding $1~\mathrm{MeV}$ were significantly increased. Using a $1~\mathrm{PW}$, $\sim25~\mathrm{fs}$ laser at a moderate intensity of $\sim10^{20}~\mathrm{W/cm^2}$, a high neutron yield of up to $(8.33\pm0.84)\times10^{6}~\mathrm{n/sr/J}$ was detected from the LiD converter via $^7\mathrm{Li}(p,n)$ and $\mathrm{D}(p,n+p)$ nuclear reactions. Self-consistent integrated simulations reproduced the experimental results and predicted that with a Be converter, a forward pulsed neutron source with an unprecedented yield per joule of $3.67\times10^{7}~\mathrm{n/sr/J}$ can be obtained under identical laser conditions. This type of neutron source is favorable for applications that require a high repetition rate utilizing compact and economical laser systems.

We construct and systematically assess four outer-crust equations of state based on relativistic nuclear mass models and a machine-learning mass table. Our aim is to quantify the sensitivity of the equilibrium composition and thermodynamic properties of the outer crust to the underlying nuclear input, and to evaluate how these differences propagate to neutron-star configurations that are particularly sensitive to crustal properties. Equilibrium sequences of nuclei were determined by minimizing the Gibbs free energy per baryon for cold, catalyzed matter in $β$-equilibrium. The resulting outer-crust equations of state were then employed in neutron-star structure calculations near the minimum-mass limit, where global observables are especially sensitive to the low-density equation of state. The four nuclear models predict different equilibrium sequences, last bound nuclei, and neutron-drip properties. These differences are confined to the deepest layers of the outer crust, beyond current experimental mass coverage. Nevertheless, they propagate only weakly to crust-dominated neutron-star configurations: the gravitational mass, radius, crustal thickness, and fractional moment of inertia differ by less than one percent among the models considered. Modern nuclear-mass models provide consistent outer-crust equations of state for neutron-star applications. Although the detailed composition near neutron drip remains model dependent, the corresponding uncertainties have only a minor impact on the global properties of crust-dominated neutron stars. Therefore, these outer-crust equations of state provide a robust low-density description for astrophysical modelling and for future extensions toward unified neutron-star equations of state.

Identifying the physical mechanism driving blazar flares remains a central challenge in high-energy astrophysics. We show that the energy dependence of the standard deviation of the polarization angle variability ($σ_\text{PA}$) provides a powerful and robust discriminator of blazar flaring mechanisms. Using particle-in-cell-integrated polarized radiative transfer simulations, we perform to-date the most rigorous statistical analyses of polarization variability. We demonstrate that magnetic reconnection and magnetized turbulence imprint qualitatively distinct energy dependence of $σ_\text{PA}$ that directly reflect their different magnetic field evolution and particle transport. Reconnection predicts higher $σ_\text{PA}$ with higher photon energy till the synchrotron spectral peak, whereas turbulence produces nearly flat $σ_\text{PA}$ across the synchrotron spectral component. These trends are resilient to realistic observational limitations. Applying our results to optical and IXPE data of Mrk~421 and 1ES~1959+650, we find strong evidence for reconnection-driven flares embedded in a turbulent blazar zone. Energy-dependent $σ_\text{PA}$ emerges as a decisive new probe of particle acceleration in relativistic jets.

The Dark Energy Spectroscopic Instrument (DESI) provides an unprecedented opportunity to test deviations from general relativity (GR) that introduce a new physical scale within its redshift range. Using the connection between a Yukawa-like potential and the Hu-Sawicki $f(R)$ model, we place strong constraints on the range of a hypothetical fifth force mediated by a massive scalar field. We analyze the power spectrum measurements from DESI Data Release 1 using a baseline EFT model that employs the fkpt approach for the loop integrals. We find no evidence for deviations from GR and obtain the constraint $\log_{10} |f_{R_0}| < -4.59$ (95\% C.L.). This corresponds to an upper bound at redshift zero on the scale at which corrections to GR become important, $λ< 17.81$ Mpc, or equivalently, a lower bound on the mass of the additional gravitational mediator of $m_φ> 3.60 \times 10^{-31}$ eV. We find that the modified gravity parameter $f_{R_0}$ is largely orthogonal to the cosmological parameters in the model, such that no additional projection effects relative to the GR case are introduced in this Full-Shape analysis. Furthermore, a second modified gravity parameter, the power index $n$, which modulates the time-variation of the associated mass, is found to be consistent with previous analyses that fixed it to unity. Adding DESI BAO data or other cosmological probes does not significantly change these results. The conclusions remain similar if the background evolution is described by evolving dark energy instead of a cosmological constant. Additionally, we test the robustness of the baseline model by varying the maximum wavenumber used in the Full-Shape analysis and analyzing the DESI targets separately. Finally, we analyze the degeneracies between the modified-gravity parameters and the sum of neutrino masses.

The 21cm emission line from neutral hydrogen (HI) contained within galaxies provides a way to make accurate spectroscopic redshift determinations in the radio part of the spectrum. Large radio arrays such as SKA-MID are coming online that will have the sensitivity and survey time required to catalogue hundreds of thousands to millions of HI galaxies, opening up the possibility of studying the cosmological large scale structure using this technique. The expected number counts and clustering properties of the galaxies are still quite poorly understood however. We use three different simulated galaxy catalogues to predict the properties of the HI galaxy distribution that SKA-MID will be able to observe, along with estimates of the error on these predictions due to modelling uncertainty. The simulations in question are from S$^3$-SAX (semi-analytic models based on the Millennium dark matter-only simulation); GAEA (an updated semi-analytic model partially calibrated on hydrodynamical simulations); and IllustrisTNG (a hydrodynamical simulation). We present predictions for galaxy number counts as a function of sensitivity cut and redshift, and use these to forecast the cosmological performance of a proposed SKA-MID cosmological survey. Finally, we fit a halo occupation distribution model to low-redshift angular correlation functions to constrain clustering properties of multiple sub-volumes of the simulations to gain insight into the expected variation (sample variance) over smaller survey areas.

Observing the Epoch of Reionisation using 21cm radio interferometry has proven to be a challenging task. Extraction of the extremely faint redshifted signal is complicated by the presence of bright foregrounds, radio frequency interference (RFI), and systematic artefacts. We discuss the challenge of accounting for systematic effects, particularly cable reflections, that appear in the visibility data obtained from 21cm interferometers. Cable reflections cause attenuated copies of the foreground signal to appear outside the 'foreground wedge' region in which foreground contamination is supposed to be localised. We build on the hydra-pspec Gibbs sampler to implement a model of the systematics as a multiplicative effect in delay-fringe rate space. We include this model in the inference of the joint posterior distribution, in addition to the 21cm signal, its power spectrum, and foregrounds. This allows the systematics contribution to be marginalised, rather than filtering it out and causing additional signal loss. We demonstrate the method on simulated visibility data for a single baseline, showing that the 21cm delay power spectrum can be recovered well regardless of the location of the systematics in delay-fringe rate space. Our implementation is suitable for modelling other multiplicative factors on the visibilities, e.g. residual gain errors.

This is a brief overview of the connection between neutron skin thickness in finite nuclei and the equation of state of neutron-rich matter, with applications to neutron stars. Multiple experimental probes are discussed, including dipole polarizability, parity-violating electron scattering, heavy-ion fragmentation, quasi-free scattering, and ultraperipheral collisions. A consistent picture emerges from Bayesian analyses combining experimental data and energy density functionals, providing constraints on the symmetry energy and its slope.

In this paper, we introduce MujicΛ (Mapping the Universe with Jax-based Initial Condition ReconstrΛction), an optimization-based framework for reconstructing initial conditions from realistic galaxy spectroscopic redshift surveys. Unlike standard optimization-based approaches, MujicΛ augments the L-BFGS algorithm with a projection operator and rank-order matching to enforce Gaussianity of the initial conditions and substantially improve robustness to incomplete survey geometries. We validate MujicΛ on a mock lightcone catalog derived from semi-analytic models applied to the Millennium simulation. We construct a differentiable forward model that incorporates a fast particle-mesh simulation at megaparsec resolution and a comprehensive treatment of observational effects and survey incompleteness. MujicΛ reaches good agreement with the true density field down to the scale of the forward model, while maintaining consistency with the Gaussian prior through the projection step. It also broadly recovers the cosmic web classification, underscoring its value for deciphering environmental information in galaxy evolution studies. Beyond its key role in next-generation constrained simulations, the methodology offers a practical way to generate initial guesses and speed up field-level inference, especially for upcoming large-scale galaxy surveys.

The Italian quantum eye (IQUEYE) is a fast photon counter based on the single photon avalanche diode detectors and capable of preserving a ~0.5 ns/h accuracy photon time of arrival. IQUEYE was originally developed for intensity interferometry experiments, but now its scientific scope has been extended towards ultra fast astronomy, including optical pulsars, millisecond pulsars and the enigmatic fast radio bursts. IQUEYE's capabilities are mainly restricted by the number of photons detected, a quantity that scales with the collector size of an optical telescope. Through the visitor instrument program at Gemini South (Cerro Pachón, Chile) we brought IQUEYE to the 8.1-m dish, reaching an order magnitude sensitivity increased from previous operations. At Gemini South we installed IQUEYE to observe giant pulse emitters, millisecond pulsars, and transitional millisecond pulsars for over 40 hours in the span of a week.

Recent JWST observations have revealed that some galaxies at $z \gtrsim 7$ generally exhibit relatively flat ultraviolet (UV) attenuation curves and a weak UV bump. These features suggest that the first dust grains formed rapidly, possibly originating from core-collapse supernovae (SNe). We investigate the time evolution of grain size distributions and extinction curves in the early phase of dust enrichment for different parameters of progenitor stars, rotation velocities, metallicity, and interstellar medium densities, including the effect of the reverse shock. We model a single starburst system assuming an initial mass function. Extinction curves are calculated from the grain size distribution for each dust species. The total dust-to-stellar mass ratio at $30 \,\mathrm{Myr}$ is $M_\mathrm{dust}/M_\star \sim 10^{-3}$ before the passage of the reverse shock, but we find it to be at most $M_\mathrm{dust}/M_\star \sim 10^{-5}$ due to the destruction effect of the reverse shock. This effect destroys grains smaller than $\sim 10\,\mathrm{nm}$ and makes amorphous carbon the dominant species, resulting in a flatter extinction curve with a wide bump at $2500\,\mathrm{\mathring{A}}$ compared to the no-reverse shock models. We find that our models are consistent with the observed attenuation curve and emissivity of high-redshift galaxies and show that the reverse shock processing significantly affects dust enrichment and grain properties such as extinction curves and emissivity in supernova yields for high-redshift galaxies.

We report a detailed analysis of GAMA 376183, a powerful, heavily obscured active galactic nucleus (AGN) hosted by a low-mass galaxy ($M_\star \approx 10^{10}~M_{\odot}$) likely experiencing a galaxy merger. The source was initially identified due to its remarkably strong [Ne v] $\lambda3426$ emission, exhibiting a rest-frame equivalent width (EW) of $\approx 48$ A. We present $\sim100$ ks Nuclear Spectroscopic Telescope Array follow-up observations, confirming its heavily obscured nature with a column density (in $\mathrm{cm^{-2}}$) of $\log N_\mathrm{H} = 23.3^{+0.4}_{-1.2}$ and an intrinsic $2$--$10$ keV luminosity (in $\mathrm{erg~s^{-1}}$) of $\log L_\mathrm{X,int} = 42.92^{+0.24}_{-0.20}$. GAMA 376183 thus represents one of the few known heavily obscured AGNs in low-mass galaxies. Its estimated Eddington ratio is $λ_\mathrm{Edd}\approx0.8$, indicative of rapid black-hole growth. High-resolution optical images reveal a disturbed, likely merging morphology, while its multiwavelength spectral energy distribution indicates a recent starburst in its host galaxy. These pieces of evidence suggest that the ongoing merger has triggered both the heavily obscured, Eddington-limited accretion and the starburst, making GAMA 376183 a rare observed case in low-mass galaxies. Overall, this unique source demonstrates that (i) [Ne v] can help identify heavily obscured low-mass AGNs, and (ii) the merger-driven coevolution framework established for massive galaxies may also extend to low-mass galaxies.

We investigate the dynamical origin and kinematic signatures of spiral structure in an N-body simulation of an isolated barred galaxy, assessing whether invariant manifold theory provides a consistent dynamical framework to disentangle the disc particle populations and to identify those that genuinely build, trace, and sustain the spiral arms. We compute the Jacobi energy of disc particles and classify them relative to the energies of the equilibrium points, thereby isolating manifold-compatible orbits. We analyse their spatial distribution and velocity structure to characterise spiral-related streaming motions. The Jacobi constant provides a physically motivated dynamical separator that reveals three distinct kinematic populations: (i) low-energy particles on nearly circular orbits populating most of the disc, (ii) high-energy particles associated with banana orbits, and (iii) manifold-compatible particles originating near the bar and following transit orbits along the spiral arms. Only the manifold-compatible population generates the prominent outward-migrating ridge observed in the R - v_phi plane and reproduces the characteristic spiral streaming pattern. In contrast, the low-energy population exhibits a global quasi-circular motion with small perturbations induced by the self-gravity of the spiral structure. Our results demonstrate that the spiral arms are dynamically traced by the manifold-compatible population, which forms the backbone of the structure and drives effective radial transport. The bulk of low-energy disc particles responds to the spiral perturbation similarly to the traditional density wave picture, enhancing the density contrast caused by the invariant-manifold compatible particles. In this framework, barred spiral arms emerge as material structures sustained by manifold-guided transport, with the surrounding disc behaving as a system of material density waves.

The radial profile of binary fraction may vary with environment and is of significant importance for studying the formation mechanisms of binary stars and their dynamical evolution within globular clusters (GCs) and galaxies. However, existing studies remain limited to the Milky Way and its neighboring galaxies. Leveraging the method proposed by Zhang et al. for estimating the variation of binary fraction from integrated spectral features, we analyze a sample of 513 elliptical galaxies drawn from the Mapping Nearby Galaxies at Apache Point Observatory (MaNGA) survey to measure their radial binary fraction profiles. Our results show that after accounting for the effect induced by radial variations in the stellar population (SP), the median SP-subtracted binary fraction, $r_{\rm b,sub}^{\rm med}$, becomes approximately flat. For nearly all elliptical galaxies in our sample, the variation in binary fraction relative to the galaxy center at $1R_e$ is less than 5%. No clear correlation is found between the binary fraction gradient and the gradients of SP properties. Moreover, we also compare differences between ultraviolet (UV) upturn and non-UV upturn galaxies. The overall binary fraction profiles and SP properties of the non-UV upturn galaxies in our sample are comparable to those of the UV upturn galaxies. This similarity may arise from the presence of residual star formation (RSF) in the non-UV upturn systems.

Be X-ray binaries (BeXRBs) are key laboratories to constrain binary interaction processes such as mass transfer, angular-momentum transport, and natal kicks. The Small Magellanic Cloud (SMC), hosting a nearly complete and well-characterized BeXRB population, offers a unique opportunity to test these physical processes at low metallicity. We aim to identify the combination of binary-evolution parameters that simultaneously reproduces the observed number and the joint distribution of orbital period and optical magnitude of SMC BeXRBs. We performed an extensive grid analysis of binary population-synthesis models exploring different mass transfer efficiencies, angular-momentum transport prescriptions and Roche-lobe overflow stability criteria. We also considered the impact of natal kicks, and that of the propeller effect of rotating magnetic fields of neutron stars. Synthetic populations obtained with the binary population synthesis code \textsc{sevn} are statistically compared to observations using likelihood-based methods applied to the orbital period and $V$-band magnitude distributions, together with requirements on the total number of systems. We find that models in which mass transfer via Roche-lobe overflow is assumed to be always stable and angular momentum is recycled back into the orbit through tides when the accretor approaches critical rotation provide the best match to observations. Our best-fitting models favor low natal kicks ($\lesssim 100\ \rm km\ s^{-1}$), a moderate mass transfer efficiency ($f_{\rm MT} \simeq 0.6$), a minimum Be threshold spin close to critical rotation, and a strong suppression of accretion onto neutron stars due to the propeller effect. Specifically, the observable population is highly sensitive to the treatment of the propeller effect, which regulates the X-ray luminosity of wide, low-accretion-rate systems.

The Euclid Wide Survey (EWS) will cover the majority of the extragalactic sky with a resolution similar to the Hubble Space Telescope. This unprecedented data set will introduce a new era of precision cosmology. However, systematic effects need to be controlled better than ever. One of the sources of systematic uncertainties in weak gravitational lensing are biases introduced during the shear measurement. Determining these biases precisely allows the calibration of cosmological measurements to within Euclid's required accuracy. The simulations that are used to determine such biases, need to resemble the real observations. In this work, we aim to learn distributions of galaxy shape parameters from real Euclid data and use the new information to augment the morphological information in the Flagship galaxy mock catalogue. The morphology is extracted using single and double-Sérsic model fits to the real data, for which we use SourceXtractor++. We train our pipeline on deep Euclid observations of a field with rich auxiliary data and then use it to simulate EWS-like data. In these simulations we compare the multiplicative bias between the morphology from the Flagship catalogue, the trained single-Sérsic morphology, and the trained double-Sérsic morphology. We find that the image simulations with the updated morphology result in a percent-level change in the multiplicative shear bias compared to the original morphology from Flagship. This bias exceeds Euclid's tight error budget by a factor of five and underlines the need for this work. Furthermore, we study the sensitivity of the multiplicative bias to key morphological parameters and show that our approach satisfies the requirements for the cosmology analysis with the first data release of Euclid.

Context. Growing evidence is revealing the crucial role of binarity in massive star evolution. This affects evolution models and demands a refinement of the available observational constraints. Aims. To investigate the possible evolutionary origins of a sample of 117 Galactic O-type stars with luminosity classes V to III and projected rotational velocities below 150 ${\rm km}\,{\rm s}^{-1}$. Methods: We extend previous quantitative spectroscopic analyses performed within the framework of the IACOB project and obtain N abundance estimates. We investigate correlations between these abundances and other stellar parameters. As a reference, we use predictions from single-star evolution models computed using different physical prescriptions. Results. We identify differences in the N abundance distributions corresponding to three He abundance regimes (He-low, He-normal, and He-rich). For the He-normal group, the N abundance distribution peaks slightly above the expected birth value and extends up to $ε_{\rm N}$=8.4 dex. For these stars, we find overall agreement with single-star evolutionary models that include efficient internal mixing and assume moderate-to-low initial rotation. In contrast, the He-rich group exhibits a bimodal N abundance distribution, with one peak at $\sim$8.1 dex and a second more enriched peak around $\sim$8.5 dex; none of these stars are consistent with predictions from single-star evolutionary models. We argue that He-rich stars are most plausibly explained as binary products. Furthermore, despite N abundance in He-normal stars with LC IV and V are reproduced by single stellar evolutionary models with efficient mixing, the same models predict a higher N abundance than observed for stars with LC III. This indicates that rotational mixing alone is unable to explain the observed distribution of N abundances among stars with normal He abundances.

Differential abundances of C, O, Mg, Al, Ca, Sc, Ti, V, Cr, Mn, Fe, Co, Ni, Zn, Y, and Zr were determined from high signal-to-noise VLT/UVES spectra for 25 turnoff stars with -2.4 < [Fe/H] < -1.3. Effective temperatures were obtained from profiles of the H_beta line and surface gravities via Gaia parallaxes. The analysis of the spectra were based on 1D model atmospheres assuming LTE, but 3D non-LTE corrections were applied for several elements. The dispersion in linear fits to the [X/Fe]-[Fe/H] relations is around a factor of two smaller than found in previous studies. After corrections for measurement errors, the 1-sigma intrinsic dispersion of [X/Fe] at a given metallicity is 0.09 dex for Y and Zr, 0.05-0.07 dex for C, O, and Al, 0.03-0.05 dex for Mg, Ca, Sc, Ti, V, Mn, and Zn, and <0.03 dex for Cr, Co, and Ni. Strong correlations between the residuals in the [X/Fe]-[Fe/H] fits are found for the alpha-capture elements Mg, Al, Ca, Sc, and Ti and between the residuals for Y and Zr. Correlations of the residuals in the [X/Fe]-[Fe/H] fits with effective temperature can be explained as due to differential atomic diffusion between elements, but its contribution to the scatter of [X/Fe] is of minor importance. Probably, both stochastic effects in sampling the IMF of CCSNe and differences in the TypeIa to CCSNe enrichment ratio between star-forming regions need to be considered in order to explain the intrinsic dispersion of [X/Fe].

Fast radio bursts (FRBs) are enigmatic millisecond-duration radio transients whose physical origins remain debated. To shed light on this, we analyze the CHIME/FRB Catalog 2. By using the probability distribution of dispersion measured (DM) derived from the IllustrisTGN simulation, we compute the pseudo-redshift with $1σ$ error for each FRB. To derive the FRB luminosity function and event rate, we employ a non-parametric statistical method. Building upon Efron-Petrosian method, we find strong luminosity evolution with redshift, well described by $L_0 \propto (1+z)^{6.38}$. After de-evolving this trend, we apply Lynden-Bell's $C^-$ method to derive the comoving FRB formation rate which is found to decline rapidly at high redshift, following $ρ(z) \propto (1+z)^{-5.38 \pm 0.02}$. We also test the robustness of our results by considering the upper and lower limits of pseudo-redshifts, and different flux limits of CHIME. Similar results are found. This steep decline is inconsistent with a direct tracing of the cosmic star formation rate, but closely resembles the redshift evolution of short gamma-ray bursts-systems linked to compact object mergers. Our results support that the origin of FRBs is associated with old populations, such as neutron stars and black holes.

Large spectroscopic surveys rely on automated pipelines to deliver homogeneous stellar labels, but a substantial fraction of observations are at low signal-to-noise ratio (S/N), where label estimates become imprecise or are omitted. In APOGEE, these low-S/N spectra visits sample faint and distant populations -- the bulge, outer halo, and satellite systems -- yet still encode recoverable chemical information. We present TwinSpecNet (TSN), a paired-learning framework that exploits APOGEE's multi-visit observing strategy: by training on empirical low-/high-S/N spectral twins of the same stars, TSN learns to suppress stochastic noise while preserving the ASPCAP label scale. TSN employs a Vision Transformer encoder with dual objectives: reconstructing high-S/N flux from low-S/N visits and predicting stellar parameters and abundances with calibrated uncertainties. TSN reduces label scatter relative to visit-level ASPCAP for S/N<60 visits. TSN reproduces the ASPCAP scale with residual scatters of $σ$< 19 K in $T_{\mathrm{eff}}$, $σ\sim$0.06 dex in $\log g$, and $σ\sim$0.03 dex in Fe/H. TSN tightens intra-cluster abundance dispersions, recovers cleaner chemical sequences in inner-disk and bulge and satellite samples, and improves C/N-based age precision for APOKASC giants from 1.70 to 1.59 Gyr. By learning survey-specific noise patterns from repeated observations, TSN demonstrates how empirical paired learning can extend the chemical reach of existing spectroscopic data, providing a template applicable to other multi-visit surveys.

Cataclysmic variables can show rapid increases in optical flux. Intermediate polars (IPs), a subset with strong magnetic fields that disrupt the inner accretion disc, have been thought to possess truncated discs that rarely undergo the disc-instability outbursts seen in dwarf novae. However, the discovery of micronovae and magnetic-gating bursts suggests that such events may occur even without a fully developed disc. Using data from the All-Sky Automated Survey for Supernovae (ASAS-SN), we identify a previously unrecognised population of short-timescale optical outbursts in IPs. Initial energy estimates indicate that at least half of these bursts may be consistent with micronovae, though limited cadence reduces our ability to classify each event with high confidence. These detections should therefore be regarded as evidence of short outbursts in IPs rather than definitive micronova identifications. Our results show that such bursts are more common in IPs than assumed and may include a substantial fraction of micronovae. Simulations reveal that if micronovae occur once per year at regular intervals, up to 30\% of the shortest bursts could be missed over a 10-year observing baseline. Under the same assumptions, and assuming all IPs display micronovae, we would expect 50--70\% of IPs to show these bursts, yet only 14\% of known IPs in our sample do so. This discrepancy suggests that not all IPs undergo micronovae. Overall, this work establishes the first comprehensive set of short burst detections in IPs, providing a foundation for future investigations and a list of candidate micronova systems for follow-up analysis.

PSR J0437-4715 is a gamma-ray millisecond pulsar, which has been detected by Fermi-LAT. For understanding the nature, we analyze the GeV gamma-ray data obtained with Fermi-LAT around the pulsar region. Based on the pulsar timing ephemeris, we derived the gamma-ray pulse profile and defined on-pulse and off-pulse phase intervals. A binned likelihood analysis was performed to investigate the spectral properties of the pulsar across different phase ranges. No spatial extension was detected for off-pulse, with an upper limit radius of 0.12 degree. We further investigate the relationship between gamma-ray luminosity, X-ray luminosity, and bow-shock radius for a sample of pulsars with detected bow-shock PWN. The relationship between gamma-ray luminosity and X-ray luminosity is explored. The conversion efficiency from spin-down power to GeV emission of outer gap model is consistent with a termination shock located close to the pulsar. We discuss the potential nature of off-pulse GeV emission and the connection to bow shock pulsar wind nebulae.

In recent years, the obsessive interest in the observation of TMC-1 has brought a boost in our knowledge of the chemistry of cold dark clouds. The number of molecules detected in this particular cloud has been more than doubled. Two observational programmes, GOTHAM and QUIJOTE, are responsible for this spectacular achievement. Here we provide an overall view of QUIJOTE, which is a line survey carried out in the Q band (31-50 GHz) with the Yebes 40m radiotelescope, summarize the actual observational status of TMC-1, and discuss the chemistry of this remarkable source. We highlight the successes and failures of state-of-the-art chemical models to describe its chemical composition, with a particular emphasis on the origin of PAHs, which is yet far from being understood.

Millisecond magnetars are widely suggested as the central engines powering hydrogen-poor superluminous supernovae (SLSNe). These magnetars primarily lose huge rotational energy through gravitational wave radiation (GWR) and magnetic dipole radiation (MDR), with MDR serving as an energy source for SLSNe. We study the evolution of the magnetar spin, magnetic inclination angle, and the resulting thermal radiative luminosity of the SLSNe, where the impacts of the nucleon-nucleon short-range correlation, the mass and initial spin of the magnetar, and the density-dependent symmetry energy of the dense nuclear matter on the evolution are discussed. The relativistic mean-field theory is employed to calculate the nuclear matter properties, and we particularly concentrate on the time- and space-dependent bulk viscosity which is crucial for the magnetic inclination angle evolution. It is found that the nucleon-nucleon short-range correlation weakens the damping of bulk viscosity of dense matter and therefore inhibits the growth of magnetic inclination angle, and it reduces the MDR (GWR) peak luminosity of a canonical magnetar by several times while it raises the peak thermal radiation luminosity of SLSNe by several times. For magnetars with nonrotating mass obviously lower than the $ 1.4 M_\odot$ with slow initial rotation, the magnetic inclination angle is more likely to evolve towards 0 degrees quickly, and these magnetars are not suitable as the central engine for SLSNe. Within the "family" of FSUGarnet interaction, a stiffer symmetry energy gives a lower threshold of direct Urca process and hence gives a much larger bulk viscosity coefficient, and thus it promotes the growth of the magnetic inclination angle and the GWR for canonical stars but reduces the peak brightness of SLSNe significantly.

We investigate the relation between neutral atomic hydrogen (HI) and dark matter halo mass (HIHM) using observations from the Deep Investigation of Neutral Gas Origins (DINGO) pilot survey 100h data, combined with spectroscopic data from the Galaxy and Mass Assembly (GAMA) survey and photometric data from the Wide Area VISTA Extragalactic Survey (WAVES) photometric catalog. We employ a combination of direct detections and spectral stacking to probe the HI content of halos across a wide mass range ($10^{10.5} \lesssim M_\mathrm{h}/M_\odot \lesssim 10^{14.5}$). By incorporating WAVES photometric members on top of the existing GAMA group catalog, we present a novel approach of extending stacking analyses beyond spectroscopic completeness limits, enabling recovery of satellite HI content otherwise missed. We find that the HIHM relation exhibits a double power-law form, with a turnover near $M_\mathrm{h} \sim 10^{11.2} \text{ M}_\odot$. Central galaxies dominate the halo HI budget below $M_\mathrm{h} \sim 6 \times 10^{12} \text{ M}_\odot$, while satellites dominate at higher halo masses. Including photometric members increases the measured HI content in halos above $10^{13} \text{ M}_\odot$ by a factor of 1.5-3, highlighting the importance of gas-rich satellites in the group and cluster regime. Comparison with previous group-stacking studies shows that low-surface brightness galaxies, and intra-group HI structures contribute only a minor fraction to the total HI mass in group and cluster halos, as the summed galaxy HI masses are consistent with the total halo HI content.

Several Lyman-alpha (Ly$α$) emitters deep into the reionization era exhibit double-peaked Ly$α$ emission profiles, raising the question of how the intergalactic medium can transmit photons blueward of the Ly$α$ resonance at such high redshifts. To investigate this, we compute Ly$α$ transmission along sightlines originating from galaxies in the Cosmic Dawn III simulation and identify cases that closely reproduce the observed double-peaked emission. In these cases, the sightlines intersect highly underdense voids located a few comoving megaparsecs from the source galaxy. These voids allow photons emitted blueward of Ly$α$ to redshift through resonance without scattering while traversing them. The low opacity arises because the neutral hydrogen density scales with the square of the underlying gas density under ionization equilibrium, making sufficiently underdense regions with $\lesssim30~\%$ of cosmic mean density highly transmissive. Such voids naturally occur in the fluctuating cosmic density field, even in the vicinity of galaxies, and can also be associated with transmissive spikes in the Ly$α$ forest. We find that the global probability of observing double-peaked emission is $\sim3\times10^{-3}$ during reionization at an 80\% global ionization fraction, while no cases are found at 60\% ionization. We also find that this probability depends sensitively on the local ionizing background intensity, increasing by $\sim10^{4}$ for a tenfold increase in intensity. These results suggest that the fraction of double-peaked Ly$α$ emission in high-$z$ galaxies can serve as a sensitive probe of the ionizing background during the late stages of cosmic reionization.

Quiescent galaxies in the distant universe exhibit a range of gas content that may indicate a variety of quenching processes are at play. Mapping the distribution and kinematics of the gas can illuminate its origins, but nearly all such observations have been unresolved. We present JWST/NIRSpec IFU observations of MRG-M0138, a gravitationally lensed, massive quiescent galaxy at $z\sim2$ observed as part of the INQUEST-JWST survey. We use Na I D absorption, which we detect in excess of the stellar absorption over most of the galaxy, to trace the kinematics and spatial distribution of the neutral gas in 219 spatial bins. The gas exhibits clear rotation that is kinematically aligned with the stellar disk. Both the gas and dust have a complex spatial structure, including an off-nuclear clump, a dust lane, and patches in the outer disk. The non-equilibrium distribution suggests that the gas was accreted. Analysis of the galaxy's star formation history supports this interpretation by indicating a rejuvenation event 500 Myrs ago. We identify two plausibly associated galaxies and suggest that tidal interactions are a likely source of the accreted gas. Our results indicate that some of the variation in gas content among early quiescent galaxies is not related to differences in gas consumption timescales. The detection of a gas clump at a projected distance of $\sim90$ pc from the known supermassive black hole illustrates a mechanism to fuel the episodic AGN feedback that may maintain quiescence.

We perform numerical simulations of particle acceleration in relativistic, self-driven turbulent magnetic reconnection using the MHD-PIC method. We systematically investigate the dependence of the non-thermal particle spectral exponent on the plasma $β$. We find that particle acceleration proceeds in two stages: an initial, efficient first-order Fermi phase where momentum gains are comparable in parallel and perpendicular directions, followed by a slower drift-dominated phase. The power-law slope of the non-thermal spectrum is established during the Fermi phase, as found in previous studies. Our results demonstrate a systematic steepening of the accelerated particle energy spectrum with increasing $β$. We derive empirical scaling relations: the spectral exponent $α\propto β^{0.5}$ in the relativistic regime, compared to $α\propto β^{0.3}$ in the non-relativistic case. This marked difference is rooted in relativistic physics: the increased inertial mass density ($ρh$) in high-$β$ plasmas acts as an energy sink, reducing the Alfvén velocity and thereby altering the dynamics of magnetic energy release and its partition efficiency. The derived scaling provides a unified physical framework for interpreting the diversity of non-thermal radiation spectra observed in astrophysical sources, including black hole corona X-ray flares, gamma-ray bursts, and active galactic nucleus jets.

The scenario of variations in accreting process around central black hole has been widely accepted as the preferred physical origin of changing-look active galactic nuclei (CLAGN), rather than obscuration effects by moving dust clouds. In this manuscript, after analyzing long-term photometric variability in Type-1.8 AGN SDSS J1346+1736 with apparent broad H$α$ but very weak broad H$β$, robust clues for obscurations can be confirmed with confidence level higher than 10$σ$. Then, based on obscurations related to moving dust clouds, from dark state with E(B-V)$\sim$1.26 to bright state with E(B-V)$\sim$0.55, apparent both broad H$α$ and broad H$β$ can be clearly expected. Meanwhile, the expected line luminosity from reddening corrected continuum luminosity is consistent with the reddening corrected line luminosity, to support the obscuration effects. Furthermore, through symmetric features in the light curves, probability is only 3.2\% in SDSS J1346+1736 to support the scenario of variations of accretion rates traced by CAR process. Therefore, besides the more and more preferred scenario of variations in accreting process, the scenario of moving dust clouds could be more reasonable in some CLAGN with apparent variations in optical colors.

This review explores the physical mechanisms driving the evolution of low- and intermediate-mass binary star systems, with particular emphasis on emerging mechanisms that challenge classical paradigms. We begin by describing the principal formation channels and orbital properties of binary systems. A critical reassessment of the Roche lobe formalism is presented, focusing on systems with eccentric orbits and asynchronous rotation, where deviations from traditional approximations become significant. We then review current theoretical models of mass and angular momentum exchange via Roche-lobe overflow, incorporating results from recent hydrodynamical simulations of wind accretion. The review also reports advances in tidal dissipation theory. Finally, we explore mechanisms capable of sustaining or exciting orbital eccentricity, including perturbations induced by mass transfer and interactions with circumbinary disks. These discussions aim to outline underexplored facets of binary evolution, offering new perspectives for theoretical and observational studies.

Gamma-ray bursts (GRBs) are the most energetic bursts of light in our universe, and rapid progenitor association of these events can lead to targeted and optimized follow-up observations, ultimately providing better insights about the physics involved. In this note, we investigate a semi-supervised machine learning algorithm, that utilizes label propagation, as a classification method. Using a dataset of 2512 GRBs we evaluate the method's ability to assign probabilistic class memberships based on a subset of events with known progenitors. Further analysis is ongoing to improve the method and future progress will be made to refine the classification algorithm and the dataset.

The annihilation of dark matter (DM) particles is expected to produce Standard Model particles, providing a potential indirect signature of DM. The clumpy substructure of DM haloes amplifies the expected annihilation signal, an effect commonly quantified by the subhalo boost factor. Standard semi-analytic models usually treat this boost as a universal function of host-halo mass, neglecting systematic variations induced by the large-scale environment. In this work, we extend this framework by incorporating the influence of the cosmic web on subhalo populations. Using simulation-calibrated, environment-dependent ratios for host-halo concentrations, the subhalo mass function, and internal-structure proxies of subhalos based on the $V_{\max}$--$R_{\max}$ relation, we compute environment-conditioned boost predictions for haloes residing in filaments, walls, and voids. Our main result is the boost factor at fixed host-halo mass, expressed relative to the cosmic-mean prediction, $B(M,\mathrm{env})/B_{\mathrm{CM}}(M)$. We find a clear environmental modulation: in the fiducial distance-dependent model, filament haloes show a mass-dependent transition from a $\sim 15\%$ suppression at the low-mass end to a modest enhancement of $\sim 12\%$ for massive hosts, wall haloes remain intermediate, while void haloes stay suppressed by roughly $30$--$33\%$ across the explored host-mass range. These results should be interpreted as deterministic model predictions obtained by propagating environment-dependent ingredient ratios through two standard semi-analytic boost frameworks. We provide an environment-aware prescription for subhalo boosts, together with modular environmental corrections that may also be useful in indirect-detection forecasts, strong-lensing mass modeling, and related halo-population applications.

Little is currently known about the large-scale environments of Green Pea (GP) and Blueberry (BB) galaxies, which are low-mass, compact systems with extreme specific star-formation rates (sSFR). Their environments are inherently linked to their formation mechanism, and they may serve as crucial local analogues for high-redshift, reionizing galaxies. This paper aims to investigate the clustering properties of GPs and BBs, leveraging large-scale survey data to quantify their spatial distribution relative to the broader galaxy population. We here investigate a sample of these galaxies, consisting of 339 GPs $\rm (0.1 < z \le 0.33)$ and 56 BBs $\rm (0 < z \le 0.1)$, whose clustering properties we analyse relative to an extensive control sample derived from the SDSS MPA-JHU DR8 catalogue, binned by stellar mass and sSFR. We use the number of neighbours within a 5 Mpc radius as a proxy for environmental density, i.e. clustering, and employ a pair-matching and bootstrapping methodology to ensure statistical robustness. We observe that galaxy clustering depends strongly on star-formation activity, with passive galaxies being more clustered than their high star-formation rate counterparts, with GPs and BBs lying at the extreme end of this relation, exhibiting the lowest neighbour counts among all subsamples. The nearest neighbours of BBs also tend to have lower masses than other classes of dwarf galaxies. GPs and BBs predominantly reside in isolated, low-density environments, suggesting that their intense starbursts are unlikely to be triggered by common environmental processes such as mergers or starburst cycles. Their low metallicities and weak clustering instead support scenarios in which recent starbursts are driven by internal processes or pristine gas accretion, reinforcing their role as nearby analogues of young, low-mass galaxies in the early Universe.

High-redshift radio galaxies (HzRGs) are among the most massive galaxies in the Universe and sites of extreme active galactic nuclei (AGN) feedback processes, powering energetic radio jets. They are typically embedded in giant Ly$α$ halos that are known to extend over $100\,\text{kpc}$ into the circumgalactic medium (CGM). In this paper, we target the Ly$α$ halos around four high-redshift radio galaxies in a redshift range of 3.1 < z < 4.5 using high-resolution spectroscopy from the Ultraviolet Echelle Spectograph (UVES) at the VLT, focusing on absorption features in the Ly$α$ emission that trace neutral hydrogen (HI) systems. We compare the UVES data to Multi Unit Spectroscopic Explorer (MUSE) observations of the same targets and find that the higher spectral resolution of UVES ($Δv \approx 12\,\text{km} \text{s}^{-1}$) allows for a more complete identification of absorbers and reveals the splitting of deep absorbers into multiple components. We identify between 6 and 14 absorbers for each target in our sample with column densities of $N_{\text{HI}} = 10^{12}-10^{17}\,\text{cm}^{-2}$. About 70 % of the absorbers can be spatially resolved along the radio jet axis, showing minimal variation in column densities over extents of more than 30 kpc. Our results indicate that a fraction of the absorbers may be physically associated with the host systems. Complementary JWST/NIRSpec observations of two of the targets, 4C+03.24 and TNJ0205, reveal potential outflows in the ionized interstellar medium (ISM). We discuss a kinematic link between the [OIII]-emitting gas and the cool halo gas as traced by Ly$α$, suggesting a common outflow origin.

Baryonic physics is anticipated to be a major source of systematic uncertainty in current and future large-scale cosmological surveys. We investigate how baryonic effects on halo density profiles vary with secondary halo properties at fixed halo mass, using the large-volume MillenniumTNG hydrodynamical simulation and its dark matter-only counterpart. We focus on the impact of halo concentration and large-scale environment on the ratio of density profiles of matched halos in the hydrodynamical and dark matter-only simulations. At redshift $z = 0.0$, we find a strong dependence on halo concentration, especially at lower halo mass ($12.5 < \log(M_h/h^{-1}M_{\odot}) < 13.0$), where more concentrated halos exhibit weaker inner enhancement and stronger intermediate-radius suppression at fixed halo mass, with variations reaching $\sim 15\%$ at small scales and decreasing toward larger scales. This trend weakens and reverses at higher halo mass. In contrast, the secondary dependence on large-scale environment is weaker ($\sim 2\%$) and largely scale-independent, with halos in denser regions exhibiting slightly weaker intermediate suppression. By separating internal profile redistribution from total mass suppression, we show that concentration impacts both components, whereas the environmental dependence is primarily associated with an overall mass shift. These secondary dependencies persist at $ z = 0.5$ and correlate with variations in internal baryonic properties. We examine additional halo properties, including halo spin and velocity dispersion, and find significant secondary dependence. Overall, our results highlight the important role of secondary halo properties in modulating baryonic effects on halo density profiles, with potential implications for future modeling efforts.

The bifurcated age-metallicity relation of globular clusters (GCs) in the Milky Way (MW) shows that GCs are either originated in situ or accreted into the Galaxy from former satellites of the MW. The effects of the Galactic tidal field can leave signatures on the dynamical evolution and structural properties of GCs. We present a homogeneous census of dynamical ages for a sample of 93 GCs in the MW, coupled with the knowledge of their common progenitors from the chemo-dynamical parameters from the Gaia mission, unavailable some years ago. We found that the majority of accreted GCs (61%) is dynamically young. This percentage drops to 38% for in situ GCs. Excluding the enigmatic low-energy (LE) GCs, with ambiguous origin, the fraction of dynamically young ex situ GCs raises to 70%. A two tail Kolmogorov-Smirnov test shows that the distribution of dynamical ages of LE GCs cannot be distinguished from the distribution of in situ bulge and disc GCs. Yet, the LE GCs are firmly located on the satellite branch of the age-metallicity relation. An explanation may be that the progenitor of LE GCs plunged very early into the MW, so that the gravitational field of the MW had time enough to act on the associated GCs. The dynamical ages offer a statistically robust evidence corroborating the scenario of an early accretion proposed by several recent studies on the origin of LE GCs.

We construct an analytical model of Population III star formation that connects the cosmological radiation background to sub-AU protostellar disk fragmentation, a dynamic range inaccessible to any single simulation. Our approach is based on combining separate models of the disparate relevant scales: from the cosmological environment to the host-halo scale, from the halo scale to the scale of the star-forming cloud, and from the cloud scale to the fragmenting, accreting protostellar disk. Individually and collectively, the models agree well with the predictions of state of the art simulations, while remaining computationally inexpensive and physically transparent. As an example of the applicability of the model, we study the effects of varying the Lyman-Werner flux on the Pop. III star formation efficiency. We show that depending on the halo properties and the strength of the dissociating radiation field, the halo-scale Pop. III star formation efficiency varies by more than two orders of magnitude from $\varepsilon_{\rm SFE,H} \approx 10^{-3}$ to $\varepsilon_{\rm SFE, H} \approx 0.5$. The abrupt transitions between hydrogen-deuteride cooling (in low virial temperature mini-halos subjected to low radiation backgrounds), molecular hydrogen cooling (at intermediate temperatures and radiation intensities), and atomic cooling (in higher temperature halos exposed to strong radiation fields) produces sharp features in the halo-scale star formation efficiency as a function of the halo properties. Meanwhile, at the scale of individual star-forming clouds, the star formation efficiency is $\varepsilon_{\rm SFE,c} \gtrsim 0.2$. That is, pristine gas in a halo is converted into unstable clouds at a wide range of efficiencies, and these unstable clouds are efficiently converted into Pop. III stars.

Context. Anomalous microwave emission (AME) is a component of interstellar medium emission peaking at 10-60 GHz. Its polarization is both a CMB foreground and a probe of the alignment physics of very small dust grains. Aims. We quantify when the purely rotational electric-dipole emission from very small interstellar grains (spinning dust/AME) can become measurably polarized, and we quantify related UV/optical/IR polarization diagnostics. Methods. We develop a quantum-mechanical symmetric-top model for an axisymmetric very small grain and express polarized emission and absorption coefficients in terms of irreducible density-matrix moments. Alignment is driven by anisotropic illumination; we solve a simplified two-manifold pumping model and compute (polarized) emission and absorption signatures for different dipole configurations and grain-size distributions. Results. Anisotropic illumination can generate polarized spinning-dust emission; in optimal geometries, polarization fractions near the emission peak reach the percent level, while more modest anisotropies reduce the polarization fraction. The same physics predicts polarized UV/optical/IR absorption that can be appreciable in strongly illuminated environments, whereas IR vibrational emission is predicted to be negligibly polarized. Conclusions. Spinning-dust polarization depends on radiation anisotropy, dipole geometry, and the competition between alignment pumping and depolarization by rotational emission. Joint constraints from AME polarization and UV/optical/IR absorption polarimetry provide a direct test of alignment by anisotropic radiation fields and help bound polarized microwave foregrounds.

Ground-based gravitational wave (GW) observatories have detected approximately 200 binary black hole (BH) mergers. The astrophysical origin of these events are debated, with evidence suggesting that at least a subset originated from dynamic environments characterized by frequent close encounters. Accretion disks in active galactic nuclei (AGNs) are of particular interest, as certain observed features could be more readily produced within such environments. In this paper, we investigate the expected properties of mergers in these environments, and their dependence on various parameters, using one-dimensional $N$-body simulations combined with a comprehensive semi-analytical model. In our fiducial model, the distributions of masses and mass ratios ($q$) are similar to those observed. However, they depend strongly on the lifetime and density of the AGN disk and on the number and accretion efficiency of BHs, with higher masses predicted as these quantities increase. The most massive mergers, such as GW231123, can be produced either by efficient gas accretion or by hierarchical mergers among $\geq 3$ generations of BHs. The observed negative correlation between $q$ and the average effective spin ($χ_{\rm eff}$), along with the positive correlation between $χ_{\rm eff}$ and the chirp mass ($M_{\rm chirp}$), can be explained by a combination of efficient gas accretion, which promotes spin alignment, and hierarchical mergers, which produce high-$|χ_{\rm eff}|$ and low-$q$ binaries. Hierarchical mergers can also explain the negative correlation between $q$ and the dispersion of $χ_{\rm eff}$, as well as the positive correlation between $|χ_{\rm eff}|$ and $M_{\rm chirp}$. We present a comprehensive study on how the expected distribution of each of these quantities depends on model parameters and assumptions, which will aid the interpretation of observed GW population properties.

Magnetized accretion onto spinning black holes can accumulate a large magnetic flux across the event horizon and launch a pair of relativistic jets via the Blandford-Znajek mechanism. In the magnetically saturated (arrested) state, excess magnetic flux is ejected from the black hole in episodic magnetic flux eruptions, which result in a significant yet temporary reduction of jet power. We analyze results of a high-resolution 3D general-relativistic magneto-hydro-dynamic numerical simulation of geometrically thick magnetically saturated accretion onto a high-spin Kerr black hole for a single cycle of magnetic flux eruption and accumulation. We show that following an eruption, a weakened jet develops a strong helical distortion with distinct structure of magnetic fields - the poloidal field along the jet core is unaffected by the eruption; while toroidal field lines, ejected from the black hole during the eruption and later re-advected onto it, form poloidal `bypasses' along the inner jet sheath. Such a distortion may appear in sources fed by geometrically thick accretion flows as an asymmetric superluminal knot, strongly interacting with the jet sheath along an oblique working surface. The jet section re-powered by magnetic flux re-accumulated on the black hole is tilted by a few degrees, implying significant variations in radiation boost towards observers of BL Lac blazars. The intrinsic structure of the jet spine is consistent with axisymmetric semi-analytical models.

The enigmatic population of ``Little Red Dots'' (LRDs) sit at the center of some of the largest debates in extragalactic astronomy today. The source(s) of ionizing emission and the physical scale over which it governs is still largely unknown. We show for the first time spectroscopic variability in a z ~ 7 LRD. Comparing a recently obtained 10.2 hr JWST/NIRSpec F290LP/G395M spectrum via the C3PO survey to an 8.4 hr F290LP/G395M spectrum taken 99 days earlier (~13 rest-days) via the THRILS survey, we find a ~30% $ difference in the continuum and broad-line flux, and a 42% difference between [OIII]5008 flux in the two epochs. Through rigorous testing, we confirm that such differences are not the result of differing MSA slit placements on source nor merely flux calibration offsets. These results are further corroborated by both a similar continuum and [OIII]5008 flux differences found in NIRSpec prism/clear observations of the source at an epoch taken approximately a year earlier than the THRILS observations via RUBIES and an additional observation fortuitously taken during the THRILS epoch (within a rest-day) via the CAPERS survey. Assuming LRDs are a type of accreting black hole system, this implies direct sight-lines must exist from the accretion disk to the surrounding nebular gas on scales beyond the broad-line region, and thus any high-density gas interpretations must allow for covering fractions < 100%. Furthermore, these results show the [OIII] line emission is likely not galaxy process-dominated, with a significant population of the narrow-line emitting gas closest to the broad-line region being directly ionized by the LRD. Finally, these results highlight the need for new approaches in inferring black hole properties of these systems, accounting for the lack of significant ionization via star formation, and/or exploring more exotic host-galaxy conditions at these early epochs.

Galaxy clusters grow through the matter accretion from the cosmic web, mainly along filaments. We aim to characterize the gas accretion onto clusters, focusing on the role of filaments in driving anisotropic inflows and thermodynamic properties, as it remains a key challenge for cosmology. In this study, we analyzed 415 galaxy clusters from the IllustrisTNG-300 hydrodynamical simulation at $z=0$. Anisotropic signatures are highlighted by probing both isotropically and anisotropically (gas in filaments only), the radial profiles of gas properties (including temperature, entropy, density, and pressure), and the radial velocity distributions. Our results highlight two distinct regimes of gas accretion depending on the cluster-centric distances. In the cluster environment ($\sim$ 2-4$R_{200}$), fast infalling warm gas tunneled by cosmic filaments enters the warm-hot circumcluster medium, but filaments remain colder due to their slow thermalization with the surrounding, generating transverse temperature gradients. At the cluster outskirts ($\sim$ 1-2$R_{200}$), gas infalling along filaments enters the hot intracluster medium, with a strong tangential velocity gradient. Warm gas tends to penetrate clusters from filaments, while hot gas is preferentially ejected beyond them. The mass and dynamical state of clusters significantly impact these accretion features, with relaxed and massive clusters exhibiting stronger and more extended temperature discontinuities. Overall, this work emphasizes a coherent picture of anisotropic gas accretion from filaments onto clusters. While virial shocks tend to be observed near the cluster boundary, especially at the filament-cluster interface. We do not find strong evidence of accretion shocks around filaments, suggesting slow thermalization of filament gas as it enters the dense warm-hot circum-cluster environment.

We describe the public release of $>2.3$ petabytes of data from the FLAMINGO cosmological simulations. The suite consists of hydrodynamical simulations that include radiative cooling, star formation, stellar mass loss and the resulting chemical enrichment, supernova feedback, and two implementations of AGN feedback. Neutrinos are simulated explicitly using particles. Data products include snapshots, halo and galaxy catalogues, HEALPix all-sky lightcone maps, particle data for lightcone maps, and power spectra. The FLAMINGO set includes 22 hydrodynamical simulations. In addition, there are 16 gravity-only simulations, including the $10080^3$ particles FLAMINGO-10k run, with initial conditions that match those of the corresponding hydrodynamical runs. The fiducial hydrodynamical simulations span three numerical resolutions that have each been calibrated to reproduce the present-day galaxy stellar mass function and gas fractions in low-redshift clusters. Other simulations systematically vary the galaxy stellar mass function, cluster gas fractions, cosmology (including neutrino masses), and/or the nature of dark matter, in volumes of 1Gpc$^3$. The release includes hitherto unpublished simulations that use extra dark matter particles. While we provide a facility for downloading complete simulation outputs, we recognise that for many users this will not be possible due to limited local storage or network bandwidth. We implement a web service that enables users to explore available outputs and selectively download datasets or parts of datasets.