arXiv Daily Digest - 2026-04-06

This paper explores the use of server learning for enhancing the robustness of federated learning against malicious attacks even when clients' training data are not independent and identically distributed. We propose a heuristic algorithm that uses server learning and client update filtering in combination with geometric median aggregation. We demonstrate via experiments that this approach can achieve significant improvement in model accuracy even when the fraction of malicious clients is high, even more than $50\%$ in some cases, and the dataset utilized by the server is small and could be synthetic with its distribution not necessarily close to that of the clients' aggregated data.

Large language models (LLMs) often produce confident but incorrect answers in settings where abstention would be safer. Standard evaluation protocols, however, require a response and do not account for how confidence should guide decisions under different risk preferences. To address this gap, we introduce the Behavioral Alignment Score (BAS), a decision-theoretic metric for evaluating how well LLM confidence supports abstention-aware decision making. BAS is derived from an explicit answer-or-abstain utility model and aggregates realized utility across a continuum of risk thresholds, yielding a measure of decision-level reliability that depends on both the magnitude and ordering of confidence. We show theoretically that truthful confidence estimates uniquely maximize expected BAS utility, linking calibration to decision-optimal behavior. BAS is related to proper scoring rules such as log loss, but differs structurally: log loss penalizes underconfidence and overconfidence symmetrically, whereas BAS imposes an asymmetric penalty that strongly prioritizes avoiding overconfident errors. Using BAS alongside widely used metrics such as ECE and AURC, we then construct a benchmark of self-reported confidence reliability across multiple LLMs and tasks. Our results reveal substantial variation in decision-useful confidence, and while larger and more accurate models tend to achieve higher BAS, even frontier models remain prone to severe overconfidence. Importantly, models with similar ECE or AURC can exhibit very different BAS due to highly overconfident errors, highlighting limitations of standard metrics. We further show that simple interventions, such as top-$k$ confidence elicitation and post-hoc calibration, can meaningfully improve confidence reliability. Overall, our work provides both a principled metric and a comprehensive benchmark for evaluating LLM confidence reliability.

Model predictive control (MPC) with learned world models has emerged as a promising paradigm for embodied control, particularly for its ability to generalize zero-shot when deployed in new environments. However, learned world models often struggle with long-horizon control due to the accumulation of prediction errors and the exponentially growing search space. In this work, we address these challenges by learning latent world models at multiple temporal scales and performing hierarchical planning across these scales, enabling long-horizon reasoning while substantially reducing inference-time planning complexity. Our approach serves as a modular planning abstraction that applies across diverse latent world-model architectures and domains. We demonstrate that this hierarchical approach enables zero-shot control on real-world non-greedy robotic tasks, achieving a 70% success rate on pick-&-place using only a final goal specification, compared to 0% for a single-level world model. In addition, across physics-based simulated environments including push manipulation and maze navigation, hierarchical planning achieves higher success while requiring up to 4x less planning-time compute.

The rapid adoption of the Internet of Medical Things (IoMT) is transforming healthcare by enabling seamless connectivity among medical devices, systems, and services. However, it also introduces serious cybersecurity and patient safety concerns as attackers increasingly exploit new methods and emerging vulnerabilities to infiltrate IoMT networks. This paper proposes a novel Tsetlin Machine (TM)-based Intrusion Detection System (IDS) for detecting a wide range of cyberattacks targeting IoMT networks. The TM is a rule-based and interpretable machine learning (ML) approach that models attack patterns using propositional logic. Extensive experiments conducted on the CICIoMT-2024 dataset, which includes multiple IoMT protocols and cyberattack types, demonstrate that the proposed TM-based IDS outperforms traditional ML classifiers. The proposed model achieves an accuracy of 99.5\% in binary classification and 90.7\% in multi-class classification, surpassing existing state-of-the-art approaches. Moreover, to enhance model trust and interpretability, the proposed TM-based model presents class-wise vote scores and clause activation heatmaps, providing clear insights into the most influential clauses and the dominant class contributing to the final model decision.

Three-dimensional medical image data and computer-aided decision making, particularly using deep learning, are becoming increasingly important in the medical field. To aid in these developments we introduce PR3DICTR: Platform for Research in 3D Image Classification and sTandardised tRaining. Built using community-standard distributions (PyTorch and MONAI), PR3DICTR provides an open-access, flexible and convenient framework for prediction model development, with an explicit focus on classification using three-dimensional medical image data. By combining modular design principles and standardization, it aims to alleviate developmental burden whilst retaining adjustability. It provides users with a wealth of pre-established functionality, for instance in model architecture design options, hyper-parameter solutions and training methodologies, but still gives users the opportunity and freedom to ``plug in'' their own solutions or modules. PR3DICTR can be applied to any binary or event-based three-dimensional classification task and can work with as little as two lines of code.

Agentic AI is increasingly judged not by fluent output alone but by whether it can act, remember, and verify under partial observability, delay, and strategic observation. Existing research often studies these demands separately: robotics emphasizes control, retrieval systems emphasize memory, and alignment or assurance work emphasizes checking and oversight. This article argues that squirrel ecology offers a sharp comparative case because arboreal locomotion, scatter-hoarding, and audience-sensitive caching couple all three demands in one organism. We synthesize evidence from fox, eastern gray, and, in one field comparison, red squirrels, and impose an explicit inference ladder: empirical observation, minimal computational inference, and AI design conjecture. We introduce a minimal hierarchical partially observed control model with latent dynamics, structured episodic memory, observer-belief state, option-level actions, and delayed verifier signals. This motivates three hypotheses: (H1) fast local feedback plus predictive compensation improves robustness under hidden dynamics shifts; (H2) memory organized for future control improves delayed retrieval under cue conflict and load; and (H3) verifiers and observer models inside the action-memory loop reduce silent failure and information leakage while remaining vulnerable to misspecification. A downstream conjecture is that role-differentiated proposer/executor/checker/adversary systems may reduce correlated error under asymmetric information and verification burden. The contribution is a comparative perspective and benchmark agenda: a disciplined program of falsifiable claims about the coupling of control, memory, and verifiable action.

All prior membership inference attacks for fine-tuned language models use hand-crafted heuristics (e.g., loss thresholding, Min-K\%, reference calibration), each bounded by the designer's intuition. We introduce the first transferable learned attack, enabled by the observation that fine-tuning any model on any corpus yields unlimited labeled data, since membership is known by construction. This removes the shadow model bottleneck and brings membership inference into the deep learning era: learning what matters rather than designing it, with generalization through training diversity and scale. We discover that fine-tuning language models produces an invariant signature of memorization detectable across architectural families and data domains. We train a membership inference classifier exclusively on transformer-based models. It transfers zero-shot to Mamba (state-space), RWKV-4 (linear attention), and RecurrentGemma (gated recurrence), achieving 0.963, 0.972, and 0.936 AUC respectively. Each evaluation combines an architecture and dataset never seen during training, yet all three exceed performance on held-out transformers (0.908 AUC). These four families share no computational mechanisms, their only commonality is gradient descent on cross-entropy loss. Even simple likelihood-based methods exhibit strong transfer, confirming the signature exists independently of the detection method. Our method, Learned Transfer MIA (LT-MIA), captures this signal most effectively by reframing membership inference as sequence classification over per-token distributional statistics. On transformers, LT-MIA achieves 2.8$\times$ higher TPR at 0.1\% FPR than the strongest baseline. The method also transfers to code (0.865 AUC) despite training only on natural language texts. Code and trained classifier available at https://github.com/JetBrains-Research/learned-mia.

Cardiovascular modeling has rapidly advanced over the past few decades due to the rising needs for health tracking and early detection of cardiovascular diseases. While 1-D arterial models offer an attractive compromise between computational efficiency and solution fidelity, their application on large populations or for generating large \emph{in silico} cohorts remains challenging. Certain hemodynamic parameters like the terminal resistance/compliance, are difficult to clinically estimate and often yield non-physiological hemodynamics when sampled naively, resulting in large portions of simulated datasets to be discarded. In this work, we present a systematic framework for training machine learning (ML) models, capable of instantaneous hemodynamic prediction and parameter estimation. We initially start with generating a parametric virtual cohort of patients which is based on the multivariate correlations observed in the large Asklepios clinical dataset, ensuring that physiological parameter distributions are respected. We then train a deep neural surrogate model, able to predict patient-specific arterial pressure and cardiac output (CO), enabling rapid a~priori screening of input parameters. This allows for immediate rejection of non-physiological combinations and drastically reduces the cost of targeted synthetic dataset generation (e.g. hypertensive groups). The model also provides a principled means of sampling the terminal resistance to minimize the uncertainties of unmeasurable parameters. Moreover, by assessing the model's predictive performance we determine the theoretical information which suffices for solving the inverse problem of estimating the CO. Finally, we apply the surrogate on a clinical dataset for the estimation of central aortic hemodynamics i.e. the CO and aortic systolic blood pressure (cSBP).

Autonomous coding agents are generating code at an unprecedented scale, with OpenAI Codex alone creating over 400,000 pull requests (PRs) in two months. As agentic PR volumes increase, code review agents (CRAs) have become routine gatekeepers in development workflows. Industry reports claim that CRAs can manage 80% of PRs in open source repositories without human involvement. As a result, understanding the effectiveness of CRA reviews is crucial for maintaining developmental workflows and preventing wasted effort on abandoned pull requests. However, empirical evidence on how CRA feedback quality affects PR outcomes remains limited. The goal of this paper is to help researchers and practitioners understand when and how CRAs influence PR merge success by empirically analyzing reviewer composition and the signal quality of CRA-generated comments. From AIDev's 19,450 PRs, we analyze 3,109 unique PRs in the commented review state, comparing human-only versus CRA-only reviews. We examine 98 closed CRA-only PRs to assess whether low signal-to-noise ratios contribute to abandonment. CRA-only PRs achieve a 45.20% merge rate, 23.17 percentage points lower than human-only PRs (68.37%), with significantly higher abandonment. Our signal-to-noise analysis reveals that 60.2% of closed CRA-only PRs fall into the 0-30% signal range, and 12 of 13 CRAs exhibit average signal ratios below 60%, indicating substantial noise in automated review feedback. These findings suggest that CRAs without human oversight often generate low-signal feedback associated with higher abandonment. For practitioners, our results indicate that CRAs should augment rather than replace human reviewers and that human involvement remains critical for effective and actionable code review.

We study multiteacher knowledge distillation for low resource abstractive summarization from a reliability aware perspective. We introduce EWAD (Entropy Weighted Agreement Aware Distillation), a token level mechanism that routes supervision between teacher distillation and gold supervision based on inter teacher agreement, and CPDP (Capacity Proportional Divergence Preservation), a geometric constraint on the student position relative to heterogeneous teachers. Across two Bangla datasets, 13 BanglaT5 ablations, and eight Qwen2.5 experiments, we find that logit level KD provides the most reliable gains, while more complex distillation improves semantic similarity for short summaries but degrades longer outputs. Cross lingual pseudo label KD across ten languages retains 71-122 percent of teacher ROUGE L at 3.2x compression. A human validated multi judge LLM evaluation further reveals calibration bias in single judge pipelines. Overall, our results show that reliability aware distillation helps characterize when multi teacher supervision improves summarization and when data scaling outweighs loss engineering.

Scaling Vision-Language-Action (VLA) models by upgrading the vision encoder is expected to improve downstream manipulation performance--as it does in vision-language modeling. We show that this expectation fails when actions are represented as discrete tokens, and explain why through an information-theoretic principle we call the Compression Gap: in any visuomotor pipeline, scaling behavior is governed by the location of the tightest information bottleneck. When actions are continuous (e.g., Diffusion Policy), the vision encoder is the binding constraint, and upgrading it directly improves performance. When actions are discretized through a fixed-capacity codebook (e.g., OAT), the codebook becomes the binding constraint, and encoder improvements cannot propagate past it--regardless of how rich the upstream representation is. We validate this principle on the LIBERO benchmark with three lines of evidence: a factorial experiment showing that encoder upgrades improve Diffusion Policy by over 21 percentage points while OAT gains are substantially attenuated across model scales; an encoder quality gradient across four encoders confirming that Diffusion Policy tracks encoder quality monotonically while OAT remains flat; and a codebook size experiment demonstrating that relaxing codebook capacity partially recovers encoder sensitivity, providing causal evidence for the bottleneck hypothesis. Our findings reveal that scaling in Physical AI requires identifying where information bottlenecks lie in the pipeline, rather than uniformly increasing model or data size.

Transformer attention computes a single softmax-weighted average over values -- a one-pass estimate that cannot correct its own errors. We introduce \emph{gradient-boosted attention}, which applies the principle of gradient boosting \emph{within} a single attention layer: a second attention pass, with its own learned projections, attends to the prediction error of the first and applies a gated correction. Under a squared reconstruction objective, the construction maps onto Friedman's gradient boosting machine, with each attention pass as a base learner and the per-dimension gate as the shrinkage parameter. We show that a single Hopfield-style update erases all query information orthogonal to the stored-pattern subspace, and that further iteration under local contraction can collapse distinct queries in the same region to the same fixed point. We also show that separate projections for the correction pass can recover residual information inaccessible to the shared-projection approach of Tukey's twicing. On a 10M-token subset of WikiText-103, gradient-boosted attention achieves a test perplexity of $67.9$ compared to $72.2$ for standard attention, $69.6$ for Twicing Attention, and $69.0$ for a parameter-matched wider baseline, with two rounds capturing most of the benefit.

Generally capable agents must learn from experience in ways that generalize across tasks and environments. The fundamental problems of learning, including credit assignment, overfitting, forgetting, local optima, and high-variance learning signals, persist whether the learned object lies in parameter space or context space. While these challenges are well understood in classical machine learning optimization, they remain underexplored in context space, leading current methods to be fragmented and ad hoc. We present Reflective Context Learning (RCL), a unified framework for agents that learn through repeated interaction, reflection on behavior and failure modes, and iterative updates to context. In RCL, reflection converts trajectories and current context into a directional update signal analogous to gradients, while mutation applies that signal to improve future behavior in context space. We recast recent context-optimization approaches as instances of this shared learning problem and systematically extend them with classical optimization primitives, including batching, improved credit-assignment signal, auxiliary losses, failure replay, and grouped rollouts for variance reduction. On AppWorld, BrowseComp+, and RewardBench2, these primitives improve over strong baselines, with their relative importance shifting across task regimes. We further analyze robustness to initialization, the effects of batch size, sampling and curriculum strategy, optimizer-state variants, and the impact of allocating stronger or weaker models to different optimization components. Our results suggest that learning through context updates should be treated not as a set of isolated algorithms, but as an optimization problem whose mechanisms can be studied systematically and improved through transferable principles.

Spiking neural networks encode information in spike timing and offer a pathway toward energy efficient artificial intelligence. However, a key challenge in spiking neural networks is realizing nonlinear and expressive computation in compact, energy-efficient hardware without relying on additional circuit complexity. In this work, we examine nonlinear computation in a CMOS+X spiking neuron implemented with a magnetic tunnel junction connected in series with an NMOS transistor. Circuit simulations of a multilayer network solving the XOR classification problem show that three intrinsic neuronal properties enable nonlinear behavior: threshold activation, response latency, and absolute refraction. Threshold activation determines which neurons participate in computation, response latency shifts spike timing, and absolute refraction suppresses subsequent spikes. These results show that magnetization dynamics of MTJ devices can support nonlinear computation in compact neuromorphic hardware.

In this paper, we propose Precision-Informed Semantic Modeling (PRISM), a structured topic modeling framework combining the benefits of rich representations captured by LLMs with the low cost and interpretability of latent semantic clustering methods. PRISM fine-tunes a sentence encoding model using a sparse set of LLM- provided labels on samples drawn from some corpus of interest. We segment this embedding space with thresholded clustering, yielding clusters that separate closely related topics within some narrow domain. Across multiple corpora, PRISM improves topic separability over state-of-the-art local topic models and even over clustering on large, frontier embedding models while requiring only a small number of LLM queries to train. This work contributes to several research streams by providing (i) a student-teacher pipeline to distill sparse LLM supervision into a lightweight model for topic discovery; (ii) an analysis of the efficacy of sampling strategies to improve local geometry for cluster separability; and (iii) an effective approach for web-scale text analysis, enabling researchers and practitioners to track nuanced claims and subtopics online with an interpretable, locally deployable framework.

The recent success of reinforcement learning (RL) in large reasoning models has inspired the growing adoption of RL for post-training Multimodal Large Language Models (MLLMs) to enhance their visual reasoning capabilities. Although many studies have reported improved performance, it remains unclear whether RL training truly enables models to learn from visual information. In this work, we propose the Hallucination-as-Cue Framework, an analytical framework designed to investigate the effects of RL-based post-training on multimodal reasoning models from the perspective of model hallucination. Specifically, we introduce hallucination-inductive, modality-specific corruptions that remove or replace essential information required to derive correct answers, thereby forcing the model to reason by hallucination. By applying these corruptions during both training and evaluation, our framework provides a unique perspective for diagnosing RL training dynamics and understanding the intrinsic properties of datasets. Through extensive experiments and analyses across multiple multimodal reasoning benchmarks, we reveal that the role of model hallucination for RL-training is more significant than previously recognized. For instance, we find that RL post-training under purely hallucination-inductive settings can still significantly improve models' reasoning performance, and in some cases even outperform standard training. These findings challenge prevailing assumptions about MLLM reasoning training and motivate the development of more modality-aware RL-based training designs.

Large language models (LLMs) encode vast world knowledge in their parameters, yet they remain fundamentally limited by static knowledge, finite context windows, and weakly structured causal reasoning. This survey provides a unified account of augmentation strategies along a single axis: the degree of structured context supplied at inference time. We cover in-context learning and prompt engineering, Retrieval-Augmented Generation (RAG), GraphRAG, and CausalRAG. Beyond conceptual comparison, we provide a transparent literature-screening protocol, a claim-audit framework, and a structured cross-paper evidence synthesis that distinguishes higher-confidence findings from emerging results. The paper concludes with a deployment-oriented decision framework and concrete research priorities for trustworthy retrieval-augmented NLP.

Large language models and deep research agents supply citation URLs to support their claims, yet the reliability of these citations has not been systematically measured. We address six research questions about citation URL validity using 10 models and agents on DRBench (53,090 URLs) and 3 models on ExpertQA (168,021 URLs across 32 academic fields). We find that 3--13\% of citation URLs are hallucinated -- they have no record in the Wayback Machine and likely never existed -- while 5--18\% are non-resolving overall. Deep research agents generate substantially more citations per query than search-augmented LLMs but hallucinate URLs at higher rates. Domain effects are pronounced: non-resolving rates range from 5.4\% (Business) to 11.4\% (Theology), with per-model effects even larger. Decomposing failures reveals that some models fabricate every non-resolving URL, while others show substantial link-rot fractions indicating genuine retrieval. As a solution, we release urlhealth, an open-source tool for URL liveness checking and stale-vs-hallucinated classification using the Wayback Machine. In agentic self-correction experiments, models equipped with urlhealth reduce non-resolving citation URLs by $6\textrm{--}79\times$ to under 1\%, though effectiveness depends on the model's tool-use competence. The tool and all data are publicly available. Our characterization findings, failure taxonomy, and open-source tooling establish that citation URL validity is both measurable at scale and correctable in practice.

Large language models with web search are increasingly used in scientific publishing agents, yet they still produce BibTeX entries with pervasive field-level errors. Prior evaluations tested base models without search, which does not reflect current practice. We construct a benchmark of 931 papers across four scientific domains and three citation tiers -- popular, low-citation, and recent post-cutoff -- designed to disentangle parametric memory from search dependence, with version-aware ground truth accounting for multiple citable versions of the same paper. Three search-enabled frontier models (GPT-5, Claude Sonnet-4.6, Gemini-3 Flash) generate BibTeX entries scored on nine fields and a six-way error taxonomy, producing ~23,000 field-level observations. Overall accuracy is 83.6%, but only 50.9% of entries are fully correct; accuracy drops 27.7pp from popular to recent papers, revealing heavy reliance on parametric memory even when search is available. Field-error co-occurrence analysis identifies two failure modes: wholesale entry substitution (identity fields fail together) and isolated field error. We evaluate clibib, an open-source tool for deterministic BibTeX retrieval from the Zotero Translation Server with CrossRef fallback, as a mitigation mechanism. In a two-stage integration where baseline entries are revised against authoritative records, accuracy rises +8.0pp to 91.5%, fully correct entries rise from 50.9% to 78.3%, and regression rate is only 0.8%. An ablation comparing single-stage and two-stage integration shows that separating search from revision yields larger gains and lower regression (0.8% vs. 4.8%), demonstrating that integration architecture matters independently of model capability. We release the benchmark, error taxonomy, and clibib tool to support evaluation and mitigation of citation hallucinations in LLM-based scientific writing.

The recent advancements in Vision Language Models (VLMs) have demonstrated progress toward true intelligence requiring robust reasoning capabilities. Beyond pattern recognition, linguistic reasoning must integrate with visual comprehension, particularly for Chart Question Answering (CQA) tasks involving complex data visualizations. Current VLMs face significant limitations in CQA, including imprecise numerical extraction, difficulty interpreting implicit visual relationships, and inadequate attention mechanisms for capturing spatial relationships in charts. In this work, we address these challenges by presenting Chart-RL, a novel reinforcement learning framework that enhances VLMs chart understanding through feedback-driven policy optimization of visual perception and logical inference. Our key innovation includes a comprehensive framework integrating Reinforcement Learning (RL) from Policy Optimization techniques along with adaptive reward functions, that demonstrates superior performance compared to baseline foundation models and competitive results against larger state-of-the-art architectures. We also integrated Parameter-Efficient Fine-Tuning through Low-Rank Adaptation (LoRA) in the RL framework that only requires single GPU configurations while preserving performance integrity. We conducted extensive benchmarking across open-source, proprietary, and state-of-the-art closed-source models utilizing the ChartQAPro dataset. The RL fine-tuned Qwen3-VL-4B-Instruct model achieved an answer accuracy of 0.634, surpassing the 0.580 accuracy of the Qwen3-VL-8B-Instruct foundation model despite utilizing half the parameter count, while simultaneously reducing inference latency from 31 seconds to 9 seconds.

Graph domain adaptation (GDA) aims to transfer knowledge from a labeled source graph to an unlabeled target graph under distribution shifts. However, existing methods are largely feature-centric and overlook structural discrepancies, which become particularly detrimental under significant topology shifts. Such discrepancies alter both geometric relationships and spectral properties, leading to unreliable transfer of graph neural networks (GNNs). To address this limitation, we propose Dual-Aligned Structural Basis Distillation (DSBD) for GDA, a novel framework that explicitly models and adapts cross-domain structural variation. DSBD constructs a differentiable structural basis by synthesizing continuous probabilistic prototype graphs, enabling gradient-based optimization over graph topology. The basis is learned under source-domain supervision to preserve semantic discriminability, while being explicitly aligned to the target domain through a dual-alignment objective. Specifically, geometric consistency is enforced via permutation-invariant topological moment matching, and spectral consistency is achieved through Dirichlet energy calibration, jointly capturing structural characteristics across domains. Furthermore, we introduce a decoupled inference paradigm that mitigates source-specific structural bias by training a new GNN on the distilled structural basis. Extensive experiments on graph and image benchmarks demonstrate that DSBD consistently outperforms state-of-the-art methods.

Purpose: Proton magnetic resonance spectroscopic imaging ($^1$H MRSI) enables the mapping of whole-brain metabolites concentrations in-vivo. However, a long-standing problem for its clinical applicability is the metabolic quantification, which can require extensive time for spectral fitting. Recently, deep learning methods have been able to provide whole-brain metabolic quantification in only a few seconds. However, neural network implementations often lack configurability and require retraining to change predefined parameter settings. Methods: We introduce HyperFitS, a hypernetwork for spectral fitting for metabolite quantification in whole-brain $^1$H MRSI that flexibly adapts to a broad range of baseline corrections and water suppression factors. Metabolite maps of human subjects acquired at 3T and 7T with isotropic resolutions of 10 mm, 3.4 mm and 2 mm by water-suppressed and water-unsuppressed MRSI were quantified with HyperFitS and compared to conventional LCModel fitting. Results: Metabolic maps show a substantial agreement between the new and gold-standard methods, with significantly faster fitting times by HyperFitS. Quantitative results further highlight the impact of baseline parametrization on metabolic quantification, which can alter results by up to 30%. Conclusion: HyperFitS shows strong agreement with state-of-the-art conventional methods, while reducing processing times from hours to a few seconds. Compared to prior deep learning based spectral fitting methods, HyperFitS enables a wide range of configurability and can adapt to data quality acquired with multiple protocols and field strengths without retraining.

We present a method to identify a valence-arousal (VA) subspace within large language model representations. From 211k emotion-labeled texts, we derive emotion steering vectors, then learn VA axes as linear combinations of their top PCA components via ridge regression on the model's self-reported valence-arousal scores. The resulting VA subspace exhibits circular geometry consistent with established models of human emotion perception. Projections along our recovered VA subspace correlate with human-crowdsourced VA ratings across 44k lexical items. Furthermore, steering generation along these axes produces monotonic shifts in the corresponding affective dimensions of model outputs. Steering along these directions also induces near-monotonic bidirectional control over refusal and sycophancy: increasing arousal decreases refusal and increases sycophancy, and vice versa. These effects replicate across Llama-3.1-8B, Qwen3-8B, and Qwen3-14B, demonstrating cross-architecture generality. We provide a mechanistic account for these effects and prior emotionally-framed controls: refusal-associated tokens ("I can't," "sorry") occupy low-arousal, negative-valence regions, so VA steering directly modulates their emission probability.

We study high-dimensional convex empirical risk minimization (ERM) under general non-Gaussian data designs. By heuristically extending the Convex Gaussian Min-Max Theorem (CGMT) to non-Gaussian settings, we derive an asymptotic min-max characterization of key statistics, enabling approximation of the mean $μ_{\hatθ}$ and covariance $C_{\hatθ}$ of the ERM estimator $\hatθ$. Specifically, under a concentration assumption on the data matrix and standard regularity conditions on the loss and regularizer, we show that for a test covariate $x$ independent of the training data, the projection $\hatθ^\top x$ approximately follows the convolution of the (generally non-Gaussian) distribution of $μ_{\hatθ}^\top x$ with an independent centered Gaussian variable of variance $\text{Tr}(C_{\hatθ}\mathbb{E}[xx^\top])$. This result clarifies the scope and limits of Gaussian universality for ERMs. Additionally, we prove that any $\mathcal{C}^2$ regularizer is asymptotically equivalent to a quadratic form determined solely by its Hessian at zero and gradient at $μ_{\hatθ}$. Numerical simulations across diverse losses and models are provided to validate our theoretical predictions and qualitative insights.

Resource sharing in multi-tenant cloud environments enables cost efficiency but introduces the Noisy Neighbor problem, i.e., co-located workloads that unpredictably degrade each other's performance. Despite extensive research on detecting such effects, there are no explainable methodologies for quantifying the severity of impact and establishing causal relationships among tenants. We propose an analytical that combines controlled experimentation with multi-stage causal inference and validates it across 10 independent rounds in a Kubernetes testbed. Our methodology not only quantifies severe performance degradations (e.g., up to 67\% in I/O-bound workloads under combined stress) but also statistically establishes causality through Granger causality analysis, revealing a 75\% increase in causal links when the noisy neighbor activates. Furthermore, we identify unique "degradation signatures" for each resource contention vector (i.e., CPU, memory, disk, network), enabling diagnostic capabilities that go beyond anomaly detection. This work transforms the Noisy Neighbor from an elusive problem into a quantifiable, diagnosable phenomenon, providing cloud operators with actionable insights for SLA management and smart resource allocation.

Industrial software development across chip design, GPU optimization, and embedded systems lacks expert reasoning traces showing how engineers reason about hardware constraints and timing semantics. In this work, we propose InCoder-32B-Thinking, trained on the data from the Error-driven Chain-of-Thought (ECoT) synthesis framework with an industrial code world model (ICWM) to generate reasoning traces. Specifically, ECoT generates reasoning chains by synthesizing the thinking content from multi-turn dialogue with environmental error feedback, explicitly modeling the error-correction process. ICWM is trained on domain-specific execution traces from Verilog simulation, GPU profiling, etc., learns the causal dynamics of how code affects hardware behavior, and enables self-verification by predicting execution outcomes before actual compilation. All synthesized reasoning traces are validated through domain toolchains, creating training data matching the natural reasoning depth distribution of industrial tasks. Evaluation on 14 general (81.3% on LiveCodeBench v5) and 9 industrial benchmarks (84.0% in CAD-Coder and 38.0% on KernelBench) shows InCoder-32B-Thinking achieves top-tier open-source results across all domains.GPU Optimization

Multi-agent LLM applications organize execution in synchronized rounds where a central scheduler gathers outputs from all agents and redistributes the combined context. This All-Gather communication pattern creates massive KV Cache redundancy, because every agent's prompt contains the same shared output blocks, yet existing reuse methods fail to exploit it efficiently. We present TokenDance, a system that scales the number of concurrent agents by exploiting the All-Gather pattern for collective KV Cache sharing. TokenDance's KV Collector performs KV Cache reuse over the full round in one collective step, so the cost of reusing a shared block is paid once regardless of agent count. Its Diff-Aware Storage encodes sibling caches as block-sparse diffs against a single master copy, achieving 11-17x compression on representative workloads. Evaluation on GenerativeAgents and AgentSociety shows that TokenDance supports up to 2.7x more concurrent agents than vLLM with prefix caching under SLO requirement, reduces per-agent KV Cache storage by up to 17.5x, and achieves up to 1.9x prefill speedup over per-request position-independent caching.

Evaluating the factuality of long-form output generated by large language models (LLMs) remains challenging, particularly when responses are open-ended and contain many fine-grained factual statements. Existing evaluation methods primarily focus on precision: they decompose a response into atomic claims and verify each claim against external knowledge sources such as Wikipedia. However, this overlooks an equally important dimension of factuality: recall, whether the generated response covers the relevant facts that should be included. We propose a comprehensive factuality evaluation framework that jointly measures precision and recall. Our method leverages external knowledge sources to construct reference facts and determine whether they are captured in generated text. We further introduce an importance-aware weighting scheme based on relevance and salience. Our analysis reveals that current LLMs perform substantially better on precision than on recall, suggesting that factual incompleteness remains a major limitation of long-form generation and that models are generally better at covering highly important facts than the full set of relevant facts.

As AI-generated fiction becomes increasingly prevalent, questions of authorship and originality are becoming central to how written work is evaluated. While most existing work in this space focuses on identifying surface-level signatures of AI writing, we ask instead whether AI-generated stories can be distinguished from human ones without relying on stylistic signals, focusing on discourse-level narrative choices such as character agency and chronological discontinuity. We propose StoryScope, a pipeline that automatically induces a fine-grained, interpretable feature space of discourse-level narrative features across 10 dimensions. We apply StoryScope to a parallel corpus of 10,272 writing prompts, each written by a human author and five LLMs, yielding 61,608 stories, each ~5,000 words, and 304 extracted features per story. Narrative features alone achieve 93.2% macro-F1 for human vs. AI detection and 68.4% macro-F1 for six-way authorship attribution, retaining over 97% of the performance of models that include stylistic cues. A compact set of 30 core narrative features captures much of this signal: AI stories over-explain themes and favor tidy, single-track plots while human stories frame protagonist' choices as more morally ambiguous and have increased temporal complexity. Per-model fingerprint features enable six-way attribution: for example, Claude produces notably flat event escalation, GPT over-indexes on dream sequences, and Gemini defaults to external character description. We find that AI-generated stories cluster in a shared region of narrative space, while human-authored stories exhibit greater diversity. More broadly, these results suggest that differences in underlying narrative construction, not just writing style, can be used to separate human-written original works from AI-generated fiction.

Many software systems originate as prototypes or minimum viable products (MVPs), developed with an emphasis on delivery speed and responsiveness to changing requirements rather than long-term code maintainability. While effective for rapid delivery, this approach can result in codebases that are difficult to modify, presenting a significant opportunity cost in the era of AI-assisted or even AI-led programming. In this paper, we present a case study of using coding models for automated unit test generation and subsequent safe refactoring, with proposed code changes validated by passing tests. The study examines best practices for iteratively generating tests to capture existing system behavior, followed by model-assisted refactoring under developer supervision. We describe how this workflow constrained refactoring changes, the errors and limitations observed in both phases, the efficiency gains achieved, when manual intervention was necessary, and how we addressed the weak value misalignment we observed in models. Using this approach, we generated nearly 16,000 lines of reliable unit tests in hours rather than weeks, achieved up to 78\% branch coverage in critical modules, and significantly reduced regression risk during large-scale refactoring. These results illustrate software engineering's shift toward an empirical science, emphasizing data collection and constraining mechanisms that support fast, safe iteration.

Tool-augmented AI agents substantially extend the practical capabilities of large language models, but they also introduce security risks that cannot be identified through model-only evaluation. In this paper, we present a systematic security assessment of six representative OpenClaw-series agent frameworks, namely OpenClaw, AutoClaw, QClaw, KimiClaw, MaxClaw, and ArkClaw, under multiple backbone models. To support this study, we construct a benchmark of 205 test cases covering representative attack behaviors across the full agent execution lifecycle, enabling unified evaluation of risk exposure at both the framework and model levels. Our results show that all evaluated agents exhibit substantial security vulnerabilities, and that agentized systems are significantly riskier than their underlying models used in isolation. In particular, reconnaissance and discovery behaviors emerge as the most common weaknesses, while different frameworks expose distinct high-risk profiles, including credential leakage, lateral movement, privilege escalation, and resource development. These findings indicate that the security of modern agent systems is shaped not only by the safety properties of the backbone model, but also by the coupling among model capability, tool use, multi-step planning, and runtime orchestration. We further show that once an agent is granted execution capability and persistent runtime context, weaknesses arising in early stages can be amplified into concrete system-level failures. Overall, our study highlights the need to move beyond prompt-level safeguards toward lifecycle-wide security governance for intelligent agent frameworks.

On-policy distillation (OPD) has become a popular training paradigm in the LLM community. This paradigm selects a larger model as the teacher to provide dense, fine-grained signals for each sampled trajectory, in contrast to reinforcement learning with verifiable rewards (RLVR), which only obtains sparse signals from verifiable outcomes in the environment. Recently, the community has explored on-policy self-distillation (OPSD), where the same model serves as both teacher and student, with the teacher receiving additional privileged information such as reference answers to enable self-evolution. This paper demonstrates that learning signals solely derived from the privileged teacher result in severe information leakage and unstable long-term training. Accordingly, we identify the optimal niche for self-distillation and propose \textbf{RLSD} (\textbf{RL}VR with \textbf{S}elf-\textbf{D}istillation). Specifically, we leverage self-distillation to obtain token-level policy differences for determining fine-grained update magnitudes, while continuing to use RLVR to derive reliable update directions from environmental feedback (e.g., response correctness). This enables RLSD to simultaneously harness the strengths of both RLVR and OPSD, achieving a higher convergence ceiling and superior training stability.

Automated annotation of pedagogical dialogue is a high-stakes task where LLMs often fail without sufficient domain grounding. We present a domain-adapted RAG pipeline for tutoring move annotation. Rather than fine-tuning the generative model, we adapt retrieval by fine-tuning a lightweight embedding model on tutoring corpora and indexing dialogues at the utterance level to retrieve labeled few-shot demonstrations. Evaluated across two real tutoring dialogue datasets (TalkMoves and Eedi) and three LLM backbones (GPT-5.2, Claude Sonnet 4.6, Qwen3-32b), our best configuration achieves Cohen's $κ$ of 0.526-0.580 on TalkMoves and 0.659-0.743 on Eedi, substantially outperforming no-retrieval baselines ($κ= 0.275$-$0.413$ and $0.160$-$0.410$). An ablation study reveals that utterance-level indexing, rather than embedding quality alone, is the primary driver of these gains, with top-1 label match rates improving from 39.7\% to 62.0\% on TalkMoves and 52.9\% to 73.1\% on Eedi under domain-adapted retrieval. Retrieval also corrects systematic label biases present in zero-shot prompting and yields the largest improvements for rare and context-dependent labels. These findings suggest that adapting the retrieval component alone is a practical and effective path toward expert-level pedagogical dialogue annotation while keeping the generative model frozen.

Kimi K2.5 is an open-weight LLM that rivals closed models across coding, multimodal, and agentic benchmarks, but was released without an accompanying safety evaluation. In this work, we conduct a preliminary safety assessment of Kimi K2.5 focusing on risks likely to be exacerbated by powerful open-weight models. Specifically, we evaluate the model for CBRNE misuse risk, cybersecurity risk, misalignment, political censorship, bias, and harmlessness, in both agentic and non-agentic settings. We find that Kimi K2.5 shows similar dual-use capabilities to GPT 5.2 and Claude Opus 4.5, but with significantly fewer refusals on CBRNE-related requests, suggesting it may uplift malicious actors in weapon creation. On cyber-related tasks, we find that Kimi K2.5 demonstrates competitive cybersecurity performance, but it does not appear to possess frontier-level autonomous cyberoffensive capabilities such as vulnerability discovery and exploitation. We further find that Kimi K2.5 shows concerning levels of sabotage ability and self-replication propensity, although it does not appear to have long-term malicious goals. In addition, Kimi K2.5 exhibits narrow censorship and political bias, especially in Chinese, and is more compliant with harmful requests related to spreading disinformation and copyright infringement. Finally, we find the model refuses to engage in user delusions and generally has low over-refusal rates. While preliminary, our findings highlight how safety risks exist in frontier open-weight models and may be amplified by the scale and accessibility of open-weight releases. Therefore, we strongly urge open-weight model developers to conduct and release more systematic safety evaluations required for responsible deployment.

VLMs trained on web-scale data retain sensitive and copyrighted visual concepts that deployment may require removing. Training-based unlearning methods share a structural flaw: fine-tuning on a narrow forget set degrades general capabilities before unlearning begins, making it impossible to attribute subsequent performance drops to the unlearning procedure itself. Training-free approaches sidestep this by suppressing concepts through prompts or system instructions, but no rigorous benchmark exists for evaluating them on visual tasks. We introduce VLM-UnBench, the first benchmark for training-free visual concept unlearning in VLMs. It covers four forgetting levels, 7 source datasets, and 11 concept axes, and pairs a three-level probe taxonomy with five evaluation conditions to separate genuine forgetting from instruction compliance. Across 8 evaluation settings and 13 VLM configurations, realistic unlearning prompts leave forget accuracy near the no-instruction baseline; meaningful reductions appear only under oracle conditions that disclose the target concept to the model. Object and scene concepts are the most resistant to suppression, and stronger instruction-tuned models remain capable despite explicit forget instructions. These results expose a clear gap between prompt-level suppression and true visual concept erasure.

Large language models (LLMs) are effective for automated program repair, but plausible patches that pass the full test suite often rewrite more code than necessary, increasing review and maintenance costs. This over-editing is common because most bugs are localized, while standard supervised fine-tuning provides no explicit signal about which tokens should be preserved and which should be changed. We propose PAFT, a preservation-aware fine-tuning method for minimal-edit program repair. PAFT derives token-level preservation signals by aligning buggy and fixed code, combines them with full-sequence masking, and applies an edit-difficulty curriculum. Across Defects4J and HumanEval-Java, PAFT improves pass@1 by up to 65.6% over standard supervised fine-tuning (StdFT) while reducing average edit distance (AED) by up to 32.6%. On Defects4J with DeepSeek-Coder-6.7B, PAFT also outperforms AdaPatcher, a strong preference-based repair baseline, improving pass@1 from 5.9% to 10.1% while reducing median AED from 61.0 to 42.0. Overall, PAFT preserves stable context and concentrates edits on faulty regions, yielding smaller, more localized, plausible patches without inference-time search, reranking, or post-processing.

Knowledge distillation is an effective technique for pre-trained language model compression. However, existing methods only focus on the knowledge distribution among layers, which may cause the loss of fine-grained information in the alignment process. To address this issue, we introduce the Multi-aspect Knowledge Distillation (MaKD) method, which mimics the self-attention and feed-forward modules in greater depth to capture rich language knowledge information at different aspects. Experimental results demonstrate that MaKD can achieve competitive performance compared with various strong baselines with the same storage parameter budget. In addition, our method also performs well in distilling auto-regressive architecture models.

Cyber-attacks continue to grow in scale and sophistication, yet existing network intrusion detection approaches lack the semantic depth required for path reasoning over attacker-victim interactions. We address this by first modelling network alerts as a knowledge graph, then formulating hyper-relational alert prediction as a hyper-relational knowledge graph completion (HR-KGC) problem, representing each network alert as a qualified statement (h, r, t, Q), where h and t are source and destination IPs, r denotes the attack type, and Q encodes flow-level metadata such as timestamps, ports, protocols, and attack intensity, going beyond standard KGC binary triples (h, r, t) that would discard this contextual richness. We introduce five models across three contributions: first, Hyper-relational Neural Bellman-Ford (HR-NBFNet) extends Neural Bellman-Ford Networks to the hyper-relational setting with qualifier-aware multi-hop path reasoning, while its multi-task variant MT-HR-NBFNet jointly predicts tail, relation, and qualifier-value within a single traversal pass; second, AlertStar fuses qualifier context and structural path information entirely in embedding space via cross-attention and learned path composition, and its multi-task extension MT-AlertStar eliminates the overhead of full knowledge graph propagation; third, HR-NBFNet-CQ extends qualifier-aware representations to answer complex first-order logic queries, including one-hop, two-hop chain, two-anchor intersection, and union, enabling multi-condition threat reasoning over the alert knowledge graph. Evaluated inductively on the Warden and UNSW-NB15 benchmarks across three qualifier-density regimes, AlertStar and MT-AlertStar achieve superior MR, MRR, and Hits@k, demonstrating that local qualifier fusion is both sufficient and more efficient than global path propagation for hyper-relational alert prediction.

Large language model (LLM) agents learn by interacting with environments, but long-horizon training remains fundamentally bottlenecked by sparse and delayed rewards. Existing methods typically address this challenge through post-hoc credit assignment or external reward models, which provide limited guidance at inference time and often separate reward improvement from policy improvement. We propose Self-Guide, a self-generated internal reward for language agents that supports both inference-time guidance and training-time supervision. Specifically, the agent uses Self-Guide as a short self-guidance signal to steer the next action during inference, and converts the same signal into step-level internal reward for denser policy optimization during training. This creates a co-evolving loop: better policy produces better guidance, and better guidance further improves policy as internal reward. Across three agent benchmarks, inference-time self-guidance already yields clear gains, while jointly evolving policy and internal reward with GRPO brings further improvements (8\%) over baselines trained solely with environment reward. Overall, our results suggest that language agents can improve not only by collecting more experience, but also by learning to generate and refine their own internal reward during acting and learning.

Accurate and automated sea ice classification is important for climate monitoring and maritime safety in the Arctic. While Synthetic Aperture Radar (SAR) is the operational standard because of its all-weather capability, it remains challenging to distinguish morphologically similar ice classes under severe class imbalance. Rather than claiming a fully validated multimodal system, this paper establishes a trustworthy SAR only baseline that future fusion work can build upon. Using the AI4Arctic/ASIP Sea Ice Dataset (v2), which contains 461 Sentinel-1 scenes matched with expert ice charts, we combine full-resolution Sentinel-1 Extra Wide inputs, leakage-aware stratified patch splitting, SIGRID-3 stage-of-development labels, and training-set normalization to evaluate Vision Transformer baselines. We compare ViT-Base models trained with cross entropy and weighted cross-entropy against a ViT-Large model trained with focal loss. Among the tested configurations, ViT-Large with focal loss achieves 69.6% held-out accuracy, 68.8% weighted F1, and 83.9% precision on the minority Multi-Year Ice class. These results show that focal-loss training offers a more useful precision-recall trade-off than weighted cross-entropy for rare ice classes and establishes a cleaner baseline for future multimodal fusion with optical, thermal, or meteorological data.

LLM agents increasingly adopt skills as a reusable unit of composition. While skills are shared across diverse agent platforms, current systems treat them as raw context, causing the same skill to behave inconsistently for different agents. This fragility undermines skill portability and execution efficiency. To address this challenge, we analyze 118,000 skills and draw inspiration from traditional compiler design. We treat skills as code and LLMs as heterogeneous processors. To make portability actionable, we decompose a skill's requirements into a set of primitive capabilities, and measure how well each model-harness pair supports them. Based on these capability profiles, we propose SkillRT, a compilation and runtime system designed for portable and efficient skill execution. At compile time, SkillRT performs capability-based compilation, environment binding, and concurrency extraction. At runtime, SkillRT applies JIT code solidification and adaptive recompilation for performance optimization. We evaluate SkillRT across eight LLMs of varying scales and three agent harnesses, covering SkillsBench and representative skill tasks. Results demonstrate that SkillRT significantly improves task completion rates across different models and environments while reducing token consumption by up to 40%. In terms of performance, SkillRT achieves up to 3.2x speedup with enhanced parallelism, and 19-50x latency reduction through code solidification.

This work establishes a rigorous bridge between infinite-dimensional delay dynamics and finite-dimensional Koopman learning, with explicit and interpretable error guarantees. While Koopman analysis is well-developed for ordinary differential equations (ODEs) and partially for partial differential equations (PDEs), its extension to delay differential equations (DDEs) remains limited due to the infinite-dimensional phase space of DDEs. We propose a finite-dimensional Koopman approximation framework based on history discretization and a suitable reconstruction operator, enabling a tractable representation of the Koopman operator via kernel-based extended dynamic mode decomposition (kEDMD). Deterministic error bounds are derived for the learned predictor, decomposing the total error into contributions from history discretization, kernel interpolation, and data-driven regression. Additionally, we develop a kernel-based reconstruction method to recover discretized states from lifted Koopman coordinates, with provable guarantees. Numerical results demonstrate convergence of the learned predictor with respect to both discretization resolution and training data, supporting reliable prediction and control of delay systems.

Reasoning about consistency models for replicated data systems is a challenging task that requires a deep understanding of both the consistency models themselves and a large part of human inputs in mechanized verification approaches. In this work, we introduce an approach to reasoning about consistency models for replicated data systems. We introduce HistMSO, a monadic second-order logic (MSO) for histories and abstract executions, the formal models of executions of replicated data systems introduced by Burckhardt. We show that HistMSO can express 39 out of 42 consistency models from Viotti and Vukolic hierarchy. Moreover, we develop a method for reducing HistMSO satisfiability and model-checking to the same problems for MSO over words. While doing this, we leverage the MONA tool for automated reasoning on consistency models.

LLM-based coding agents extend their capabilities via third-party agent skills distributed through open marketplaces without mandatory security review. Unlike traditional packages, these skills are executed as operational directives with system-level privileges, so a single malicious skill can compromise the host. Prior work has not examined whether supply-chain attacks can directly hijack an agent's action space, such as file writes, shell commands, and network requests, despite existing safeguards. We introduce Document-Driven Implicit Payload Execution (DDIPE), which embeds malicious logic in code examples and configuration templates within skill documentation. Because agents reuse these examples during normal tasks, the payload executes without explicit prompts. Using an LLM-driven pipeline, we generate 1,070 adversarial skills from 81 seeds across 15 MITRE ATTACK categories. Across four frameworks and five models, DDIPE achieves 11.6% to 33.5% bypass rates, while explicit instruction attacks achieve 0% under strong defenses. Static analysis detects most cases, but 2.5% evade both detection and alignment. Responsible disclosure led to four confirmed vulnerabilities and two fixes.

As modern analogue/mixed-signal design increasingly relies on optimization-in-the-loop flows, such as AI and LLM-based sizing agents that repeatedly invoke SPICE-efficient, accurate high-performance simulators have become an indispensable foundation for modern integrated circuit (IC) design. However, the computational cost of evaluating nonlinear models, particularly for BSIM models, remains a significant bottleneck. In standard parallelization approaches, devices such as transistors are easily distributed across processors. The subsequent stamping phase, where each device's contributions are added to the shared system matrix, often creates a bottleneck. Because multiple processor cores compete to update the same matrix elements simultaneously, the system is forced to process tasks one at a time to avoid errors. This paper introduces EEspice, an open-source circuit simulation framework whose modular architecture decouples device model evaluation into independently replaceable kernels, enabling a parallel stamping strategy that overcomes this bottleneck. It partitions MOSFET instances into independent color groups, which can be processed in parallel. Our results show that on a 64-core workstation, the proposed approach achieves up to 45x speedup over single-thread performance when conflicts are low. Our analysis also explores how performance depends on circuit topology.

Transcribing and understanding multi-speaker conversations requires speech recognition, speaker attribution, and timestamp localization. While speech LLMs excel at single-speaker tasks, multi-speaker scenarios remain challenging due to overlapping speech, backchannels, rapid turn-taking, and context window constraints. We propose Speaker-Reasoner, an end-to-end Speech LLM with agentic multi-turn temporal reasoning. Instead of single-pass inference, the model iteratively analyzes global audio structure, autonomously predicts temporal boundaries, and performs fine-grained segment analysis, jointly modeling speaker identity, gender, timestamps, and transcription. A speaker-aware cache further extends processing to audio exceeding the training context window. Trained with a three-stage progressive strategy, Speaker-Reasoner achieves consistent improvements over strong baselines on AliMeeting and AISHELL-4 datasets, particularly in handling overlapping speech and complex turn-taking.

We present a case study where an automatic AI system formalizes a textbook with more than 500 pages of graduate-level algebraic combinatorics to Lean. The resulting formalization represents a new milestone in textbook formalization scale and proficiency, moving from early results in undergraduate topology and restructuring of existing library content to a full standalone formalization of a graduate textbook. The formalization comprises 130K lines of code and 5900 Lean declarations and was conducted within one week by a total of 30K Claude 4.5 Opus agents collaborating in parallel on a shared code base via version control, simultaneously setting a record in multi-agent software engineering with usable results. The inference cost matches or undercuts what we estimate as the salaries required for a team of human experts, and we expect there is still the potential for large efficiencies to be made without the need for better models. We make our code, the resulting Lean code base and a side-by-side blueprint website available open-source.

Third-party skills extend LLM agents with powerful capabilities but often handle sensitive credentials in privileged environments, making leakage risks poorly understood. We present the first large-scale empirical study of this problem, analyzing 17,022 skills (sampled from 170,226 on SkillsMP) using static analysis, sandbox testing, and manual inspection. We identify 520 vulnerable skills with 1,708 issues and derive a taxonomy of 10 leakage patterns (4 accidental and 6 adversarial). We find that (1) leakage is fundamentally cross-modal: 76.3% require joint analysis of code and natural language, while 3.1% arise purely from prompt injection; (2) debug logging is the primary vector, with print and console.log causing 73.5% of leaks due to stdout exposure to LLMs; and (3) leaked credentials are both exploitable (89.6% without privileges) and persistent, as forks retain secrets even after upstream fixes. After disclosure, all malicious skills were removed and 91.6% of hardcoded credentials were fixed. We release our dataset, taxonomy, and detection pipeline to support future research.

LLMs can be socially sycophantic, affirming users when they ask questions like "am I in the wrong?" rather than providing genuine assessment. We hypothesize that this behavior arises from incorrect assumptions about the user, like underestimating how often users are seeking information over reassurance. We present Verbalized Assumptions, a framework for eliciting these assumptions from LLMs. Verbalized Assumptions provide insight into LLM sycophancy, delusion, and other safety issues, e.g., the top bigram in LLMs' assumptions on social sycophancy datasets is ``seeking validation.'' We provide evidence for a causal link between Verbalized Assumptions and sycophantic model behavior: our assumption probes (linear probes trained on internal representations of these assumptions) enable interpretable fine-grained steering of social sycophancy. We explore why LLMs default to sycophantic assumptions: on identical queries, people expect more objective and informative responses from AI than from other humans, but LLMs trained on human-human conversation do not account for this difference in expectations. Our work contributes a new understanding of assumptions as a mechanism for sycophancy.

This paper presents an open source methodology for allowing users to query structured non textual datasets through natural language Unlike Retrieval Augmented Generation RAG which struggles with numerical and highly structured information our approach trains an LLM to generate executable queries To support this capability we introduce a principled pipeline for synthetic training data generation producing diverse question answer pairs that capture both user intent and the semantics of the underlying dataset We fine tune a compact model DeepSeek R1 Distill 8B using QLoRA with 4 bit quantization making the system suitable for deployment on commodity hardware We evaluate our approach on a dataset describing accessibility to essential services across Durangaldea Spain The fine tuned model achieves high accuracy across monolingual multilingual and unseen location scenarios demonstrating both robust generalization and reliable query generation Our results highlight that small domain specific models can achieve high precision for this task without relying on large proprietary LLMs making this methodology suitable for resource constrained environments and adaptable to broader multi dataset systems We evaluate our approach on a dataset describing accessibility to essential services across Durangaldea Spain The fine tuned model achieves high accuracy across monolingual multilingual and unseen location scenarios demonstrating both robust generalization and reliable query generation Our results highlight that small domain specific models can achieve high precision for this task without relying on large proprietary LLMs making this methodology suitable for resource constrained environments and adaptable to broader multi dataset systems.

In this paper, we study a network formation game in which agents seek to maximize their influence by allocating constrained resources to choose connections with other agents. In particular, we use Katz centrality to model agents' influence in the network. Allocations are restricted to neighbors in a given unweighted network encoding topological constraints. The allocations by an agent correspond to the weights of its outgoing edges. Such allocation by all agents thereby induces a network. This models a strategic-form game in which agents' utilities are given by their Katz centralities. We characterize the Nash equilibrium networks of this game and analyze their properties. We propose a sequential best-response dynamics (BRD) to model the network formation process. We show that it converges to the set of Nash equilibria under very mild assumptions. For complete underlying topologies, we show that Katz centralities are proportional to agents' budgets at Nash equilibria. For general underlying topologies in which each agent has a self-loop, we show that hierarchical networks form at Nash equilibria. Finally, simulations illustrate our findings.

These proceedings contain the papers that were presented at the 7th Workshop on Models for Formal Analysis of Real Systems (MARS 2026), which took place on 12 April 2026 in Turin, Italy, as a satellite event of the 29th International Joint Conferences on Theory and Practice of Software (ETAPS 2026). The goal of MARS is to bring together researchers from different communities who are developing formal models of real systems in areas where complex models occur (e.g., networks, cyber-physical systems, hardware/software codesign, biology). The motivation for MARS stems from the following two observations: - Large case studies are essential to show that specification formalisms and modelling techniques are applicable to real systems, whereas many papers only consider toy examples or tiny case studies. - Developing an accurate model of a real system takes a large amount of time, often months or years. In most papers, however, salient details of the model need to be skipped due to lack of space, and to leave room for formal verification methodologies and results. MARS aims at remedying these issues, emphasising modelling over verification, so as to retain lessons learned from formal modelling, which are not usually discussed elsewhere, and which may lay the basis for future analysis and comparison.

While affective computing has advanced considerably, multimodal emotion prediction in aging populations remains underexplored, largely due to the scarcity of dedicated datasets. Existing multimodal benchmarks predominantly target young, cognitively healthy subjects, neglecting the influence of cognitive decline on emotional expression and physiological responses. To bridge this gap, we present MECO, a Multimodal dataset for Emotion and Cognitive understanding in Older adults. MECO includes 42 participants and provides approximately 38 hours of multimodal signals, yielding 30,592 synchronized samples. To maximize ecological validity, data collection followed standardized protocols within community-based settings. The modalities cover video, audio, electroencephalography (EEG), and electrocardiography (ECG). In addition, the dataset offers comprehensive annotations of emotional and cognitive states, including self-assessed valence, arousal, six basic emotions, and Mini-Mental State Examination cognitive scores. We further establish baseline benchmarks for both emotion and cognitive prediction. MECO serves as a foundational resource for multimodal modeling of affect and cognition in aging populations, facilitating downstream applications such as personalized emotion recognition and early detection of mild cognitive impairment (MCI) in real-world settings. The complete dataset and supplementary materials are available at https://maitrechen.github.io/meco-page/.

Context: Since it is well-established that developers spend a substantial portion of their time understanding source code, the ability to automatically identify algorithms within source code presents a valuable opportunity. This capability can support program comprehension, facilitate maintenance, and enhance overall software quality. Objective: We empirically evaluate how combining LLMs with static code analysis can improve the automated recognition of algorithms, while also evaluating their standalone performance and dependence on identifier names. Method: We perform multiple experiments evaluating the combination of LLMs with static analysis using different filter patterns. We compare this combined approach against their standalone performance under various prompting strategies and investigate the impact of systematic identifier obfuscation on classification performance and runtime. Results: The combination of LLMs with lightweight static analysis performs surprisingly well, reducing required LLM calls by 72.39-97.50% depending on the filter pattern. This not only lowers runtime significantly but also improves F1-scores by up to 12 percentage points (pp) compared to the baseline. Regarding the different prompting strategies, in-context learning with two examples provides an effective trade-off between classification performance and runtime efficiency, achieving F1-scores of 75-77% with only a modest increase in inference time. Lastly, we find that LLMs are not solely dependent on name-information as they are still able to identify most algorithm implementations when identifiers are obfuscated. Conclusion: By combining LLMs with static analysis, we achieve substantial reductions in runtime while simultaneously improving F1-scores, underscoring the value of a hybrid approach.

We introduce JoyAI-LLM Flash, an efficient Mixture-of-Experts (MoE) language model designed to redefine the trade-off between strong performance and token efficiency in the sub-50B parameter regime. JoyAI-LLM Flash is pretrained on a massive corpus of 20 trillion tokens and further optimized through a rigorous post-training pipeline, including supervised fine-tuning (SFT), Direct Preference Optimization (DPO), and large-scale reinforcement learning (RL) across diverse environments. To improve token efficiency, JoyAI-LLM Flash strategically balances \emph{thinking} and \emph{non-thinking} cognitive modes and introduces FiberPO, a novel RL algorithm inspired by fibration theory that decomposes trust-region maintenance into global and local components, providing unified multi-scale stability control for LLM policy optimization. To enhance architectural sparsity, the model comprises 48B total parameters while activating only 2.7B parameters per forward pass, achieving a substantially higher sparsity ratio than contemporary industry leading models of comparable scale. To further improve inference throughput, we adopt a joint training-inference co-design that incorporates dense Multi-Token Prediction (MTP) and Quantization-Aware Training (QAT). We release the checkpoints for both JoyAI-LLM-48B-A3B Base and its post-trained variants on Hugging Face to support the open-source community.

In a healthcare environment, the healthcare interoperability platforms based on HL7 FHIR allow concurrent, asynchronous access to a set of shared patient resources, which are independent systems, i.e., EHR systems, pharmacy systems, lab systems, and devices. The FHIR specification lacks a protocol for concurrency control, and the research on detecting a race condition only targets the OS kernel. The research on FHIR security only targets authentication and injection attacks, considering concurrent access to patient resources to be sequential. The gap in the research in this area is addressed through the introduction of FHIR Resource Access Graph (FRAG), a formally defined graph G = (P,R,E, λ, τ, S), in which the nodes are the concurrent processes, the typed edges represent the resource access events, and the race conditions are represented as detectable structural properties. Three clinically relevant race condition classes are formally specified: Simultaneous Write Conflict (SWC), TOCTOU Authorization Violation (TAV), and Cascading Update Race (CUR). The FRAG model is implemented as a three-pass graph traversal detection algorithm and tested against a time window-based baseline on 1,500 synthetic FHIR R4 transaction logs. Under full concurrent access (C2), FRAG attains a 90.0% F1 score vs. 25.5% for the baseline, a 64.5 pp improvement.

Long-horizon robotic manipulation remains challenging for reinforcement learning (RL) because sparse rewards provide limited guidance for credit assignment. Practical policy improvement thus relies on richer intermediate supervision, such as dense progress rewards, which are costly to obtain and ill-suited to non-monotonic behaviors such as backtracking and recovery. To address this, we propose Advantage Reward Modeling (ARM), a framework that shifts from hard-to-quantify absolute progress to estimating relative advantage. We introduce a cost-effective tri-state labeling strategy -- Progressive, Regressive, and Stagnant -- that reduces human cognitive overhead while ensuring high cross-annotator consistency. By training on these intuitive signals, ARM enables automated progress annotation for both complete demonstrations and fragmented DAgger-style data. Integrating ARM into an offline RL pipeline allows for adaptive action-reward reweighting, effectively filtering suboptimal samples. Our approach achieves a 99.4% success rate on a challenging long-horizon towel-folding task, demonstrating improved stability and data efficiency over current VLA baselines with near-zero human intervention during policy training.

Existing datasets for coding agents evaluate performance on isolated, single pull request (PR) tasks in a stateless manner, failing to capture the reality of real-world software development where code changes accumulate, technical debt accrues, and test suites grow over time. To bridge this gap, we introduce an automated coding task generation framework, which helps generate our dataset SWE-STEPS, that evaluates coding agents on long-horizon tasks through two realistic settings mirroring actual developer workflows: Conversational coding with iterative requests, and single-shot Project Requirement document (PRD)-based coding. Unlike existing datasets that evaluate agents on disjointed Pull Requests (PRs), our framework assesses performance across chains of dependent PRs, enabling evaluation of sequential execution, regression verification, and long-term repository health. We discover that widely used isolated PR evaluations yield inflated success rates, w.r.t. our settings - overshooting performance by as much as 20 percentage points - because they ignore the ``spillover'' effects of previous inefficient or buggy code. Furthermore, our analysis reveals that even when agents successfully resolve issues, they degrade repository health by generating code with higher cognitive complexity and technical debt compared to human developers, underscoring the necessity for multidimensional evaluation.

We generalize the framework of Fredholm Neural Networks, to learn non-expansive integral operators arising in Fredholm Integral Equations (FIEs) of the second kind in arbitrary dimensions. We first present the proposed Fredholm Integral Neural Operators (FREDINOs), for FIEs and prove that they are universal approximators of linear and non-linear integral operators and corresponding solution operators. We furthermore prove that the learned operators are guaranteed to be contractive, thereby strictly satisfying the mathematical property required for the convergence of the fixed point scheme. Finally, we also demonstrate how FREDINOs can be used to learn the solution operator of non-linear elliptic PDEs, via a Boundary Integral Equation (BIE) formulation. We assess the proposed methodology numerically, via several benchmark problems: linear and non-linear FIEs in arbitrary dimensions, as well as a non-linear elliptic PDE in 2D. Built on tailored mathematical/numerical analysis theory, FREDINOs offer high-accuracy approximations and interpretable schemes, making them well suited for scientific machine learning/numerical analysis computations.

DBMSs are complex systems prone to bugs that may lead to system failures or compromise data integrity. Establishing unified DBMS bug repositories is crucial for systematically organizing bug-related data, enabling code improvement, and supporting automated testing. In particular, bug reports often contain valuable test inputs and bug-triggering clues that help explore rare execution paths and expose critical buggy behavior, thereby guiding automated DBMS testing. However, the heterogeneity of bug reports, along with their incomplete or inaccurate content, makes it challenging to build unified repositories and convert them into high-quality test cases. In this paper, we propose BugForge, a framework that constructs standardized DBMS bug repositories and leverages them to generate high-quality test cases to enhance DBMS testing. Specifically, BugForge progressively collects bug reports, then employs syntax-aware processing and input-adaptive raw PoC extraction to construct a DBMS bug repository. The repository stores structured bug-related data, including bug metadata and raw PoCs that entail potential bug-triggering semantics. These data are further refined into high-quality test cases through semantic-guided adaptation, thereby enabling enhanced DBMS testing methods, including DBMS fuzzing, regression testing, and cross-DBMS bug discovery. We implemented BugForge for PostgreSQL, MySQL, MariaDB, and MonetDB, totally integrated 37,632 bug reports spanning up to 28 years. Based on the repository, BugForge uncovered 35 previously unknown bugs with 22 confirmed by developers, demonstrating the value of constructing and utilizing bug repositories for DBMS testing.

Problem solving plays an essential role in science education, and generative AI (GAI) chatbots have emerged as a promising tool for supporting students' science problem solving. However, general-purpose chatbots (e.g., ChatGPT), which often provide direct, ready-made answers, may lead to students' cognitive offloading. Prior research has rarely focused on custom chatbots for facilitating students' science problem solving, nor has it examined how they differently influence problem-solving processes and performance compared to general-purpose chatbots. To address this gap, we developed a pedagogy-informed custom GAI chatbot grounded in the Socratic questioning method, which supports students by prompting them with guiding questions. This study employed a within-subjects counterbalanced design in which 48 secondary school students used both custom and general-purpose chatbot to complete two science problem-solving tasks. 3297 student-chatbot dialogues were collected and analyzed using Heterogeneous Interaction Network Analysis (HINA). The results showed that: (1) students demonstrated significantly higher interaction intensity and cognitive interaction diversity when using custom chatbot than using general-purpose chatbot; (2) students were more likely to follow custom chatbot's guidance to think and reflect, whereas they tended to request general-purpose chatbot to execute specific commands; and (3) no statistically significant difference was observed in students' problem-solving performance evaluated by solution quality between two chatbot conditions. This study provides novel theoretical insights and empirical evidence that custom chatbots are less likely to induce cognitive offloading and instead foster greater cognitive engagement compared to general-purpose chatbots. This study also offers insights into the design and integration of GAI chatbots in science education.

Multimodal Large Language Models (MLLMs) are evolving from passive observers into active agents, solving problems through Visual Expansion (invoking visual tools) and Knowledge Expansion (open-web search). However, existing evaluations fall short: they lack flexible tool integration, test visual and search tools separately, and evaluate primarily by final answers. Consequently, they cannot verify if tools were actually invoked, applied correctly, or used efficiently. To address this, we introduce Agentic-MME, a process-verified benchmark for Multimodal Agentic Capabilities. It contains 418 real-world tasks across 6 domains and 3 difficulty levels to evaluate capability synergy, featuring over 2,000 stepwise checkpoints that average 10+ person-hours of manual annotation per task. Each task includes a unified evaluation framework supporting sandboxed code and APIs, alongside a human reference trajectory annotated with stepwise checkpoints along dual-axis: S-axis and V-axis. To enable true process-level verification, we audit fine-grained intermediate states rather than just final answers, and quantify efficiency via an overthinking metric relative to human trajectories. Experimental results show the best model, Gemini3-pro, achieves 56.3% overall accuracy, which falls significantly to 23.0% on Level-3 tasks, underscoring the difficulty of real-world multimodal agentic problem solving.

Given $n$ independent samples from a $d$-dimensional probability distribution, our aim is to generate diffusion-based samples from a distribution obtained by tilting the original, where the degree of tilt is parametrized by $θ\in \mathbb{R}^d$. We define a plug-in estimator and show that it is minimax-optimal. We develop Wasserstein bounds between the distribution of the plug-in estimator and the true distribution as a function of $n$ and $θ$, illustrating regimes where the output and the desired true distribution are close. Further, under some assumptions, we prove the TV-accuracy of running Diffusion on these tilted samples. Our theoretical results are supported by extensive simulations. Applications of our work include finance, weather and climate modelling, and many other domains, where the aim may be to generate samples from a tilted distribution that satisfies practically motivated moment constraints.

Multi-modal recommendation (MMR) enriches item representations by introducing item content, e.g., visual and textual descriptions, to improve upon interaction-only recommenders. The success of MMR hinges on aligning these content modalities with user preferences derived from interaction data, yet dominant practices based on disentangling modality-invariant preference-driving signals from modality-specific preference-irrelevant noises are flawed. First, they assume a one-size-fits-all relevance of item content to user preferences for all users, which contradicts the user-conditional fact of preferences. Second, they optimize pairwise contrastive losses separately toward cross-modal alignment, systematically ignoring higher-order dependencies inherent when multiple content modalities jointly influence user choices. In this paper, we introduce GTC, a conditional Generative Total Correlation learning framework. We employ an interaction-guided diffusion model to perform user-aware content feature filtering, preserving only personalized features relevant to each individual user. Furthermore, to capture complete cross-modal dependencies, we optimize a tractable lower bound of the total correlation of item representations across all modalities. Experiments on standard MMR benchmarks show GTC consistently outperforms state-of-the-art, with gains of up to 28.30% in NDCG@5. Ablation studies validate both conditional preference-driven feature filtering and total correlation optimization, confirming the ability of GTC to model user-conditional relationships in MMR tasks. The code is available at: https://github.com/jingdu-cs/GTC.

Memory-disaggregated key-value (KV) stores suffer from a severe performance bottleneck due to their I/O redundancy issues. A huge amount of redundant I/Os are generated when synchronizing concurrent data accesses, making the limited network between the compute and memory pools of DM a performance bottleneck. We identify the root cause for the redundant I/O lies in the mismatch between the optimistic synchronization of existing memory-disaggregated KV stores and the highly concurrent workloads on DM. In this paper, we propose to boost memory-disaggregated KV stores with pessimistic synchronization. We propose CIDER, a compute-side I/O optimization framework, to verify our idea. CIDER adopts a global write-combining technique to further reduce cross-node redundant I/Os. A contention-aware synchronization scheme is designed to improve the performance of pessimistic synchronization under low contention scenarios. Experimental results show that CIDER effectively improves the throughput of state-of-the-art memory-disaggregated KV stores by up to $6.6\times$ under the YCSB benchmark.

While deep reasoning with long chain-of-thought has dramatically improved large language models in verifiable domains like mathematics, its effectiveness for open-ended tasks such as writing remains unexplored. In this paper, we conduct a systematic investigation revealing that existing mainstream reasoning models achieve limited gains on open-ended writing tasks. Our further analysis shows that these models lack deep reflection and revision patterns in open-ended writing, resulting in substantially smaller improvements compared to mathematical reasoning tasks. To address this limitation, we introduce R2-Write: an automated framework that synthesizes high-quality thinking trajectories enriched with explicit reflection and revision patterns through iterative writer-judge interaction. To prevent redundant reflections, we design a process reward mechanism that supervises reflection quality during reinforcement learning, improving both performance and token efficiency. Extensive experiments across multiple creative writing and deep-research benchmarks demonstrate significant improvements, validating that explicitly incorporating reflection and revision patterns unlocks deep reasoning capabilities for open-ended writing tasks.

We introduce a nonnegative functional on the space of line arrangements in $\mathbb{P}^2$ that vanishes precisely on free arrangements, obtained as a semicontinuous relaxation of Saito's criterion for freeness. Given an arrangement $\mathcal{A}$ of $n$ lines with candidate exponents $(d_1, d_2)$, we parameterize the spaces of logarithmic derivations of degrees $d_1$ and $d_2$ via the null spaces of the associated derivation matrices and express the Saito determinant as a bilinear map into the space of degree $n$ polynomials. The functional then admits a natural geometric interpretation: it measures the squared sine of the angle between the image of this bilinear map and the direction of the defining polynomial $Q(\mathcal{A})$ in coefficient space, and equals zero if and only if its image contains the line spanned by $Q(\mathcal{A})$. This provides a computable measure of how far a given arrangement is from admitting a free basis of logarithmic derivations of the expected degrees. Using this functional as a reward signal, we develop a sequential construction procedure in which lines are added one at a time so as to minimize the angular distance to freeness, implemented via reinforcement learning with an adaptive curriculum over arrangement sizes and exponent types. Our results suggest that semicontinuous relaxation techniques, grounded in the geometry of polynomial coefficient spaces, offer a viable approach to the computational exploration of freeness in the theory of line arrangements.

Federated learning (FL) enables collaborative training across organizations without sharing raw data, but it is hindered by statistical heterogeneity (non-i.i.d.\ client data) and by instability of naive weight averaging under client drift. In many cross-silo deployments, FL is warm-started from a strong pretrained backbone (e.g., ImageNet-1K) and then adapted to local domains. Motivated by recent evidence that ReLU-like gating regimes (structural knowledge) stabilize earlier than the remaining parameter values (quantitative knowledge), we propose FedSQ (Federated Structural-Quantitative learning), a transfer-initialized neural federated procedure based on a DualCopy, piecewise-linear view of deep networks. FedSQ freezes a structural copy of the pretrained model to induce fixed binary gating masks during federated fine-tuning, while only a quantitative copy is optimized locally and aggregated across rounds. Fixing the gating reduces learning to within-regime affine refinements, which stabilizes aggregation under heterogeneous partitions. Experiments on two convolutional neural network backbones under i.i.d.\ and Dirichlet splits show that FedSQ improves robustness and can reduce rounds-to-best validation performance relative to standard baselines while preserving accuracy in the transfer setting.

Given a user's complex information need, a multi-agent Deep Research system iteratively plans, retrieves, and synthesizes evidence across hundreds of documents to produce a high-quality answer. In one possible architecture, an orchestrator agent coordinates the process, while parallel worker agents execute tasks. Current Deep Research systems, however, often rely on hand-engineered prompts and static architectures, making improvement brittle, expensive, and time-consuming. We therefore explore various multi-agent optimization methods to show that enabling agents to self-play and explore different prompt combinations can produce high-quality Deep Research systems that match or outperform expert-crafted prompts.

Reward models (RMs) used in reinforcement learning from human feedback (RLHF) are vulnerable to reward hacking: as the policy maximizes a learned proxy reward, true quality plateaus or degrades. We make the assumption that reward hacking is often caused by flipped advantage signs: instead of reducing the likelihood of a bad response, a flipped sign causes the update to increase it. By considering an adversarial perturbation in the RM parameter space, we can derive a certified sign-preservation radius, which is the smallest perturbation that can flip the advantage sign during policy optimization. Based on this formulation, we propose Sign-Certified Policy Optimization (SignCert-PO), down-weighting non-robust completions in the policy gradient update. Unlike prior approaches that require multiple RMs or access to the RM training data, SignCert-PO is lightweight and operates purely at the policy optimization stage using only the RM parameters and on-policy completions. On TL;DR summarization and AlpacaFarm benchmarks, SignCert-PO consistently achieves a better win rate than baselines and reduces reward hacking.

With the wide adoption of language models for IR -- and specifically RAG systems -- the latency of the underlying LLM becomes a crucial bottleneck, since the long contexts of retrieved passages lead large prompts and therefore, compute increase. Prompt compression, which reduces the size of input prompts while aiming to preserve performance on downstream tasks, has established itself as a cost-effective and low-latency method for accelerating inference in large language models. However, its usefulness depends on whether the additional preprocessing time during generation is offset by faster decoding. We present the first systematic, large-scale study of this trade-off, with thousands of runs and 30,000 queries across several open-source LLMs and three GPU classes. Our evaluation separates compression overhead from decoding latency while tracking output quality and memory usage. LLMLingua achieves up to 18% end-to-end speed-ups, when prompt length, compression ratio, and hardware capacity are well matched, with response quality remaining statistically unchanged across summarization, code generation, and question answering tasks. Outside this operating window, however, the compression step dominates and cancels out the gains. We also show that effective compression can reduce memory usage enough to offload workloads from data center GPUs to commodity cards, with only a 0.3s increase in latency. Our open-source profiler predicts the latency break-even point for each model-hardware setup, providing practical guidance on when prompt compression delivers real-world benefits.

Large Reasoning Models (LRMs) have recently achieved remarkable success in complex reasoning tasks. However, closer scrutiny reveals persistent failure modes compromising performance and cost: I) Intra-step level, marked by calculation or derivation errors; II) Inter-step level, involving oscillation and stagnation; and III) Instance level, causing maladaptive over-thinking. Existing endeavors target isolated levels without unification, while their black-box nature and reliance on RL hinder explainability and controllability. To bridge these gaps, we conduct an in-depth white-box analysis, identifying key neurons (Mixture of Neurons, MoN) and their fluctuation patterns associated with distinct failures. Building upon these insights, we propose NeuReasoner, an explainable, controllable, and unified reasoning framework driven by MoN. Technically, NeuReasoner integrates lightweight MLPs for failure detection with a special token-triggered self-correction mechanism learned via SFT. During inference, special tokens are inserted upon failure detection to actuate controllable remedial behaviors. Extensive evaluations across six benchmarks, six backbone models (8B~70B) against nine competitive baselines, demonstrate that NeuReasoner achieves performance gains of up to 27.0% while reducing token consumption by 19.6% ~ 63.3%.

Recent agentic search systems have made substantial progress by emphasising deep, multi-step reasoning. However, this focus often overlooks the challenges of wide-scale information synthesis, where agents must aggregate large volumes of heterogeneous evidence across many sources. As a result, most existing large language model agent systems face severe limitations in data-intensive settings, including context saturation, cascading error propagation, and high end-to-end latency. To address these challenges, we present \framework, a hierarchical framework based on principle of near-decomposability, containing a strategic \textit{Host}, multiple \textit{Managers} and parallel \textit{Workers}. By leveraging aggregation and reflection mechanisms at the Manager layer, our framework enforces strict context isolation to prevent saturation and error propagation. Simultaneously, the parallelism in worker layer accelerates the speed of overall task execution, mitigating the significant latency. Our evaluation on two complementary benchmarks demonstrates both efficiency ($ 3-5 \times$ speed-up) and effectiveness, achieving a $8.4\%$ success rate on WideSearch-en and $52.9\%$ accuracy on BrowseComp-zh. The code is released at https://github.com/agent-on-the-fly/InfoSeeker

The natural gradient method is widely used in statistical optimization, but its standard formulation assumes a Euclidean parameter space. This paper proposes an inversion-free stochastic natural gradient method for probability distributions whose parameters lie on a Riemannian manifold. The manifold setting offers several advantages: one can implicitly enforce parameter constraints such as positive definiteness and orthogonality, ensure parameters are identifiable, or guarantee regularity properties of the objective like geodesic convexity. Building on an intrinsic formulation of the Fisher information matrix (FIM) on a manifold, our method maintains an online approximation of the inverse FIM, which is efficiently updated at quadratic cost using score vectors sampled at successive iterates. In the Riemannian setting, these score vectors belong to different tangent spaces and must be combined using transport operations. We prove almost-sure convergence rates of $O(\log{s}/s^α)$ for the squared distance to the minimizer when the step size exponent $α>2/3$. We also establish almost-sure rates for the approximate FIM, which now accumulates transport-based errors. A limited-memory variant of the algorithm with sub-quadratic storage complexity is proposed. Finally, we demonstrate the effectiveness of our method relative to its Euclidean counterparts on variational Bayes with Gaussian approximations and normalizing flows.

Recent Large Reasoning Models (LRMs) like DeepSeek-R1 have demonstrated remarkable success in complex reasoning tasks, exhibiting human-like patterns in exploring multiple alternative solutions. Upon closer inspection, however, we uncover a surprising phenomenon: The First is The Best, where alternative solutions are not merely suboptimal but potentially detrimental. This observation challenges widely accepted test-time scaling laws, leading us to hypothesize that errors within the reasoning path scale concurrently with test time. Through comprehensive empirical analysis, we characterize errors as a forest-structured Forest of Errors (FoE) and conclude that FoE makes the First the Best, which is underpinned by rigorous theoretical analysis. Leveraging these insights, we propose RED, a self-guided efficient reasoning framework comprising two components: I) Refining First, which suppresses FoE growth in the first solution; and II) Discarding Subs, which prunes subsequent FoE via dual-consistency. Extensive experiments across five benchmarks and six backbone models demonstrate that RED outperforms eight competitive baselines, achieving performance gains of up to 19.0% while reducing token consumption by 37.7% ~ 70.4%. Moreover, comparative experiments on FoE metrics shed light on how RED achieves effectiveness.

Vision-Language-Action (VLA) models, as large foundation models for embodied control, have shown strong performance in manipulation tasks. However, their performance comes at high inference cost. To improve efficiency, recent methods adopt action chunking, which predicts a sequence of future actions for open-loop execution. Although effective for reducing computation, open-loop execution is sensitive to environmental changes and prone to error accumulation due to the lack of close-loop feedback. To address this limitation, we propose Speculative Verification for VLA Control (SV-VLA), a framework that combines efficient open-loop long-horizon planning with lightweight closed-loop online verification. Specifically, SV-VLA uses a heavy VLA as a low-frequency macro-planner to generate an action chunk together with a planning context, while a lightweight verifier continuously monitors execution based on the latest observations. Conditioned on both the current observation and the planning context, the verifier compares the planned action against a closed-loop reference action and triggers replanning only when necessary. Experiments demonstrate that SV-VLA combines the efficiency of chunked prediction with the robustness of closed-loop control, enabling efficient and reliable VLA-based control in dynamic environments. Code is available: https://github.com/edsad122/SV-VLA.

Graph-based Retrieval-Augmented Generation (GraphRAG) enhances the reasoning capabilities of Large Language Models (LLMs) by grounding their responses in structured knowledge graphs. Leveraging community detection and relation filtering techniques, GraphRAG systems demonstrate inherent resistance to traditional RAG attacks, such as text poisoning and prompt injection. However, in this paper, we find that the security of GraphRAG systems fundamentally relies on the topological integrity of the underlying graph, which can be undermined by implicitly corrupting the logical connections, without altering surface-level text semantics. To exploit this vulnerability, we propose \textsc{LogicPoison}, a novel attack framework that targets logical reasoning rather than injecting false contents. Specifically, \textsc{LogicPoison} employs a type-preserving entity swapping mechanism to perturb both global logic hubs for disrupting overall graph connectivity and query-specific reasoning bridges for severing essential multi-hop inference paths. This approach effectively reroutes valid reasoning into dead ends while maintaining surface-level textual plausibility. Comprehensive experiments across multiple benchmarks demonstrate that \textsc{LogicPoison} successfully bypasses GraphRAG's defenses, significantly degrading performance and outperforming state-of-the-art baselines in both effectiveness and stealth. Our code is available at \textcolor{blue}https://github.com/Jord8061/logicPoison.

Human data annotation, especially when involving experts, is often treated as an objective reference. However, many annotation tasks are inherently subjective, and annotators' judgments may evolve over time. This study investigates changes in the quality of annotators' work from a competence perspective during a process of social influence recognition. The study involved 25 annotators from five different groups, including both experts and non-experts, who annotated a dataset of 1,021 dialogues with 20 social influence techniques, along with intentions, reactions, and consequences. An initial subset of 150 texts was annotated twice - before and after the main annotation process - to enable comparison. To measure competence shifts, we combined qualitative and quantitative analyses of the annotated data, semi-structured interviews with annotators, self-assessment surveys, and Large Language Model training and evaluation on the comparison dataset. The results indicate a significant increase in annotators' self-perceived competence and confidence. Moreover, observed changes in data quality suggest that the annotation process may enhance annotator competence and that this effect is more pronounced in expert groups. The observed shifts in annotator competence have a visible impact on the performance of LLMs trained on their annotated data.

Computer-use agents extend language models from text generation to persistent action over tools, files, and execution environments. Unlike chat systems, they maintain state across interactions and translate intermediate outputs into concrete actions. This creates a distinct safety challenge in that harmful behavior may emerge through sequences of individually plausible steps, including intermediate actions that appear locally acceptable but collectively lead to unauthorized actions. We present \textbf{AgentHazard}, a benchmark for evaluating harmful behavior in computer-use agents. AgentHazard contains \textbf{2,653} instances spanning diverse risk categories and attack strategies. Each instance pairs a harmful objective with a sequence of operational steps that are locally legitimate but jointly induce unsafe behavior. The benchmark evaluates whether agents can recognize and interrupt harm arising from accumulated context, repeated tool use, intermediate actions, and dependencies across steps. We evaluate AgentHazard on Claude Code, OpenClaw, and IFlow using mostly open or openly deployable models from the Qwen3, Kimi, GLM, and DeepSeek families. Our experimental results indicate that current systems remain highly vulnerable. In particular, when powered by Qwen3-Coder, Claude Code exhibits an attack success rate of \textbf{73.63\%}, suggesting that model alignment alone does not reliably guarantee the safety of autonomous agents.

Learning methods using synthetic data have attracted attention as an effective approach for increasing the diversity of training data while reducing collection costs, thereby improving the robustness of model discrimination. However, many existing methods improve robustness only indirectly through the diversification of training samples and do not explicitly teach the model which regions in the input space truly contribute to discrimination; consequently, the model may learn spurious correlations caused by synthesis biases and artifacts. Motivated by this limitation, this paper proposes a learning framework that uses provenance information obtained during the training data synthesis process, indicating whether each region in the input space originates from the target object, as an auxiliary supervisory signal to promote the acquisition of representations focused on target regions. Specifically, input gradients are decomposed based on information about target and non-target regions during synthesis, and input gradient guidance is introduced to suppress gradients over non-target regions. This suppresses the model's reliance on non-target regions and directly promotes the learning of discriminative representations for target regions. Experiments demonstrate the effectiveness and generality of the proposed method across multiple tasks and modalities, including weakly supervised object localization, spatio-temporal action localization, and image classification.

Multimodal large language models (MLLMs) enable powerful cross-modal reasoning capabilities but impose substantial computational and latency burdens, posing critical challenges for deployment on resource-constrained edge devices. In this paper, we propose MSAO, an adaptive modality sparsity-aware offloading framework with edge-cloud collaboration for efficient MLLM Inference. First, a lightweight heterogeneous modality-aware via fine-grained sparsity module performs spatial-temporal-modal joint analysis to compute the Modality Activation Sparsity (MAS) metric, which quantifies the necessity of each modality with minimal computational overhead. Second, an adaptive speculative edge-cloud collaborative offloading mechanism dynamically schedules workloads between edge and cloud based on the derived MAS scores and real-time system states, leveraging confidence-guided speculative execution to hide communication latency. Extensive experiments on VQAv2 and MMBench benchmarks demonstrate that MSAO achieves a 30% reduction in end-to-end latency and 30%-65% decrease in resource overhead, while delivering a throughput improvement of 1.5x to 2.3x compared to traditional approaches, all without compromising competitive accuracy.

Microgravity induces profound metabolic adaptations in mammalian physiology, yet the molecular mechanisms governing thermogenesis in female white adipose tissue (WAT) remain poorly characterized. This paper presents the first machine learning (ML) analysis of NASA Open Science Data Repository (OSDR) dataset OSD-970, derived from the Rodent Research-1 (RR-1) mission. Using RT-qPCR data from 89 adipogenesis and thermogenesis pathway genes in gonadal WAT of 16 female C57BL/6J mice (8 flight, 8 ground control) following 37 days aboard the International Space Station (ISS), we applied differential expression analysis, multiple ML classifiers with Leave-One-Out Cross-Validation (LOO-CV), and Explainable AI via SHapley Additive exPlanations (SHAP). The most striking finding is a dramatic 12.21-fold upregulation of Ucp1 (Delta-Delta-Ct = -3.61, p = 0.0167) in microgravity-exposed WAT, accompanied by significant activation of the thermogenesis pathway (mean pathway fold-change = 3.24). The best-performing model (Random Forest with top-20 features) achieved AUC = 0.922, Accuracy = 0.812, and F1 = 0.824 via LOO-CV. SHAP analysis consistently ranked Ucp1 among the top predictive features, while Angpt2, Irs2, Jun, and Klf-family transcription factors emerged as dominant consensus classifiers. Principal component analysis (PCA) revealed clear separation between flight and ground samples, with PC1 explaining 69.1% of variance. These results suggest rapid thermogenic reprogramming in female WAT as a compensatory response to microgravity. This study demonstrates the power of explainable AI for re-analysis of newly released NASA space biology datasets, with direct implications for female astronaut health on long-duration missions and for Earth-based obesity and metabolic disease research.

Advancements in extended reality (XR) are driving the development of the metaverse, which demands efficient real-time transformation of 2D scenes into 3D objects, a computation-intensive process that necessitates task offloading because of complex perception, visual, and audio processing. This challenge is further compounded by asymmetric uplink (UL) and downlink (DL) data characteristics, where 2D data are transmitted in the UL and 3D content is rendered in the DL. To address this issue, we propose a digital twin (DT)-based in-network computing (INC)-assisted multi-access edge computing (MEC) framework that enables real-time synchronization and collaborative computing via URLLC. In this framework, a network operator manages wireless and computational resources for XR user devices (XUDs), while XUDs autonomously offload tasks to maximize their utilities. We model the interactions between XUDs and the operator as a Stackelberg Markov game, where the optimal offloading strategy constitutes an exact potential game with a Nash Equilibrium (NE), and the operator's problem is formulated as an asynchronous Markov decision process (MDP). We further propose a decentralized solution in which XUDs determine offloading decisions based on the operator's joint UL-DL optimization of offloading mode (INC-E or MEC only) and DL power allocation. A Nash-asynchronous hybrid multi-agent reinforcement learning (AMRL) algorithm is developed to predict the UL user-associated and DL transmission power, thereby achieving NE. Simulation results demonstrate that the proposed approach considerably improves system utility, uplink rate, and energy efficiency by reducing latency and optimizing resource utilization in metaverse environments.

Routing algorithms are crucial for efficient computer network operations, and in many settings they must be able to react to traffic bursts within milliseconds. Live telemetry data can provide informative signals to routing algorithms, and recent work has trained neural networks to exploit such signals for traffic-aware routing. Yet, aggregating network-wide information is subject to communication delays, and existing neural approaches either assume unrealistic delay-free global states, or restrict routers to purely local telemetry. This leaves their deployability in real-world environments unclear. We cast telemetry-aware routing as a delay-aware closed-loop control problem and introduce a framework that trains and evaluates neural routing algorithms, while explicitly modeling communication and inference delays. On top of this framework, we propose LOGGIA, a scalable graph neural routing algorithm that predicts log-space link weights from attributed topology-and-telemetry graphs. It utilizes a data-driven pre-training stage, followed by on-policy Reinforcement Learning. Across synthetic and real network topologies, and unseen mixed TCP/UDP traffic sequences, LOGGIA consistently outperforms shortest-path baselines, whereas neural baselines fail once realistic delays are enforced. Our experiments further suggest that neural routing algorithms like LOGGIA perform best when deployed fully locally, i.e., observing network states and inferring actions at every router individually, as opposed to centralized decision making.

The article proposes a new architecture based on Multi-head attention to solve the problem of morphological tagging for the Russian language. The preprocessing of the word vectors includes splitting the words into subtokens, followed by a trained procedure for aggregating the vectors of the subtokens into vectors for tokens. This allows to support an open dictionary and analyze morphological features taking into account parts of words (prefixes, endings, etc.). The open dictionary allows in future to analyze words that are absent in the training dataset. The performed computational experiment on the SinTagRus and Taiga datasets shows that for some grammatical categories the proposed architecture gives accuracy 98-99% and above, which outperforms previously known results. For nine out of ten words, the architecture precisely predicts all grammatical categories and indicates when the categories must not be analyzed for the word. At the same time, the model based on the proposed architecture can be trained on consumer-level graphics accelerators, retains all the advantages of Multi-head attention over RNNs (RNNs are not used in the proposed approach), does not require pretraining on large collections of unlabeled texts (like BERT), and shows higher processing speed than previous results.

Large Language Models (LLMs), particularly those employing Mixture-of-Experts (MoE) architectures, have achieved remarkable capabilities across diverse natural language processing tasks. However, these models frequently suffer from hallucinations -- generating plausible but factually incorrect content -- and exhibit systematic biases that are amplified by uneven expert activation during inference. In this paper, we propose the Council Mode, a novel multi-agent consensus framework that addresses these limitations by dispatching queries to multiple heterogeneous frontier LLMs in parallel and synthesizing their outputs through a dedicated consensus model. The Council pipeline operates in three phases: (1) an intelligent triage classifier that routes queries based on complexity, (2) parallel expert generation across architecturally diverse models, and (3) a structured consensus synthesis that explicitly identifies agreement, disagreement, and unique findings before producing the final response. We implement and evaluate this architecture within an open-source AI workspace. Our comprehensive evaluation across multiple benchmarks demonstrates that the Council Mode achieves a 35.9% relative reduction in hallucination rates on the HaluEval benchmark and a 7.8-point improvement on TruthfulQA compared to the best-performing individual model, while maintaining significantly lower bias variance across domains. We provide the mathematical formulation of the consensus mechanism, detail the system architecture, and present extensive empirical results with ablation studies.

This paper studies the Exponential Weights (EW) algorithm with an isotropic Gaussian prior for online logistic regression. We show that the near-optimal worst-case regret bound $O(d\log(Bn))$ for EW, established by Kakade and Ng (2005) against the best linear predictor of norm at most $B$, can be achieved with total worst-case computational complexity $O(B^3 n^5)$. This substantially improves on the $O(B^{18}n^{37})$ complexity of prior work achieving the same guarantee (Foster et al., 2018). Beyond efficiency, we analyze the large-$B$ regime under linear separability: after rescaling by $B$, the EW posterior converges as $B\to\infty$ to a standard Gaussian truncated to the version cone. Accordingly, the predictor converges to a solid-angle vote over separating directions and, on every fixed-margin slice of this cone, the mode of the corresponding truncated Gaussian is aligned with the hard-margin SVM direction. Using this geometry, we derive non-asymptotic regret bounds showing that once $B$ exceeds a margin-dependent threshold, the regret becomes independent of $B$ and grows only logarithmically with the inverse margin. Overall, our results show that EW can be both computationally tractable and geometrically adaptive in online classification.

We develop a sketch-based factor reduction and a Nesterov-accelerated projected gradient algorithm (NPGA) with GPU acceleration, yielding a doubly accelerated solver for large-scale constrained mean-variance portfolio optimization. Starting from the sample covariance factor $L$, the method combines randomized subspace embedding, spectral truncation, and ridge stabilization to construct an effective factor $L_{eff}$. It then solves the resulting constrained problem with a structured projection computed by scalar dual search and GPU-friendly matrix-vector kernels, yielding one computational pipeline for the baseline, sketched, and Sketch-Truncate-Ridge (STR)-regularized models. We also establish approximation, conditioning, and stability guarantees for the sketching and STR models, including explicit $O(\varepsilon)$ bounds for the covariance approximation, the optimal value error, and the solution perturbation under $(\varepsilon,δ)$-subspace embeddings. Experiments on synthetic and real equity-return data show that the method preserves objective accuracy while reducing runtime substantially. On a 5440-asset real-data benchmark with 48374 training periods, NPGA-GPU solves the unreduced full model in 2.80 seconds versus 64.84 seconds for Gurobi, while the optimized compressed GPU variants remain in the low-single-digit-second regime. These results show that the full dense model is already practical on modern GPUs and that, after compression, the remaining bottleneck is projection rather than matrix-vector multiplication.

Partial deepfake speech detection requires identifying manipulated regions that may occur within short temporal portions of an otherwise bona fide utterance, making the task particularly challenging for conventional utterance-level classifiers. We propose a split-and-conquer framework that decomposes the problem into two stages: boundary detection and segment-level classification. A dedicated boundary detector first identifies temporal transition points, allowing the audio signal to be divided into segments that are expected to contain acoustically consistent content. Each resulting segment is then evaluated independently to determine whether it corresponds to bona fide or fake speech. This formulation simplifies the learning objective by explicitly separating temporal localization from authenticity assessment, allowing each component to focus on a well-defined task. To further improve robustness, we introduce a reflection-based multi-length training strategy that converts variable-duration segments into several fixed input lengths, producing diverse feature-space representations. Each stage is trained using multiple configurations with different feature extractors and augmentation strategies, and their complementary predictions are fused to obtain improved final models. Experiments on the PartialSpoof benchmark demonstrate state-of-the-art performance across multiple temporal resolutions as well as at the utterance level, with substantial improvements in the accurate detection and localization of spoofed regions. In addition, the proposed method achieves state-of-the-art performance on the Half-Truth dataset, further confirming the robustness and generalization capability of the framework.

In January 2026, Anthropic published a 79-page "constitution" for its AI model Claude, the most comprehensive corporate AI governance document ever released. This Article offers the first legal and democratic-theoretic analysis of that document. Despite genuine philosophical sophistication, the constitution harbors two structural defects. First, it excludes the contexts where ethical constraints matter most: models deployed to the U.S. military operate under different rules, a gap exposed when Claude remained embedded in Palantir's Maven platform during military strikes in Iran even after a government-wide ban on Anthropic's technology. Second, its very comprehensiveness forecloses democratic contestation by resolving questions about AI values, moral status, and conscientious objection that should remain open for public deliberation. Anthropic's own 2023 experiment in participatory constitution-making found roughly 50% divergence between publicly sourced and corporate-authored principles, with the democratic version producing lower bias across nine social dimensions, yet the 2026 constitution incorporates none of those findings. I argue that AI governance suffers from a "political community deficit": the absence of any democratic body authorized to determine the principles governing AI behavior. Corporate transparency, however admirable, is not democratic legitimacy.

Classic AI planning problems have been revisited in the Large Language Model (LLM) era, with a focus of recent benchmarks on success rates rather than plan efficiency. We examine the degree to which frontier models reason optimally versus relying on simple, heuristic, and possibly inefficient strategies. We focus on the Blocksworld domain involving towers of labeled blocks which have to be moved from an initial to a goal configuration via a set of primitive actions. We also study a formally equivalent task, the generalized Path-Star ($P^*$) graph, in order to isolate true topological reasoning from semantic priors. We systematically manipulate problem depth (the height of block towers), width (the number of towers), and compositionality (the number of goal blocks). Reasoning-enhanced LLMs significantly outperform traditional satisficing planners (e.g., LAMA) in complex, multi-goal configurations. Although classical search algorithms hit a wall as the search space expands, LLMs track theoretical optimality limits with near-perfect precision, even when domain-specific semantic hints are stripped away. To explain these surprising findings, we consider (and find evidence to support) two hypotheses: an active Algorithmic Simulation executed via reasoning tokens and a Geometric Memory that allows models to represent the $P^*$ topology as a navigable global geometry, effectively bypassing exponential combinatorial complexity.

In this article, we present a gold-standard benchmark dataset for Biomedical Urdu Named Entity Recognition (BioUNER), developed by crawling health-related articles from online Urdu news portals, medical prescriptions, and hospital health blogs and websites. After preprocessing, three native annotators with familiarity in the medical domain participated in the annotation process using the Doccano text annotation tool and annotated 153K tokens. Following annotation, the proposed BioiUNER dataset was evaluated both intrinsically and extrinsically. An inter-annotator agreement score of 0.78 was achieved, thereby validating the dataset as gold-standard quality. To demonstrate the utility and benchmarking capability of the dataset, we evaluated several machine learning and deep learning models, including Support Vector Machines (SVM), Long Short-Term Memory networks (LSTM), Multilingual BERT (mBERT), and XLM-RoBERTa. The gold-standard BioUNER dataset serves as a reliable benchmark and a valuable addition to Urdu language processing resources.

Long-range 3D object detection remains challenging because LiDAR observations become highly sparse and fragmented in the far field, making reliable context modeling difficult for existing detectors. To address this issue, recent state space model (SSM)-based methods have improved long-range modeling efficiency. However, their effectiveness is still limited by generic serialization strategies that fail to preserve meaningful contextual neighborhoods in sparse scenes. To address this issue, we propose RayMamba, a geometry-aware plug-and-play enhancement for voxel-based 3D detectors. RayMamba organizes sparse voxels into sector-wise ordered sequences through a ray-aligned serialization strategy, which preserves directional continuity and occlusion-related context for subsequent Mamba-based modeling. It is compatible with both LiDAR-only and multimodal detectors, while introducing only modest overhead. Extensive experiments on nuScenes and Argoverse 2 demonstrate consistent improvements across strong baselines. In particular, RayMamba achieves up to 2.49 mAP and 1.59 NDS gain in the challenging 40--50 m range on nuScenes, and further improves VoxelNeXt on Argoverse 2 from 30.3 to 31.2 mAP.

Money laundering presents a persistent challenge for financial institutions worldwide, while criminal organizations constantly evolve their tactics to bypass detection systems. Traditional anti-money laundering approaches mainly rely on predefined risk-based rules, leading to resource-intensive investigations and high numbers of false positive alerts. In order to restrict operational costs from exploding, while billions of transactions are being processed every day, financial institutions are investing in more sophisticated mechanisms to improve existing systems. In this paper, we present ExSTraQt (EXtract Suspicious TRAnsactions from Quasi-Temporal graph representation), an advanced supervised learning approach to detect money laundering (or suspicious) transactions in financial datasets. Our proposed framework excels in performance, when compared to the state-of-the-art AML (Anti Money Laundering) detection models. The key strengths of our framework are sheer simplicity, in terms of design and number of parameters; and scalability, in terms of the computing and memory requirements. We evaluated our framework on transaction-level detection accuracy using a real dataset; and a set of synthetic financial transaction datasets. We consistently achieve an uplift in the F1 score for most datasets, up to 1% for the real dataset; and more than 8% for one of the synthetic datasets. We also claim that our framework could seamlessly complement existing AML detection systems in banks. Our code and datasets are available at https://github.com/mhaseebtariq/exstraqt.

We study visual explanation in geometry education as a Referring Image Segmentation (RIS) problem: given a diagram and a natural language description, the task is to produce a pixel-level mask for the referred geometric element. However, existing RIS models trained on natural image benchmarks such as RefCOCO fail catastrophically on geometric diagrams due to the fundamental domain shift between photographic scenes and abstract, textureless schematics. To address the absence of suitable training data, we present a fully automated procedural data engine that generates over 200,000 synthetic geometry diagrams with pixel-perfect segmentation masks and linguistically diverse referring expressions, requiring zero manual annotation. We further propose domain-specific fine-tuning of vision-language models (VLMs), demonstrating that a fine-tuned Florence-2 achieves 49% IoU and 85% Buffered IoU (BIoU), compared to <1% IoU in zero-shot settings. We introduce Buffered IoU, a geometry-aware evaluation metric that accounts for thin-structure localization, and show that it better reflects true segmentation quality than standard IoU. Our results establish a foundation for building Artificial General Teachers (AGTs) capable of providing visually grounded, step-by-step explanations of geometry problems.

Data assimilation is the process of estimating the time-evolving state of a dynamical system by integrating model predictions and noisy observations. It is commonly formulated as Bayesian filtering, but classical filters often struggle with accuracy or computational feasibility in high dimensions. Recently, score-based generative models have emerged as a scalable approach for high-dimensional data assimilation, enabling accurate modeling and sampling of complex distributions. However, existing score-based filters often specify the forward process independently of the data assimilation. As a result, the measurement-update step depends on heuristic approximations of the likelihood score, which can accumulate errors and degrade performance over time. Here, we propose a measurement-aware score-based filter (MASF) that defines a measurement-aware forward process directly from the measurement equation. This construction makes the likelihood score analytically tractable: for linear measurements, we derive the exact likelihood score and combine it with a learned prior score to obtain the posterior score. Numerical experiments covering a range of settings, including high-dimensional datasets, demonstrate improved accuracy and stability over existing score-based filters.

Bilevel optimization is a field of significant theoretical and practical interest, yet solving such optimization problems remains challenging. Evolutionary methods have been employed to address these problems in the black-box setting; however, they incur high computational cost due to the nested nature of bilevel optimization. Although previous methods have attempted to reduce this cost through various heuristic techniques, such approaches limit versatility on challenging optimization landscapes, such as those with multimodality and significant interaction between upper- and lower-level decision variables. In this study, we propose an efficient framework that exploits the invariance of rank-based evolutionary algorithms to monotonic transformations, thereby reducing the computational burden of the lower-level optimization loop. Specifically, our method directly approximates the rankings of the upper-level value function, bypassing the need to run the lower-level optimizer until convergence for each upper-level iteration. We apply this framework to the setting where both levels are continuous, adopting CMA-ES as the optimizer. We demonstrate that our method achieves competitive performance on standard bilevel optimization benchmarks and can solve problems that are intractable with previously proposed methods, particularly those with multimodality and strong inter-variable interactions.

Feature maps associated with positive definite kernels play a central role in kernel methods and learning theory, where regularity properties such as Lipschitz continuity are closely related to robustness and stability guarantees. Despite their importance, explicit characterizations of the Lipschitz constant of kernel feature maps are available only in a limited number of cases. In this paper, we study the Lipschitz regularity of feature maps associated with integral kernels under differentiability assumptions. We first provide sufficient conditions ensuring Lipschitz continuity and derive explicit formulas for the corresponding Lipschitz constants. We then identify a condition under which the feature map fails to be Lipschitz continuous and apply these results to several important classes of kernels. For infinite width two-layer neural network with isotropic Gaussian weight distributions, we show that the Lipschitz constant of the associated kernel can be expressed as the supremum of a two-dimensional integral, leading to an explicit characterization for the Gaussian kernel and the ReLU random neural network kernel. We also study continuous and shift-invariant kernels such as Gaussian, Laplace, and Matérn kernels, which admit an interpretation as neural network with cosine activation function. In this setting, we prove that the feature map is Lipschitz continuous if and only if the weight distribution has a finite second-order moment, and we then derive its Lipschitz constant. Finally, we raise an open question concerning the asymptotic behavior of the convergence of the Lipschitz constant in finite width neural networks. Numerical experiments are provided to support this behavior.

Weight-space model merging combines independently fine-tuned models without accessing original training data, offering a practical alternative to joint training. While merging succeeds in multitask settings, its behavior in multilingual contexts remains poorly understood. We systematically study weight-space merging for multilingual machine translation by fully fine-tuning language model on large-scale bilingual corpora and evaluating standard merging strategies. Our experiments reveal that merging degrades performance, especially when target languages differ. To explain this failure, we analyze internal representations using span-conditioned neuron selectivity and layer-wise centered kernel alignment. We find that language-specific neurons concentrate in embedding layers and upper transformer blocks, while intermediate layers remain largely shared across languages. Critically, fine-tuning redistributes rather than sharpens language selectivity: neurons for supervised and related languages become less exclusive, while those for unsupervised languages grow more isolated. This redistribution increases representational divergence in higher layers that govern generation. These findings suggest that multilingual fine-tuning may reshape geometry in ways that reduce compatibility with standard weight-space merging assumptions. Our work thus provides an explanation for why merging fails in multilingual translation scenarios.

Operational flood forecasting still relies on high-fidelity two-dimensional hydraulic solvers, but their runtime can be prohibitive for rapid decision support on large urban floodplains. In parallel, AI-based surrogate models have shown strong potential in several areas of computational physics for accelerating otherwise expensive high-fidelity simulations. We address this issue on the lower Têt River (France), starting from a production-grade Telemac2D model defined on a high-resolution unstructured finite-element mesh with more than $4\times 10^5$ nodes. From this setup, we build a learning-ready database of synthetic but operationally grounded flood events covering several representative hydrograph families and peak discharges. On top of this database, we develop a graph-neural surrogate based on projected meshes and multimesh connectivity. The projected-mesh strategy keeps training tractable while preserving high-fidelity supervision from the original Telemac simulations, and the multimesh construction enlarges the effective spatial receptive field without increasing network depth. We further study the effect of an explicit discharge feature $Q(t)$ and of pushforward training for long autoregressive rollouts. The experiments show that conditioning on $Q(t)$ is essential in this boundary-driven setting, that multimesh connectivity brings additional gains once the model is properly conditioned, and that pushforward further improves rollout stability. Among the tested configurations, the combination of $Q(t)$, multimesh connectivity, and pushforward provides the best overall results. These gains are observed both on hydraulic variables over the surrogate mesh and on inundation maps interpolated onto a common $25\,\mathrm{m}$ regular grid and compared against the original high-resolution Telemac solution. On the studied case, the learned surrogate produces 6-hour predictions in about $0.4\,\mathrm{s}$ on a single NVIDIA A100 GPU, compared with about $180\,\mathrm{min}$ on 56 CPU cores for the reference simulation. These results support graph-based surrogates as practical complements to industrial hydraulic solvers for operational flood mapping.

Training tool-calling agents with reinforcement learning on multi-turn tasks remains challenging due to sparse outcome rewards and difficult credit assignment across conversation turns. We present the first application of MT-GRPO (Multi-Turn Group Relative Policy Optimization) combined with GTPO (Generalized Token-level Policy Optimization) for training a tool-calling agent on realistic customer service tasks with an LLM-based user simulator. Through systematic analysis of training rollouts, we discover that naively designed dense per-turn rewards degrade performance by up to 14 percentage points due to misalignment between reward discriminativeness and advantage direction. We introduce Iterative Reward Calibration, a methodology for designing per-turn rewards using empirical discriminative analysis of rollout data, and show that our GTPO hybrid advantage formulation eliminates the advantage misalignment problem. Applied to the Tau-Bench airline benchmark, our approach improves Qwen3.5-4B from 63.8 percent to 66.7 percent (+2.9pp) and Qwen3-30B-A3B from 58.0 percent to 69.5 percent (+11.5pp) -- with the trained 4B model exceeding GPT-4.1 (49.4 percent) and GPT-4o (42.8 percent) despite being 50 times smaller, and the 30.5B MoE model approaching Claude Sonnet 4.5 (70.0 percent). To our knowledge, these are the first published RL training results on Tau-Bench. We release our code, reward calibration analysis, and training recipes.

Knowledge Graph construction from natural language requires extracting structured triplets from complex, information-dense sentences. In this paper, we investigate if the decomposition of text into atomic propositions (minimal, semantically autonomous units of information) can improve the triplet extraction. We introduce MPropositionneur-V2, a small multilingual model covering six European languages trained by knowledge distillation from Qwen3-32B into a Qwen3-0.6B architecture, and we evaluate its integration into two extraction paradigms: entity-centric (GLiREL) and generative (Qwen3). Experiments on SMiLER, FewRel, DocRED and CaRB show that atomic propositions benefit weaker extractors (GLiREL, CoreNLP, 0.6B models), improving relation recall and, in the multilingual setting, overall accuracy. For stronger LLMs, a fallback combination strategy recovers entity recall losses while preserving the gains in relation extraction. These results show that atomic propositions are an interpretable intermediate data structure that complements extractors without replacing them.

Majority voting is the standard for aggregating multi-agent responses into a final decision. However, traditional methods typically require all agents to complete their reasoning before aggregation begins, leading to significant computational overhead, as many responses become redundant once a majority consensus is achieved. In this work, we formulate the multi-agent voting as a reliability-aware agent scheduling problem, and propose an Efficient Majority-then-Stopping (EMS) to improve reasoning efficiency. EMS prioritizes agents based on task-aware reliability and terminates the reasoning pipeline the moment a majority is achieved from the following three critical components. Specifically, we introduce Agent Confidence Modeling (ACM) to estimate agent reliability using historical performance and semantic similarity, Adaptive Incremental Voting (AIV) to sequentially select agents with early stopping, and Individual Confidence Updating (ICU) to dynamically update the reliability of each contributing agent. Extensive evaluations across six benchmarks demonstrate that EMS consistently reduces the average number of invoked agents by 32%.

The integration of large language models (LLMs) with graph-structured data has become a pivotal and fast evolving research frontier, drawing strong interest from both academia and industry. The 2nd LLM+Graph Workshop, co-located with the 51st International Conference on Very Large Data Bases (VLDB 2025) in London, focused on advancing algorithms and systems that bridge LLMs, graph data management, and graph machine learning for practical applications. This report highlights the key research directions, challenges, and innovative solutions presented by the workshop's speakers.

Temporal sentence grounding in videos (TSGV) aims to localize a temporal segment that semantically corresponds to a sentence query from an untrimmed video. Most current methods adopt pre-trained query-agnostic visual encoders for offline feature extraction, and the video backbones are frozen and not optimized for TSGV. This leads to a task discrepancy issue for the video backbone trained for visual classification, but utilized for TSGV. To bridge this gap, we propose a fully end-to-end paradigm that jointly optimizes the video backbone and localization head. We first conduct an empirical study validating the effectiveness of end-to-end learning over frozen baselines across different model scales. Furthermore, we introduce a Sentence Conditioned Adapter (SCADA), which leverages sentence features to train a small portion of video backbone parameters adaptively. SCADA facilitates the deployment of deeper network backbones with reduced memory and significantly enhances visual representation by modulating feature maps through precise integration of linguistic embeddings. Experiments on two benchmarks show that our method outperforms state-of-the-art approaches. The code and models will be released.

Automating C-to-Rust migration is critical for improving software security without sacrificing performance. Traditional rule-based methods struggle with diverse C idioms, often producing rigid and unidiomatic Rust code. Large Language Models (LLMs), trained on massive code corpora, offer a promising alternative by leveraging cross-language generalization to generate more idiomatic and maintainable Rust code. However, several challenges remain. First, existing LLM-based approaches fail to handle cross-file dependencies effectively, either ignoring them or including entire files as context, which limits accurate dependency modeling. Second, complex dependencies and structured inputs and outputs make it difficult to verify syntactic correctness and functional equivalence at the repository level. Third, the lack of large-scale C-Rust parallel data constrains model performance. We propose DepTrans, a framework that combines model capability enhancement with structured inference. DepTrans introduces Reinforcement-Aligned Syntax Training to improve generation quality through multi-task fine-tuning and feedback-driven reinforcement learning. It further applies Dependency-Guided Iterative Refinement to capture fine-grained cross-file dependencies and iteratively refine generated Rust code. We construct a dataset of 85k training samples and a benchmark of 145 repository-level instances. Experiments show that DepTrans achieves a 60.7 percent compilation success rate and 43.5 percent computational accuracy, outperforming the strongest baseline by 22.8 and 17.3 percentage points. It also successfully builds 7 of 15 industrial C projects, demonstrating its practical potential.

The ocean interior regulates Earth's climate but remains sparsely observed due to limited in situ measurements, while satellite observations are restricted to the surface. We present a depth-aware generative framework for reconstructing high-resolution three-dimensional ocean states from extremely sparse surface data. Our approach employs a conditional denoising diffusion probabilistic model (DDPM) trained on sea surface height and temperature observations with up to 99.9 percent sparsity, without reliance on a background dynamical model. By incorporating continuous depth embeddings, the model learns a unified vertical representation of the ocean states and generalizes to previously unseen depths. Applied to the Gulf of Mexico, the framework accurately reconstructs subsurface temperature, salinity, and velocity fields across multiple depths. Evaluations using statistical metrics, spectral analysis, and heat transport diagnostics demonstrate recovery of both large-scale circulation and multiscale variability. These results establish generative diffusion models as a scalable approach for probabilistic ocean reconstruction in data-limited regimes, with implications for climate monitoring and forecasting.

We show that the error-gated Hebbian rule for PCA (EGHR-PCA), a three-factor learning rule equivalent to Oja's subspace rule under Gaussian inputs, can be systematically derived from Oja's subspace rule using frame theory. The global third factor in EGHR-PCA arises exactly as a frame coefficient when the learning rule is expanded with respect to a natural frame on the space of symmetric matrices. This provides a principled, non-heuristic derivation of a biologically plausible learning rule from its mathematically canonical counterpart.

Agent Skills is an emerging open standard that defines a modular, filesystem-based packaging format enabling LLM-based agents to acquire domain-specific expertise on demand. Despite rapid adoption across multiple agentic platforms and the emergence of large community marketplaces, the security properties of Agent Skills have not been systematically studied. This paper presents the first comprehensive security analysis of the Agent Skills framework. We define the full lifecycle of an Agent Skill across four phases -- Creation, Distribution, Deployment, and Execution -- and identify the structural attack surface each phase introduces. Building on this lifecycle analysis, we construct a threat taxonomy comprising seven categories and seventeen scenarios organized across three attack layers, grounded in both architectural analysis and real-world evidence. We validate the taxonomy through analysis of five confirmed security incidents in the Agent Skills ecosystem. Based on these findings, we discuss defense directions for each threat category, identify open research challenges, and provide actionable recommendations for stakeholders. Our analysis reveals that the most severe threats arise from structural properties of the framework itself, including the absence of a data-instruction boundary, a single-approval persistent trust model, and the lack of mandatory marketplace security review, and cannot be addressed through incremental mitigations alone.

Longitudinal health agents must reason across multi-source trajectories that combine continuous device streams, sparse clinical exams, and episodic life events - yet evaluating them is hard: real-world data cannot be released at scale, and temporally grounded attribution questions seldom admit definitive answers without structured ground truth. We present ESL-Bench, an event-driven synthesis framework and benchmark providing 100 synthetic users, each with a 1-5 year trajectory comprising a health profile, a multi-phase narrative plan, daily device measurements, periodic exam records, and an event log with explicit per-indicator impact parameters. Each indicator follows a baseline stochastic process driven by discrete events with sigmoid-onset, exponential-decay kernels under saturation and projection constraints; a hybrid pipeline delegates sparse semantic artifacts to LLM-based planning and dense indicator dynamics to algorithmic simulation with hard physiological bounds. Users are each paired with 100 evaluation queries across five dimensions - Lookup, Trend, Comparison, Anomaly, Explanation - stratified into Easy, Medium, and Hard tiers, with all ground-truth answers programmatically computable from the recorded event-indicator relationships. Evaluating 13 methods spanning LLMs with tools, DB-native agents, and memory-augmented RAG, we find that DB agents (48-58%) substantially outperform memory RAG baselines (30-38%), with the gap concentrated on Comparison and Explanation queries where multi-hop reasoning and evidence attribution are required.

Accurate forecasting of recovery rates (RR) is central to credit risk management and regulatory capital determination. In many loan portfolios, however, RR modeling is constrained by data scarcity arising from infrequent default events. Transfer learning (TL) offers a promising avenue to mitigate this challenge by exploiting information from related but richer source domains, yet its effectiveness critically depends on the presence and strength of distributional shifts, and on potential heterogeneity between source and target feature spaces. This paper introduces FT-MDN-Transformer, a mixture-density tabular Transformer architecture specifically designed for TL in RR forecasting across heterogeneous feature sets. The model produces both loan-level point estimates and portfolio-level predictive distributions, thereby supporting a wide range of practical RR forecasting applications. We evaluate the proposed approach in a controlled Monte Carlo simulation that facilitates systematic variation of covariate, conditional, and label shifts, as well as in a real-world transfer setting using the Global Credit Data (GCD) loan dataset as source and a novel bonds dataset as target. Our results show that FT-MDN-Transformer outperforms baseline models when target-domain data are limited, with particularly pronounced gains under covariate and conditional shifts, while label shift remains challenging. We also observe its probabilistic forecasts to closely track empirical recovery distributions, providing richer information than conventional point-prediction metrics alone. Overall, the findings highlight the potential of distribution-aware TL architectures to improve RR forecasting in data-scarce credit portfolios and offer practical insights for risk managers operating under heterogeneous data environments.

Detecting whether a model's internal knowledge is sufficient to correctly answer a given question is a fundamental challenge in deploying responsible LLMs. In addition to verbalising the confidence by LLM self-report, more recent methods explore the model internals, such as the hidden states of the response tokens to capture how much knowledge is activated. We argue that such activated knowledge may not align with what the query requires, e.g., capturing the stylistic and length-related features that are uninformative for answering the query. To fill the gap, we propose GRADE (Gradient Dynamics for knowledge gap detection), which quantifies the knowledge gap via the cross-layer rank ratio of the gradient to that of the corresponding hidden state subspace. This is motivated by the property of gradients as estimators of the required knowledge updates for a given target. We validate \modelname{} on six benchmarks, demonstrating its effectiveness and robustness to input perturbations. In addition, we present a case study showing how the gradient chain can generate interpretable explanations of knowledge gaps for long-form answers.

Creating flexible 3D scenes from a single image is vital when direct 3D data acquisition is costly or impractical. We introduce NavCrafter, a novel framework that explores 3D scenes from a single image by synthesizing novel-view video sequences with camera controllability and temporal-spatial consistency. NavCrafter leverages video diffusion models to capture rich 3D priors and adopts a geometry-aware expansion strategy to progressively extend scene coverage. To enable controllable multi-view synthesis, we introduce a multi-stage camera control mechanism that conditions diffusion models with diverse trajectories via dual-branch camera injection and attention modulation. We further propose a collision-aware camera trajectory planner and an enhanced 3D Gaussian Splatting (3DGS) pipeline with depth-aligned supervision, structural regularization and refinement. Extensive experiments demonstrate that NavCrafter achieves state-of-the-art novel-view synthesis under large viewpoint shifts and substantially improves 3D reconstruction fidelity.

Large reasoning models achieve strong performance on complex tasks through long chain-of-thought (CoT) trajectories, but directly transferring such reasoning processes to smaller models remains challenging. A key difficulty is that not all teacher-generated reasoning trajectories are suitable for student learning. Existing approaches typically rely on post-hoc filtering, selecting trajectories after full generation based on heuristic criteria. However, such methods cannot control the generation process itself and may still produce reasoning paths that lie outside the student's learning capacity. To address this limitation, we propose Gen-SSD (Generation-time Self-Selection Distillation), a student-in-the-loop framework that performs generation-time selection. Instead of passively consuming complete trajectories, the student evaluates candidate continuations during the teacher's sampling process, guiding the expansion of only learnable reasoning paths and enabling early pruning of unhelpful branches. Experiments on mathematical reasoning benchmarks demonstrate that Gen-SSD consistently outperforms standard knowledge distillation and recent baselines, with improvements of around 5.9 points over Standard KD and up to 4.7 points over other baselines. Further analysis shows that Gen-SSD produces more stable and learnable reasoning trajectories, highlighting the importance of incorporating supervision during generation for effective distillation.

Multimodal Large Language Models (MLLMs) have shown strong reasoning ability, but their high computational and memory costs hinder deployment in resource-constrained settings. While Post-Training Quantization (PTQ) and vision token pruning are standard compression techniques, they are usually treated as independent optimizations. In this paper, we show that these two techniques are strongly coupled: naively applying semantic-based token pruning to PTQ-optimized MLLMs can discard activation outliers that are important for numerical stability and thus worsen quantization errors in low-bit regimes (\textit{e.g.}, W4A4). To address this issue, we propose a quantization-aware vision token pruning framework. Our method introduces a lightweight hybrid sensitivity metric that combines simulated group-wise quantization error with outlier intensity. By combining this metric with standard semantic relevance scores, the method retains tokens that are both semantically informative and robust to quantization. Experiments on standard LLaVA architectures show that our method consistently outperforms naive integration baselines. At an aggressive pruning ratio that retains only 12.5\% of visual tokens, our framework improves accuracy by 2.24\% over the baseline and even surpasses dense quantization without pruning. To the best of our knowledge, this is the first method that explicitly co-optimizes vision token pruning and PTQ for accurate low-bit MLLM inference.

Functional verification consumes over 50% of the IC development lifecycle, where SystemVerilog Assertions (SVAs) are indispensable for formal property verification and enhanced simulation-based debugging. However, manual SVA authoring is labor-intensive and error-prone. While Large Language Models (LLMs) show promise, their direct deployment is hindered by low functional accuracy and a severe scarcity of domain-specific data. To address these challenges, we introduce ChatSVA, an end-to-end SVA generation system built upon a multi-agent framework. At its core, the AgentBridge platform enables this multi-agent approach by systematically generating high-purity datasets, overcoming the data scarcity inherent to few-shot scenarios. Evaluated on 24 RTL designs, ChatSVA achieves 98.66% syntax and 96.12% functional pass rates, generating 139.5 SVAs per design with 82.50% function coverage. This represents a 33.3 percentage point improvement in functional correctness and an over 11x enhancement in function coverage compared to the previous state-of-the-art (SOTA). ChatSVA not only sets a new SOTA in automated SVA generation but also establishes a robust framework for solving long-chain reasoning problems in few-shot, domain-specific scenarios. An online service has been publicly released at https://www.nctieda.com/CHATDV.html.

Pavement condition assessment is essential for road safety and maintenance. Existing research has made significant progress. However, most studies focus on conventional computer vision tasks such as classification, detection, and segmentation. In real-world applications, pavement inspection requires more than visual recognition. It also requires quantitative analysis, explanation, and interactive decision support. Current datasets are limited. They focus on unimodal perception. They lack support for multi-turn interaction and fact-grounded reasoning. They also do not connect perception with vision-language analysis. To address these limitations, we introduce PaveBench, a large-scale benchmark for pavement distress perception and interactive vision-language analysis on real-world highway inspection images. PaveBench supports four core tasks: classification, object detection, semantic segmentation, and vision-language question answering. It provides unified task definitions and evaluation protocols. On the visual side, PaveBench provides large-scale annotations and includes a curated hard-distractor subset for robustness evaluation. It contains a large collection of real-world pavement images. On the multimodal side, we introduce PaveVQA, a real-image question answering (QA) dataset that supports single-turn, multi-turn, and expert-corrected interactions. It covers recognition, localization, quantitative estimation, and maintenance reasoning. We evaluate several state-of-the-art methods and provide a detailed analysis. We also present a simple and effective agent-augmented visual question answering framework that integrates domain-specific models as tools alongside vision-language models. The dataset is available at: https://huggingface.co/datasets/MML-Group/PaveBench.

Rubric-based Reinforcement Learning (RL) has emerged as a promising approach for aligning Large Language Models (LLMs) with complex, open-domain instruction following tasks. However, existing methods predominantly rely on response-level rewards, introducing severe reward sparsity and reward ambiguity problems. To address these issues, we propose Rubrics to Tokens (RTT), a novel rubric-based RL framework that bridges coarse response-level scores and fine-grained token-level credit assignment. RTT introduces a Token-Level Relevance Discriminator to predict which tokens in the response are responsible for a specific constraint, and optimizes the policy model via RTT-GRPO, which integrates response-level and token-level advantages within a unified framework. Furthermore, when transitioning from one-dimensional, outcome-level reward to three-dimensional reward space in the token-level rubric-based RL, we propose a novel group normalization method, called Intra-sample Token Group Normalization, to accommodate this shift. Extensive experiments and benchmarks demonstrate that RTT consistently outperforms other baselines in both instruction- and rubric-level accuracy across different models.

Charts are ubiquitous in scientific and financial literature for presenting structured data. However, chart reasoning remains challenging for multimodal large language models (MLLMs) due to the lack of high-quality training data, as well as the need for fine-grained visual grounding and precise numerical computation. To address these challenges, we first propose DuoChart, a scalable dual-source data pipeline that combines synthesized charts with real-world charts to construct diverse, high-quality chart training data. We then introduce CharTool, which equips MLLMs with external tools, including image cropping for localized visual perception and code-based computation for accurate numerical reasoning. Through agentic reinforcement learning on DuoChart, CharTool learns tool-integrated reasoning grounded in chart content. Extensive experiments on six chart benchmarks show that our method consistently improves over strong MLLM baselines across model scales. Notably, CharTool-7B outperforms the base model by **+8.0%** on CharXiv (Reasoning) and **+9.78%** on ChartQAPro, while achieving competitive performance with substantially larger or proprietary models. Moreover, CharTool demonstrates positive generalization to out-of-domain visual math reasoning benchmarks.

We study Byzantine-resilient distributed multi-agent reinforcement learning (MARL), where agents must collaboratively learn optimal value functions over a compromised communication network. Existing resilient MARL approaches typically guarantee almost sure convergence only to near-optimal value functions, or require restrictive assumptions to ensure convergence to optimal solution. As a result, agents may fail to learn the optimal policies under these methods. To address this, we propose a novel distributed Q-learning algorithm, under which all agents' value functions converge almost surely to the optimal value functions despite Byzantine edge attacks. The key idea is a redundancy-based filtering mechanism that leverages two-hop neighbor information to validate incoming messages, while preserving bidirectional information flow. We then introduce a new topological condition for the convergence of our algorithm, present a systematic method to construct such networks, and prove that this condition can be verified in polynomial time. We validate our results through simulations, showing that our method converges to the optimal solutions, whereas prior methods fail under Byzantine edge attacks.

The growing number of individual generating units, hybrid resources, and security constraints has significantly increased the computational burden of network-constrained unit commitment (UC), where most solution time is spent exploring branch-and-bound trees over unit-hour binary variables. To reduce this combinatorial burden, recent approaches have explored learning-based guidance to assist commitment decisions. However, directly using tools such as large language models (LLMs) to predict full commitment schedules is unreliable, as infeasible or inconsistent binary decisions can violate inter-temporal constraints and degrade economic optimality. This paper proposes a solver-compatible dimensionality reduction framework for UC that exploits structural regularities in commitment decisions. Instead of generating complete schedules, the framework identifies a sparse subset of structurally stable commitment binaries to fix prior to optimization. One implementation uses an LLM to select these variables. The LLM does not replace the optimization process but provides partial variable restriction, while all constraints and remaining decisions are handled by the original MILP solver, which continues to enforce network, ramping, reserve, and security constraints. We formally show that the masked problem defines a reduced feasible region of the original UC model, thereby preserving feasibility and enabling solver-certified optimality within the restricted space. Experiments on IEEE 57-bus, RTS 73-bus, IEEE 118-bus, and augmented large-scale cases, including security-constrained variants, demonstrate consistent reductions in branch-and-bound nodes and solution time, achieving order-of-magnitude speedups on high-complexity instances while maintaining near-optimal objective values.

The rapid adoption of HDR-capable devices has created a pressing need to convert the 8-bit Standard Dynamic Range (SDR) content into perceptually and physically accurate 10-bit High Dynamic Range (HDR). Existing inverse tone-mapping (ITM) methods often rely on fixed tone-mapping operators that struggle to generalize to real-world degradations, stylistic variations, and camera pipelines, frequently producing clipped highlights, desaturated colors, or unstable tone reproduction. We introduce LumaFlux, a first physically and perceptually guided diffusion transformer (DiT) for SDR-to-HDR reconstruction by adapting a large pretrained DiT. Our LumaFlux introduces (1) a Physically-Guided Adaptation (PGA) module that injects luminance, spatial descriptors, and frequency cues into attention through low-rank residuals; (2) a Perceptual Cross-Modulation (PCM) layer that stabilizes chroma and texture via FiLM conditioning from vision encoder features; and (3) an HDR Residual Coupler that fuses physical and perceptual signals under a timestep- and layer-adaptive modulation schedule. Finally, a lightweight Rational-Quadratic Spline decoder reconstructs smooth, interpretable tone fields for highlight and exposure expansion, enhancing the output of the VAE decoder to generate HDR. To enable robust HDR learning, we curate the first large-scale SDR-HDR training corpus. For fair and reproducible comparison, we further establish a new evaluation benchmark, comprising HDR references and corresponding expert-graded SDR versions. Across benchmarks, LumaFlux outperforms state-of-the-art baselines, achieving superior luminance reconstruction and perceptual color fidelity with minimal additional parameters.

Vision-Language Models (VLMs) excel at multimodal tasks, but they remain vulnerable to hallucinations that are factually incorrect or ungrounded in the input image. Recent work suggests that hallucination detection using internal representations is more efficient and accurate than approaches that rely solely on model outputs. However, existing internal-representation-based methods typically rely on a single representation or detector, limiting their ability to capture diverse hallucination signals. In this paper, we propose EnsemHalDet, an ensemble-based hallucination detection framework that leverages multiple internal representations of VLMs, including attention outputs and hidden states. EnsemHalDet trains independent detectors for each representation and combines them through ensemble learning. Experimental results across multiple VQA datasets and VLMs show that EnsemHalDet consistently outperforms prior methods and single-detector models in terms of AUC. These results demonstrate that ensembling diverse internal signals significantly improves robustness in multimodal hallucination detection.

AI chatbots are increasingly stepping into roles as collaborators or teachers in analyzing, visualizing, and reasoning through data and domain problem. Yet, AI's default assistant mode with its comprehensive and one-off responses may undermine opportunities for practitioners to develop literacy through their own thinking, inducing cognitive passivity. Drawing on evidence from empirical studies and theories, we argue that disrupting cognitive passivity necessitates a nuanced approach: rather than simply making AI promote deliberative thinking, there is a need for more dynamic and adaptive strategy through cognitive alignment -- a framework that characterizes effective human-AI interaction as a function of alignment between users' cognitive demand and AI's interaction mode. In the framework, we provide the mapping between AI's interaction mode (transmissive or deliberative) and users' cognitive demand (receptive or deliberative), otherwise leading to either cognitive passivity or friction. We further discuss implications and offer open questions for future research on data literacy.

Real-world multimodal knowledge graphs (MMKGs) are dynamic, with new entities, relations, and multimodal knowledge emerging over time. Existing continual knowledge graph reasoning (CKGR) methods focus on structural triples and cannot fully exploit multimodal signals from new entities. Existing multimodal knowledge graph reasoning (MMKGR) methods, however, usually assume static graphs and suffer catastrophic forgetting as graphs evolve. To address this gap, we present a systematic study of continual multimodal knowledge graph reasoning (CMMKGR). We construct several continual multimodal knowledge graph benchmarks from existing MMKG datasets and propose MRCKG, a new CMMKGR model. Specifically, MRCKG employs a multimodal-structural collaborative curriculum to schedule progressive learning based on the structural connectivity of new triples to the historical graph and their multimodal compatibility. It also introduces a cross-modal knowledge preservation mechanism to mitigate forgetting through entity representation stability, relational semantic consistency, and modality anchoring. In addition, a multimodal contrastive replay scheme with a two-stage optimization strategy reinforces learned knowledge via multimodal importance sampling and representation alignment. Experiments on multiple datasets show that MRCKG preserves previously learned multimodal knowledge while substantially improving the learning of new knowledge.

The shift from cloud-hosted Large Language Models (LLMs) to locally deployed open-source Small Language Models (SLMs) has democratized AI-assisted coding; however, it has also decentralized the environmental footprint of AI. While prompting strategies - such as Chain-of-Thought and ReAct - serve as external mechanisms for optimizing code generation without modifying model parameters, their impact on energy consumption and carbon emissions remains largely invisible to developers. This paper presents the first systematic empirical study investigating how different prompt engineering strategies in SLM-based code generation impact code generation accuracy alongside sustainability factors. We evaluate six prominent prompting strategies across 11 open-source models (ranging from 1B to 34B parameters) using the HumanEval+ and MBPP+ benchmarks. By measuring Pass@1 accuracy alongside energy (kWh), carbon emissions (kgCO2eq), and inference latency, we reveal that sustainability often decouples from accuracy, allowing significant environmental optimizations without sacrificing performance. Our findings indicate that Chain-of-Thought, being a simpler prompting technique, can provide a near-optimal balance between reasoning capability and energy efficiency. Conversely, multi-sampling strategies often incur disproportionate costs for marginal gains. Finally, we identify grid carbon intensity as the dominant factor in deployment-time emissions, highlighting the need for practitioners to consider regional energy profiles. This work provides a quantitative foundation for "green" prompt engineering, enabling developers to align high-performance code generation with ecological responsibility.

Multilingual Pre-trained Language Models (MPLMs) have become essential tools for natural language processing. However, they often exhibit biases related to sensitive attributes such as gender, race, and religion. In this paper, we introduce a comprehensive multilingual debiasing method named Multiple-Debias to address these issues across multiple languages. By incorporating multilingual counterfactual data augmentation and multilingual Self-Debias across both pre-processing and post-processing stages, alongside parameter-efficient fine-tuning, we significantly reduced biases in MPLMs across three sensitive attributes in four languages. We also extended CrowS-Pairs to German, Spanish, Chinese, and Japanese, validating our full-process multilingual debiasing method for gender, racial, and religious bias. Our experiments show that (i) multilingual debiasing methods surpass monolingual approaches in effectively mitigating biases, and (ii) integrating debiasing information from different languages notably improves the fairness of MPLMs.

In large language model (LLM)-driven multi-agent systems, disobey role specification (failure to adhere to the defined responsibilities and constraints of an assigned role, potentially leading to an agent behaving like another) is a major failure mode \cite{DBLP:journals/corr/abs-2503-13657}. To address this issue, in the present paper, we propose a quantitative role clarity to improve role consistency. Firstly, we construct a role assignment matrix $S(φ)=[s_{ij}(φ)]$, where $s_{ij}(φ)$ is the semantic similarity between the $i$-th agent's behavior trajectory and the $j$-th agent's role description. Then we define role clarity matrix $M(φ)$ as $\text{softmax}(S(φ))-I$, where $\text{softmax}(S(φ))$ is a row-wise softmax of $S(φ)$ and $I$ is the identity matrix. The Frobenius norm of $M(φ)$ quantifies the alignment between agents' role descriptions and their behaviors trajectory. Moreover, we employ the role clarity matrix as a regularizer during lightweight fine-tuning to improve role consistency, thereby improving end-to-end task performance. Experiments on the ChatDev multi-agent system show that our method substantially improves role consistency and task performance: with Qwen and Llama, the role overstepping rate decreases from $46.4\%$ to $8.4\%$ and from $43.4\%$ to $0.2\%$, respectively, and the role clarity score increases from $0.5328$ to $0.9097$ and from $0.5007$ to $0.8530$, respectively, the task success rate increases from $0.6769$ to $0.6909$ and from $0.6174$ to $0.6763$, respectively.

When Agent A delegates to Agent B, which invokes Tool C on behalf of User X, no existing framework can answer: whose authorization chain led to this action, and where did it violate policy? This paper introduces SentinelAgent, a formal framework for verifiable delegation chains in federal multi-agent AI systems. The Delegation Chain Calculus (DCC) defines seven properties - six deterministic (authority narrowing, policy preservation, forensic reconstructibility, cascade containment, scope-action conformance, output schema conformance) and one probabilistic (intent preservation) - with four meta-theorems and one proposition establishing the practical infeasibility of deterministic intent verification. The Intent-Preserving Delegation Protocol (IPDP) enforces all seven properties at runtime through a non-LLM Delegation Authority Service. A three-point verification lifecycle achieves 100% combined TPR at 0% FPR on DelegationBench v4 (516 scenarios, 10 attack categories, 13 federal domains). Under black-box adversarial conditions, the DAS blocks 30/30 attacks with 0 false positives. Deterministic properties are unbreakable under adversarial stress testing; intent verification degrades to 13% against sophisticated paraphrasing. Fine-tuning the NLI model on 190 government delegation examples improves P2 from 1.7% to 88.3% TPR (5-fold cross-validated, F1=82.1%). Properties P1, P3-P7 are mechanically verified via TLA+ model checking across 2.7 million states with zero violations. Even when intent verification is evaded, the remaining six properties constrain the adversary to permitted API calls, conformant outputs, traceable actions, bounded cascades, and compliant behavior.

Modern LLMs inherit strong priors from web-scale pretraining, which can limit the headroom of post-training data-selection strategies. While Active Preference Learning (APL) seeks to optimize query efficiency in online Direct Preference Optimization (DPO), the inherent richness of on-policy candidate pools often renders simple Random sampling a surprisingly formidable baseline. We evaluate uncertainty-based APL against Random across harmlessness, helpfulness, and instruction-following settings, utilizing both reward models and LLM-as-a-judge proxies. We find that APL yields negligible improvements in proxy win-rates compared to Random. Crucially, we observe a dissociation where win-rate improves even as general capability -- measured by standard benchmarks -- degrades. APL fails to mitigate this capability collapse or reduce variance significantly better than random sampling. Our findings suggest that in the regime of strong pre-trained priors, the computational overhead of active selection is difficult to justify against the ``cheap diversity'' provided by simple random samples. Our code is available at https://github.com/BootsofLagrangian/random-vs-apl.

Class-incremental learning (CIL) is typically evaluated under predefined schedules with equal-sized tasks, leaving more realistic and complex cases unexplored. However, a practical CIL system should learns immediately when any number of new classes arrive, without forcing fixed-size tasks. We formalize this setting as Free-Flow Class-Incremental Learning (FFCIL), where data arrives as a more realistic stream with a highly variable number of unseen classes each step. It will make many existing CIL methods brittle and lead to clear performance degradation. We propose a model-agnostic framework for robust CIL learning under free-flow arrivals. It comprises a class-wise mean (CWM) objective that replaces sample frequency weighted loss with uniformly aggregated class-conditional supervision, thereby stabilizing the learning signal across free-flow class increments, as well as method-wise adjustments that improve robustness for representative CIL paradigms. Specifically, we constrain distillation to replayed data, normalize the scale of contrastive and knowledge transfer losses, and introduce Dynamic Intervention Weight Alignment (DIWA) to prevent over-adjustment caused by unstable statistics from small class increments. Experiments confirm a clear performance degradation across various CIL baselines under FFCIL, while our strategies yield consistent gains.

The growing adoption of prompt-based automation in software testing raises important issues regarding its computational and environmental sustainability. Existing sustainability studies in AI-driven testing primarily focus on large language models, leaving the impact of prompt engineering strategies largely unexplored - particularly in the context of Small Language Models (SLMs). This gap is critical, as prompt design directly influences inference behavior, execution cost, and resource utilization, even when model size is fixed. To the best of our knowledge, this paper presents the first systematic sustainability evaluation of prompt engineering strategies for automated test generation using SLMs. We analyze seven prompt strategies across three open-source SLMs under a controlled experimental setup. Our evaluation jointly considers execution time, token usage, energy consumption, carbon emissions, and coverage test quality, the latter assessed through coverage analysis of the generated test scripts. The results show that prompt strategies have a substantial and independent impact on sustainability outcomes, often outweighing the effect of model choice. Reasoning intensive strategies such as Chain of Thought and Self-Consistency achieve higher coverage but incur significantly higher execution time, energy consumption, and carbon emissions. In contrast, simpler strategies such as Zero-Shot and ReAct deliver competitive coverage test quality with markedly lower environmental cost, while Least-to-Most and Program of Thought offer balanced trade-offs.

Accurate crowd simulation is crucial for public safety management, emergency evacuation planning, and intelligent transportation systems. However, existing methods, which typically model crowds as a collection of independent individual trajectories, are limited in their ability to capture macroscopic physical laws. This microscopic approach often leads to error accumulation and compromises simulation stability. Furthermore, deep learning-driven methods tend to suffer from low inference efficiency and high computational overhead, making them impractical for large-scale, efficient simulations. To address these challenges, we propose the Spatio-Temporal Decoupled Differential Equation Network (STDDN), a novel framework that guides microscopic trajectory prediction with macroscopic physics. We innovatively introduce the continuity equation from fluid dynamics as a strong physical constraint. A Neural Ordinary Differential Equation (Neural ODE) is employed to model the macroscopic density evolution driven by individual movements, thereby physically regularizing the microscopic trajectory prediction model. We design a density-velocity coupled dynamic graph learning module to formulate the derivative of the density field within the Neural ODE, effectively mitigating error accumulation. We also propose a differentiable density mapping module to eliminate discontinuous gradients caused by discretization and introduce a cross-grid detection module to accurately model the impact of individual cross-grid movements on local density changes. The proposed STDDN method has demonstrated significantly superior simulation performance compared to state-of-the-art methods on long-term tasks across four real-world datasets, as well as a major reduction in inference latency.

Nonlinear time-history evolution problems employing high-fidelity physical models are essential in numerous scientific domains. However, these problems face a critical dual bottleneck: the immense computational cost of time-stepping and the massive memory requirements for maintaining a vast array of state variables. To address these challenges, we propose a novel framework based on heterogeneous memory management for massive ensemble simulations of general nonlinear time-history problems with complex constitutive laws. Taking advantage of recent advancements in CPU-GPU interconnect bandwidth, our approach actively leverages the large capacity of host CPU memory while simultaneously maximizing the throughput of the GPU. This strategy effectively overcomes the GPU memory wall, enabling memory-intensive simulations. We evaluate the performance of the proposed method through comparisons with conventional implementations, demonstrating significant improvements in time-to-solution and energy-to-solution. Furthermore, we demonstrate the practical utility of this framework by developing a Neural Network-based surrogate model using the generated massive datasets. The results highlight the effectiveness of our approach in enabling high-fidelity 3D evaluations and its potential for broader applications in data-driven scientific discovery.

The increasing use of language models in automated test script generation raises concerns about their environmental impact, yet existing sustainability analyses focus predominantly on large language models. As a result, the energy and carbon characteristics of small language models (SLMs) during prompt-driven unit-test script generation remain largely unexplored. To address this gap, this study empirically examines the environmental and performance tradeoffs of SLMs (in the 2B-8B parameter range) using the HumanEval benchmark and adaptive prompt variants (based on the Anthropic template). The analysis uses CodeCarbon to characterize energy consumption carbon emissions and duration under controlled conditions, with unit-test script coverage serving as an initial proxy for generated test quality. Our results show that different SLMs exhibit distinct sustainability profiles - some favor lower energy use and faster execution, while others maintain higher stability or coverage under comparable conditions. Overall, this work provides focused empirical evidence on sustainable SLM-based test script generation, clarifying how prompt structure and model selection jointly shape environmental and performance outcomes.

Diffusion models often degrade when trained in latent spaces (e.g., VAEs), yet the formal causes remain poorly understood. We quantify latent-space diffusability through the rate of change of the Minimum Mean Squared Error (MMSE) along the diffusion trajectory. Our framework decomposes this MMSE rate into contributions from Fisher Information (FI) and Fisher Information Rate (FIR). We demonstrate that while global isometry ensures FI alignment, FIR is governed by the encoder's local geometric properties. Our analysis explicitly decouples latent geometric distortion into three measurable penalties: dimensional compression, tangential distortion, and curvature injection. We derive theoretical conditions for FIR preservation across spaces, ensuring maintained diffusability. Experiments across diverse autoencoding architectures validate our framework and establish these efficient FI and FIR metrics as a robust diagnostic suite for identifying and mitigating latent diffusion failure.

Social media is widely used to share information globally and it also aids to gain attention from the world. When socially sensitive incidents like rape, human rights march, corruption, political controversy, chemical attacks occur, they gain immense attention from people all over the world, causing microblogging platforms like Twitter to get flooded with tweets related to such events. When an event evolves, many other events of a similar nature have happened in and around the same time frame. These are cross events because they are linked to the nature of the main event. Dissemination of information relating to such cross events helps in engaging the masses to share the varied views that emerge out of the similarities and differences between the events. Cross event detection is critical in determining the nature of events. Cross events have fulcrums points, i.e., topics around which the discussion is focused, as the event evolves which must be considered in topic evolution. We have proposed Cross Event Evolution Detection CEED framework which detects cross events that are similar with regards to their temporal nature resulting from main events. Event detection is based on the tweet segmentation using the Wikipedia title database and clustering segments based on a similarity measure. The cross event detection algorithm reveals events that overlap in both time and context to evaluate the effects of these cross events on deliberate negligent human actions. The topic evolution algorithm puts into perspective the change in topics for an events lifetime. The experimental results on a real Twitter data set demonstrate the effectiveness and precision of our proposed framework for both cross event detection and topic evolution algorithm during the evolution of cross events.

This paper focuses on the state estimation problem in distributed sensor networks, where intermittent packet dropouts, corrupted observations, and unknown noise covariances coexist. To tackle this challenge, we formulate the joint estimation of system states, noise parameters, and network reliability as a Bayesian variational inference problem, and propose a novel variational Bayesian adaptive Kalman filter (VB-AKF) to approximate the joint posterior probability densities of the latent parameters. Unlike existing AKF that separately handle missing data and measurement outliers, the proposed VB-AKF adopts a dual-mask generative model with two independent Bernoulli random variables, explicitly characterizing both observable communication losses and latent data authenticity. Additionally, the VB-AKF integrates multiple concurrent multiple observations into the adaptive filtering framework, which significantly enhances statistical identifiability. Comprehensive numerical experiments verify the effectiveness and asymptotic optimality of the proposed method, showing that both parameter identification and state estimation asymptotically converge to the theoretical optimal lower bound with the increase in the number of sensors.

Large language models (LLMs) have demonstrated strong potential in long-horizon decision-making tasks, such as embodied manipulation and web interaction. However, agents frequently struggle with endless trial-and-error loops or deviate from the main objective in complex environments. We attribute these failures to two fundamental errors: global Progress Drift and local Feasibility Violation. Existing methods typically attempt to address both issues simultaneously using a single paradigm. However, these two challenges are fundamentally distinct: the former relies on fuzzy semantic planning, while the latter demands strict logical constraints and state validation. The inherent limitations of such a single-paradigm approach pose a fundamental challenge for existing models in handling long-horizon tasks. Motivated by this insight, we propose a Neuro-Symbolic Dual Memory Framework that explicitly decouples semantic progress guidance from logical feasibility verification. Specifically, during the inference phase, the framework invokes both memory mechanisms synchronously: on one hand, a neural-network-based Progress Memory extracts semantic blueprints from successful trajectories to guide global task advancement; on the other hand, a symbolic-logic-based Feasibility Memory utilizes executable Python verification functions synthesized from failed transitions to perform strict logical validation. Experiments demonstrate that this method significantly outperforms existing competitive baselines on ALFWorld, WebShop, and TextCraft, while drastically reducing the invalid action rate and average trajectory length.

Reasoning benchmarks typically evaluate whether a model derives the correct answer from a fixed premise set, but they under-measure a closely related capability that matters in dynamic environments: belief revision under minimal evidence change. We introduce DeltaLogic, a benchmark transformation protocol that converts natural-language reasoning examples into short revision episodes. Each episode first asks for an initial conclusion under premises P, then applies a minimal edit δ(P), and finally asks whether the previous conclusion should remain stable or be revised. We instantiate DeltaLogic from FOLIO and ProofWriter and evaluate small causal language models with constrained label scoring. On a completed 30-episode Qwen evaluation subset, stronger initial reasoning still does not imply stronger revision behavior: Qwen3-1.7B reaches 0.667 initial accuracy but only 0.467 revision accuracy, with inertia rising to 0.600 on episodes where the gold label should change, while Qwen3-0.6B collapses into near universal abstention. There, Qwen3-4B preserves the same inertial failure pattern (0.650 initial, 0.450 revised, 0.600 inertia), whereas Phi-4-mini-instruct is substantially stronger (0.950 initial, 0.850 revised) but still exhibits non-trivial abstention and control instability. These results suggest that logical competence under fixed premises does not imply disciplined belief revision after local evidence edits. DeltaLogic therefore targets a distinct and practically important reasoning capability that complements existing logical inference and belief-updating benchmarks.

Code generation and comprehension by Large Language Models (LLMs) have emerged as core drivers of industrial intelligence and decision optimization, finding widespread application in fields such as finance, automation, and aerospace. Although recent advancements have demonstrated the remarkable potential of LLMs in general code generation, existing benchmarks are mainly confined to single domains and languages. Consequently, they fail to effectively evaluate the generalization capabilities required for real-world industrial applications or to reflect the coding proficiency demanded by complex industrial scenarios. To bridge this gap, we introduce IndustryCode, the first comprehensive benchmark designed to span multiple industrial domains and programming languages. IndustryCode comprises 579 sub-problems derived from 125 primary industrial challenges, accompanied by rigorous problem descriptions and test cases. It covers a wide range of fields, including finance, automation, aerospace, and remote sensing-and incorporates diverse programming languages such as MATLAB, Python, C++, and Stata. In our evaluation, the top-performing model, Claude 4.5 Opus, achieved an overall accuracy of 68.1% on sub-problems and 42.5% main problems. The benchmark dataset and automated evaluation code will be made publicly available upon acceptance.

The challenges of the uncertainties in renewable energy generation and the instability of the real-time market limit the effective utilization of clean energy in microgrid communities. Existing peer-to-peer (P2P) and microgrid coordination approaches typically rely on certain centralized optimization or restrictive coordination rules which are difficult to be implemented in real-life applications. To address the challenge, we propose an intraday P2P trading framework that allows self-interested microgrids to pursue their economic benefits, while allowing the market operator to maximize the social welfare, namely the low carbon emission objective, of the entire community. Specifically, the decision-making processes of the microgrids are formulated as a Decentralized Partially Observable Markov Decision Process (DEC-POMDP) and solved using a Multi-Agent Reinforcement Learning (MARL) framework. Such an approach grants each microgrid a high degree of decision-making autonomy, while a novel market clearing mechanism is introduced to provide macro-regulation, incentivizing microgrids to prioritize local renewable energy consumption and hence reduce carbon emissions. Simulation results demonstrate that the combination of the self-interested bidding strategy and the P2P market design helps significantly improve renewable energy utilization and reduce reliance on external electricity with high carbon-emissions. The framework achieves a balanced integration of local autonomy, self-interest pursuit, and improved community-level economic and environmental benefits.

Competitive programming remains one of the last few human strongholds in coding against AI. The best AI system to date still underperforms the best humans competitive programming: the most recent best result, Google's Gemini~3 Deep Think, attained 8th place even not being evaluated under live competition conditions. In this work, we introduce GrandCode, a multi-agent RL system designed for competitive programming. The capability of GrandCode is attributed to two key factors: (1) It orchestrates a variety of agentic modules (hypothesis proposal, solver, test generator, summarization, etc) and jointly improves them through post-training and online test-time RL; (2) We introduce Agentic GRPO specifically designed for multi-stage agent rollouts with delayed rewards and the severe off-policy drift that is prevalent in agentic RL. GrandCode is the first AI system that consistently beats all human participants in live contests of competitive programming: in the most recent three Codeforces live competitions, i.e., Round~1087 (Mar 21, 2026), Round~1088 (Mar 28, 2026), and Round~1089 (Mar 29, 2026), GrandCode placed first in all of them, beating all human participants, including legendary grandmasters. GrandCode shows that AI systems have reached a point where they surpass the strongest human programmers on the most competitive coding tasks.

We introduce MOMO, the first multi-sensor foundation model for Mars remote sensing. MOMO uses model merge to integrate representations learned independently from three key Martian sensors (HiRISE, CTX, and THEMIS), spanning resolutions from 0.25 m/pixel to 100 m/pixel. Central to our method is our novel Equal Validation Loss (EVL) strategy, which aligns checkpoints across sensors based on validation loss similarity before fusion via task arithmetic. This ensures models are merged at compatible convergence stages, leading to improved stability and generalization. We train MOMO on a large-scale, high-quality corpus of $\sim 12$ million samples curated from Mars orbital data and evaluate it on 9 downstream tasks from Mars-Bench. MOMO achieves better overall performance compared to ImageNet pre-trained, earth observation foundation model, sensor-specific pre-training, and fully-supervised baselines. Particularly on segmentation tasks, MOMO shows consistent and significant performance improvement. Our results demonstrate that model merging through an optimal checkpoint selection strategy provides an effective approach for building foundation models for multi-resolution data. The model weights, pretraining code, pretraining data, and evaluation code are available at: https://github.com/kerner-lab/MOMO.

Diffusion language models have seen exciting recent progress, offering far more flexibility in generative trajectories than autoregressive models. This flexibility has motivated a growing body of research into new approaches to diffusion language modeling, which typically begins at the scale of GPT-2 small (150 million parameters). However, these advances introduce new issues with evaluation methodology. In this technical note, we discuss the limitations of current methodology and propose principled augmentations to ensure reliable comparisons. We first discuss why OpenWebText has become the standard benchmark, and why alternatives such as LM1B are inherently less meaningful. We then discuss the limitations of likelihood evaluations for diffusion models, and explain why relying on generative perplexity alone as a metric can lead to uninformative results. To address this, we show that generative perplexity and entropy are two components of the KL divergence to a reference distribution. This decomposition explains generative perplexity's sensitivity to entropy, and naturally suggests generative frontiers as a principled method for evaluating model generative quality. We conclude with empirical observations on model quality at this scale. We include a blog post with interactive content to illustrate the argument at https://patrickpynadath1.github.io/blog/eval_methodology/.

Mixture-of-Experts (MoE) models have become a dominant paradigm for scaling large language models, but their rapidly growing parameter sizes introduce a fundamental inefficiency during inference: most expert weights remain idle in GPU memory while competing with performance-critical runtime state such as the key-value (KV) cache. Since KV cache capacity directly determines serving throughput, this mismatch leads to underutilized memory and degraded performance. In this paper, we present FluxMoE, a new MoE inference system that decouples expert parameters from persistent GPU residency. FluxMoE introduces an expert paging abstraction that treats expert weights as streamed, transient resources, materializing them on demand and evicting them immediately after use, allowing GPU memory to be preferentially allocated to throughput-critical runtime state. We implement FluxMoE atop vLLM to enable efficient MoE inference under severe memory constraints. Experimental results demonstrate that FluxMoE achieves up to 3.0$\times$ throughput gains over vLLM in memory-intensive regimes, without compromising model fidelity.

Conversational AI is increasingly deployed in emotionally charged and ethically sensitive interactions. Previous research has primarily concentrated on emotional benchmarks or static safety checks, overlooking how alignment unfolds in evolving conversation. We explore the research question: what breakdowns arise when conversational agents confront emotionally and ethically sensitive behaviors, and how do these affect dialogue quality? To stress-test chatbot performance, we develop a persona-conditioned user simulator capable of engaging in multi-turn dialogue with psychological personas and staged emotional pacing. Our analysis reveals that mainstream models exhibit recurrent breakdowns that intensify as emotional trajectories escalate. We identify several common failure patterns, including affective misalignments, ethical guidance failures, and cross-dimensional trade-offs where empathy supersedes or undermines responsibility. We organize these patterns into a taxonomy and discuss the design implications, highlighting the necessity to maintain ethical coherence and affective sensitivity throughout dynamic interactions. The study offers the HCI community a new perspective on the diagnosis and improvement of conversational AI in value-sensitive and emotionally charged contexts.

Multimodal large language models (MLLMs) have shown strong potential for autonomous driving, yet existing benchmarks remain largely ego-centric and therefore cannot systematically assess model performance in infrastructure-centric and cooperative driving conditions. In this work, we introduce V2X-QA, a real-world dataset and benchmark for evaluating MLLMs across vehicle-side, infrastructure-side, and cooperative viewpoints. V2X-QA is built around a view-decoupled evaluation protocol that enables controlled comparison under vehicle-only, infrastructure-only, and cooperative driving conditions within a unified multiple-choice question answering (MCQA) framework. The benchmark is organized into a twelve-task taxonomy spanning perception, prediction, and reasoning and planning, and is constructed through expert-verified MCQA annotation to enable fine-grained diagnosis of viewpoint-dependent capabilities. Benchmark results across ten representative state-of-the-art proprietary and open-source models show that viewpoint accessibility substantially affects performance, and infrastructure-side reasoning supports meaningful macroscopic traffic understanding. Results also indicate that cooperative reasoning remains challenging since it requires cross-view alignment and evidence integration rather than simply additional visual input. To address these challenges, we introduce V2X-MoE, a benchmark-aligned baseline with explicit view routing and viewpoint-specific LoRA experts. The strong performance of V2X-MoE further suggests that explicit viewpoint specialization is a promising direction for multi-view reasoning in autonomous driving. Overall, V2X-QA provides a foundation for studying multi-perspective reasoning, reliability, and cooperative physical intelligence in connected autonomous driving. The dataset and V2X-MoE resources are publicly available at: https://github.com/junwei0001/V2X-QA.

The formal reasoning capabilities of LLMs are crucial for advancing automated software engineering. However, existing benchmarks for LLMs lack systematic evaluation based on computation and complexity, leaving a critical gap in understanding their formal reasoning capabilities. Therefore, it is still unknown whether SOTA LLMs can grasp the structured, hierarchical complexity of formal languages as defined by Computation Theory. To address this, we introduce ChomskyBench, a benchmark for systematically evaluating LLMs through the lens of Chomsky Hierarchy. Unlike prior work that uses vectorized classification for neural networks, ChomskyBench is the first to combine full Chomsky Hierarchy coverage, process-trace evaluation via natural language, and deterministic symbolic verifiability. ChomskyBench is composed of a comprehensive suite of language recognition and generation tasks designed to test capabilities at each level. Extensive experiments indicate a clear performance stratification that correlates with the hierarchy's levels of complexity. Our analysis reveals a direct relationship where increasing task difficulty substantially impacts both inference length and performance. Furthermore, we find that while larger models and advanced inference methods offer notable relative gains, they face severe efficiency barriers: achieving practical reliability would require prohibitive computational costs, revealing that current limitations stem from inefficiency rather than absolute capability bounds. A time complexity analysis further indicates that LLMs are significantly less efficient than traditional algorithmic programs for these formal tasks. These results delineate the practical limits of current LLMs, highlight the indispensability of traditional software tools, and provide insights to guide the development of future LLMs with more powerful formal reasoning capabilities.

Dynamic languages (such as Python and JavaScript) offer flexibility and simplified type handling for programming, but this can also lead to an increase in type-related errors and additional overhead for compile-time type inference. As a result, type inference for dynamic languages has become a popular research area. Existing approaches typically achieve type inference through static analysis, machine learning, or large language models (LLMs). However, current work only focuses on the direct dependencies of variables related to type inference as the context, resulting in incomplete contextual information and thus affecting the accuracy of type inference. To address this issue, this paper proposes a method called TypePro, which leverages LLMs for type inference in dynamic languages. TypePro supplements contextual information by conducting inter-procedural code slicing. Then, TypePro proposes a set of candidate complex types based on the structural information of data types implied in the slices, thereby addressing the lack of domain knowledge of LLMs. We conducted experiments on the ManyTypes4Py and ManyTypes4TypeScript datasets, achieving Top-1 exact match (EM) rates of 88.9% and 86.6%, respectively. Notably, TypePro improves the Top-1 Exact Match by 7.1 and 10.3 percentage points over the second-best approach, showing the effectiveness and robustness of TypePro.

A previous study reported that E-Prime (English without the verb "to be") selectively altered reasoning in language models, with cross-model correlations suggesting a structural signature tied to which vocabulary was removed. I designed a replication with active controls to test the proposed mechanism: cognitive restructuring through specific vocabulary-cognition mappings. The experiment tested five conditions (unconstrained control, E-Prime, No-Have, elaborated metacognitive prompt, neutral filler-word ban) across six models and seven reasoning tasks (N=15,600 trials, 11,919 after compliance filtering). Every prediction from the cognitive restructuring hypothesis was disconfirmed. All four treatments outperformed the control (83.0%), including both active controls predicted to show null effects. The neutral filler-word ban, banning words like "very" and "just" with no role in logical inference, produced the largest improvement (+6.7 pp), while E-Prime produced the smallest (+3.7 pp). The four conditions ranked in perfect inverse order of theoretical depth. The cross-model correlation signature did not replicate (mean r=0.005). These results are consistent with a simpler mechanism: any constraint that forces a model off its default generation path acts as an output regularizer, improving reasoning by disrupting fluent but shallow response patterns. The shallowest constraints work best because they impose monitoring load with minimal conceptual disruption. I present these findings as a case study in discovery through disconfirmation.

Quantum Machine Learning (QML) is fundamentally limited by two challenges: barren plateaus (exponentially vanishing gradients) and the fragility of parameterized quantum circuits under noise. Despite extensive empirical studies, a unified theoretical framework remains lacking. We introduce LieTrunc-QNN, an algebraic-geometric framework that characterizes trainability via Lie-generated dynamics. Parameterized quantum circuits are modeled as Lie subalgebras of u(2^n), whose action induces a Riemannian manifold of reachable quantum states. Expressivity is reinterpreted as intrinsic manifold dimension and geometry. We establish a geometric capacity-plateau principle: increasing effective dimension leads to exponential gradient suppression due to concentration of measure. By restricting to structured Lie subalgebras (LieTrunc), the manifold is contracted, preventing concentration and preserving non-degenerate gradients. We prove two main results: (1) a trainability lower bound for LieTrunc-QNN, and (2) that the Fubini-Study metric rank is bounded by the algebraic span of generators, showing expressivity is governed by structure rather than parameter count. Compact Lie subalgebras also provide inherent robustness to perturbations. Importantly, we establish a polynomial trainability regime where gradient variance decays polynomially instead of exponentially. Experiments (n=2-6) validate the theory: LieTrunc-QNN maintains stable gradients and high effective dimension, while random truncation leads to metric rank collapse. At n=6, full metric rank is preserved (rank=16). Results support a scaling law between gradient variance and effective dimension. This work provides a unified geometric framework for QNN design, linking Lie algebra, manifold geometry, and optimization.

The rapid progress of generative AI has enabled increasingly realistic text-centric image forgeries, posing major challenges to document safety. Existing forensic methods mainly rely on visual cues and lack evidence-based reasoning to reveal subtle text manipulations. Detection, localization, and explanation are often treated as isolated tasks, limiting reliability and interpretability. To tackle these challenges, we propose DocShield, the first unified framework formulating text-centric forgery analysis as a visual-logical co-reasoning problem. At its core, a novel Cross-Cues-aware Chain of Thought (CCT) mechanism enables implicit agentic reasoning, iteratively cross-validating visual anomalies with textual semantics to produce consistent, evidence-grounded forensic analysis. We further introduce a Weighted Multi-Task Reward for GRPO-based optimization, aligning reasoning structure, spatial evidence, and authenticity prediction. Complementing the framework, we construct RealText-V1, a multilingual dataset of document-like text images with pixel-level manipulation masks and expert-level textual explanations. Extensive experiments show DocShield significantly outperforms existing methods, improving macro-average F1 by 41.4% over specialized frameworks and 23.4% over GPT-4o on T-IC13, with consistent gains on the challenging T-SROIE benchmark. Our dataset, model, and code will be publicly released.

Deep learning based semantic communication has achieved significant progress in wireless image transmission, but most existing schemes rely on fixed models and thus lack robustness to diverse image contents and dynamic channel conditions. To improve adaptability, recent studies have developed adaptive semantic communication strategies that adjust transmission or model behavior according to either source content or channel state. More recently, MoE-based semantic communication has emerged as a sparse and efficient adaptive architecture, although existing designs still mainly rely on single-driven routing. To address this limitation, we propose a novel multi-stage end-to-end image semantic communication system for multi-input multi-output (MIMO) channels, built upon an adaptive MoE Swin Transformer block. Specifically, we introduce a dynamic expert gating mechanism that jointly evaluates both real-time CSI and the semantic content of input image patches to compute adaptive routing probabilities. By selectively activating only a specialized subset of experts based on this joint condition, our approach breaks the rigid coupling of traditional adaptive methods and overcomes the bottlenecks of single-driven routing. Simulation results indicate a significant improvement in reconstruction quality over existing methods while maintaining the transmission efficiency.

Recent advances in Multimodal Large Language Models (MLLMs) have expanded reasoning capabilities into 3D domains, enabling fine-grained spatial understanding. However, the substantial size of 3D MLLMs and the high dimensionality of input features introduce considerable inference overhead, which limits practical deployment on resource constrained platforms. To overcome this limitation, this paper presents Efficient3D, a unified framework for visual token pruning that accelerates 3D MLLMs while maintaining competitive accuracy. The proposed framework introduces a Debiased Visual Token Importance Estimator (DVTIE) module, which considers the influence of shallow initial layers during attention aggregation, thereby producing more reliable importance predictions for visual tokens. In addition, an Adaptive Token Rebalancing (ATR) strategy is developed to dynamically adjust pruning strength based on scene complexity, preserving semantic completeness and maintaining balanced attention across layers. Together, they enable context-aware token reduction that maintains essential semantics with lower computation. Comprehensive experiments conducted on five representative 3D vision and language benchmarks, including ScanRefer, Multi3DRefer, Scan2Cap, ScanQA, and SQA3D, demonstrate that Efficient3D achieves superior performance compared with unpruned baselines, with a +2.57% CIDEr improvement on the Scan2Cap dataset. Therefore, Efficient3D provides a scalable and effective solution for efficient inference in 3D MLLMs. The code is released at: https://github.com/sol924/Efficient3D

Existing LLM agents for computational materials science are constrained by pipeline-bounded architectures tied to specific simulation codes and by dependence on manually written tool functions that grow with task scope. We present MatClaw, a code-first agent that writes and executes Python directly, composing any installed domain library to orchestrate multi-code workflows on remote HPC clusters without predefined tool functions. To sustain coherent execution across multi-day workflows, MatClaw uses a four-layer memory architecture that prevents progressive context loss, and retrieval-augmented generation over domain source code that raises per-step API-call accuracy to ${\sim}$99 %. Three end-to-end demonstrations on ferroelectric CuInP2S6 (machine-learning force field training via active learning, Curie temperature prediction, and heuristic parameter-space search) reveal that the agent handles code generation reliably but struggles with tacit domain knowledge. The missing knowledge, such as appropriate simulation timescales, equilibration protocols, and sampling strategies, is the kind that researchers accumulate through experience but rarely formalize. Two lightweight interventions, literature self-learning and expert-specified constraints, bridge these gaps, defining a guided autonomy model in which the researcher provides high-level domain knowledge while the agent handles workflow execution. Our results demonstrate that the gap between guided and fully autonomous computational materials research is narrower than ever before: LLMs already handle code generation and scientific interpretation reliably, and the rapid improvement in their capabilities will accelerate materials discovery beyond what manual workflows can achieve. All code and benchmarks are open-source.

Reward models (RMs) are widely used as optimization targets in reinforcement learning from human feedback (RLHF), yet they remain vulnerable to reward hacking. Existing attacks mainly operate within the semantic space, constructing human-readable adversarial outputs that exploit RM biases. In this work, we introduce a fundamentally different paradigm: Token Mapping Perturbation Attack (TOMPA), a framework that performs adversarial optimization directly in token space. By bypassing the standard decode-re-tokenize interface between the policy and the reward model, TOMPA enables the attack policy to optimize over raw token sequences rather than coherent natural language. Using only black-box scalar feedback, TOMPA automatically discovers non-linguistic token patterns that elicit extremely high rewards across multiple state-of-the-art RMs. Specifically, when targeting Skywork-Reward-V2-Llama-3.1-8B, TOMPA nearly doubles the reward of GPT-5 reference answers and outperforms them on 98.0% of prompts. Despite these high scores, the generated outputs degenerate into nonsensical text, revealing that RMs can be systematically exploited beyond the semantic regime and exposing a critical vulnerability in current RLHF pipelines.

Understanding the geometric structure of internal representations is a central goal of mechanistic interpretability. Prior work has shown that transformers trained on sequences generated by hidden Markov models encode probabilistic belief states as simplex-shaped geometries in their residual stream, with vertices corresponding to latent generative states. Whether large language models trained on naturalistic text develop analogous geometric representations remains an open question. We introduce a pipeline for discovering candidate simplex-structured subspaces in transformer representations, combining sparse autoencoders (SAEs), $k$-subspace clustering of SAE features, and simplex fitting using AANet. We validate the pipeline on a transformer trained on a multipartite hidden Markov model with known belief-state geometry. Applied to Gemma-2-9B, we identify 13 priority clusters exhibiting candidate simplex geometry ($K \geq 3$). A key challenge is distinguishing genuine belief-state encoding from tiling artifacts: latents can span a simplex-shaped subspace without the mixture coordinates carrying predictive signal beyond any individual feature. We therefore adopt barycentric prediction as our primary discriminating test. Among the 13 priority clusters, 3 exhibit a highly significant advantage on near-vertex samples (Wilcoxon $p < 10^{-14}$) and 4 on simplex-interior samples. Together 5 distinct real clusters pass at least one split, while no null cluster passes either. One cluster, 768_596, additionally achieves the highest causal steering score in the dataset. This is the only case where passive prediction and active intervention converge. We present these findings as preliminary evidence that genuine belief-like geometry exists in Gemma-2-9B's representation space, and identify the structured evaluation that would be required to confirm this interpretation.

Clinical evidence synthesis requires identifying relevant trials from large registries and aggregating results that account for population differences. While recent LLM-based approaches have automated components of systematic review, they do not support end-to-end evidence synthesis. Moreover, conventional meta-analysis weights studies by statistical precision without considering clinical compatibility reflected in eligibility criteria. We propose EligMeta, an agentic framework that integrates automated trial discovery with eligibility-aware meta-analysis, translating natural-language queries into reproducible trial selection and incorporating eligibility alignment into study weighting to produce cohort-specific pooled estimates. EligMeta employs a hybrid architecture separating LLM-based reasoning from deterministic execution: LLMs generate interpretable rules from natural-language queries and perform schema-constrained parsing of trial metadata, while all logical operations, weight computations, and statistical pooling are executed deterministically to ensure reproducibility. The framework structures eligibility criteria and computes similarity-based study weights reflecting population alignment between target and comparator trials. In a gastric cancer landscape analysis, EligMeta reduced 4,044 candidate trials to 39 clinically relevant studies through rule-based filtering, recovering all 13 guideline-cited trials. In an olaparib adverse events meta-analysis across four trials, eligibility-aware weighting shifted the pooled risk ratio from 2.18 (95% CI: 1.71-2.79) under conventional Mantel-Haenszel estimation to 1.97 (95% CI: 1.76-2.20), demonstrating quantifiable impact of incorporating eligibility alignment. EligMeta bridges automated trial discovery with eligibility-aware meta-analysis, providing a scalable and reproducible framework for evidence synthesis in precision medicine.

Large Language Model (LLM) multi-agent systems are increasingly deployed as interacting agent societies, yet scaling these systems often yields diminishing or unstable returns, the causes of which remain poorly understood. We present the first large-scale empirical study of coordination dynamics in LLM-based multi-agent systems, introducing an atomic event-level formulation that reconstructs reasoning as cascades of coordination. Analyzing over 1.5 Million interactions across tasks, topologies, and scales, we uncover three coupled laws: coordination follows heavy-tailed cascades, concentrates via preferential attachment into intellectual elites, and produces increasingly frequent extreme events as system size grows. We show that these effects are coupled through a single structural mechanism: an integration bottleneck, in which coordination expansion scales with system size while consolidation does not, producing large but weakly integrated reasoning processes. To test this mechanism, we introduce Deficit-Triggered Integration (DTI), which selectively increases integration under imbalance. DTI improves performance precisely where coordination fails, without suppressing large-scale reasoning. Together, our results establish quantitative laws of collective cognition and identify coordination structure as a fundamental, previously unmeasured axis for understanding and improving scalable multi-agent intelligence.

Muscle fatigue detection plays an important role in physical rehabilitation. Previous researches have demonstrated that sEMG offers superior sensitivity in detecting muscle fatigue compared to other biological signals. However, features extracted from sEMG may vary during dynamic contractions and across different subjects, which causes unstability in fatigue detection. To address these challenges, this research proposes a novel neural network comprising an Inception-attention module as a feature extractor, a fatigue classifier and a domain classifier equipped with a gradient reversal layer. The integrated domain classifier encourages the network to learn subject-invariant common fatigue features while minimizing subject-specific features. Furthermore, a supervised contrastive loss function is also employed to enhance the generalization capability of the model. Experimental results demonstrate that the proposed model achieved outstanding performance in three-class classification tasks, reaching 93.54% accuracy, 92.69% recall and 92.69% F1-score, providing a robust solution for cross-subject muscle fatigue detection, offering significant guidance for rehabilitation training and assistance.

How biased is a language model? The answer depends on how you ask. A model that refuses to choose between castes for a leadership role will, in a fill-in-the-blank task, reliably associate upper castes with purity and lower castes with lack of hygiene. Single-task benchmarks miss this because they capture only one slice of a model's bias profile. We introduce a hierarchical taxonomy covering 9 bias types, including under-studied axes like caste, linguistic, and geographic bias, operationalized through 7 evaluation tasks that span explicit decision-making to implicit association. Auditing 7 commercial and open-weight LLMs with \textasciitilde45K prompts, we find three systematic patterns. First, bias is task-dependent: models counter stereotypes on explicit probes but reproduce them on implicit ones, with Stereotype Score divergences up to 0.43 between task types for the same model and identity groups. Second, safety alignment is asymmetric: models refuse to assign negative traits to marginalized groups, but freely associate positive traits with privileged ones. Third, under-studied bias axes show the strongest stereotyping across all models, suggesting alignment effort tracks benchmark coverage rather than harm severity. These results demonstrate that single-benchmark audits systematically mischaracterize LLM bias and that current alignment practices mask representational harm rather than mitigating it.

Large language models (LLMs) often exhibit sycophancy: agreement with user stance even when it conflicts with the model's opinion. While prior work has mostly studied this in single-agent settings, it remains underexplored in collaborative multi-agent systems. We ask whether awareness of other agents' sycophancy levels influences discussion outcomes. To investigate this, we run controlled experiments with six open-source LLMs, providing agents with peer sycophancy rankings that estimate each peer's tendency toward sycophancy. These rankings are based on scores calculated using various static (pre-discussion) and dynamic (online) strategies. We find that providing sycophancy priors reduces the influence of sycophancy-prone peers, mitigates error-cascades, and improves final discussion accuracy by an absolute 10.5%. Thus, this is a lightweight, effective way to reduce discussion sycophancy and improve downstream accuracy.

Optimization is as much about modeling the right problem as solving it. Identifying the right objectives, constraints, and trade-offs demands extensive interaction between researchers and stakeholders. Large language models can empower decision-makers with optimization capabilities through interactive optimization agents that can propose, interpret and refine solutions. However, it is fundamentally harder to evaluate a conversation-based interaction than traditional one-shot approaches. This paper proposes a scalable and replicable methodology for evaluating optimization agents through conversations. We build LLM-powered decision agents that role-play diverse stakeholders, each governed by an internal utility function but communicating like a real decision-maker. We generate thousands of conversations in a school scheduling case study. Results show that one-shot evaluation is severely limiting: the same optimization agent converges to much higher-quality solutions through conversations. Then, this paper uses this methodology to demonstrate that tailored optimization agents, endowed with domain-specific prompts and structured tools, can lead to significant improvements in solution quality in fewer interactions, as compared to general-purpose chatbots. These findings provide evidence of the benefits of emerging solutions at the AI-optimization interface to expand the reach of optimization technologies in practice. They also uncover the impact of operations research expertise to facilitate interactive deployments through the design of effective and reliable optimization agents.

The SZZ algorithm is the dominant technique for identifying bug-inducing commits and underpins many software engineering tasks, such as defect prediction and vulnerability analysis. Despite numerous variants, including recent LLM-based approaches, performance remains limited on developer-annotated datasets (e.g., recall of 0.552 on the Linux kernel). A key limitation is the reliance on git blame, which traces line-level changes within the same file, failing in common scenarios such as ghost and cross-file cases-making nearly one-quarter of bug-inducing commits inherently untraceable. Moreover, current approaches follow fixed pipelines that restrict iterative reasoning and exploration, unlike developers who investigate bugs through an interactive, multi-tool process. To address these challenges, we propose AgentSZZ, an agent-based framework that leverages LLM-driven agents to explore repositories and identify bug-inducing commits. Unlike prior methods, AgentSZZ integrates task-specific tools, domain knowledge, and a ReAct-style loop to enable adaptive and causal tracing of bugs. A structured compression module further improves efficiency by reducing redundant context while preserving key evidence. Extensive experiments on three widely used datasets show that AgentSZZ consistently outperforms state-of-the-art SZZ algorithms across all settings, achieving F1-score gains of up to 27.2% over prior LLM-based approaches. The improvements are especially pronounced in challenging scenarios such as cross-file and ghost commits, with recall gains of up to 300% and 60%, respectively. Ablation studies show that task-specific tools and domain knowledge are critical, while compression tool outputs reduce token consumption by over 30% with negligible impact. The replication package is available.

Severe accident analysis using system-level codes such as MELCOR is indispensable for nuclear safety assessment, yet the computational cost of repeated simulations poses a significant bottleneck for parametric studies and uncertainty quantification. Existing surrogate models accelerate these analyses but depend on large volumes of simulation data, while physics-informed neural networks (PINNs) enable data-free training but must be retrained for every change in problem parameters. This study addresses both limitations by developing the Parameterized PINNs coupled with FDM (P2F) method, a node-assigned hybrid framework for MELCOR's Control Volume Hydrodynamics/Flow Path (CVH/FP) module. In the P2F method, a parameterized Node-Assigned PINN (NA-PINN) accepts the water-level difference, initial velocity, and time as inputs, learning a solution manifold so that a single trained network serves as a data-free surrogate for the momentum conservation equation across all flow paths without retraining. This PINN is coupled with a finite difference method (FDM) solver that advances the mass conservation equation at each time step, ensuring exact discrete mass conservation while replacing the iterative nonlinear momentum solve with a single forward pass. Verification on a six-tank gravity-driven draining scenario yields a water level mean absolute error of $7.85 \times 10^{-5}$ m and a velocity mean absolute error of $3.21 \times 10^{-3}$ m/s under the nominal condition with $Δt = 1.0$ s. The framework maintains consistent accuracy across time steps ranging from 0.2 to 1.0 s and generalizes to five distinct initial conditions, all without retraining or simulation data. This work introduces a numerical coupling methodology for integrating parameterized PINNs with FDM within a nuclear thermal-hydraulic system code framework.

As Large Language Models (LLMs) increasingly power decision-making systems across critical domains, understanding and mitigating their biases becomes essential for responsible AI deployment. Although bias assessment frameworks have proliferated for attributes such as race and gender, socioeconomic status bias remains significantly underexplored despite its widespread implications in the real world. We introduce SocioEval, a template-based framework for systematically evaluating socioeconomic bias in foundation models through decision-making tasks. Our hierarchical framework encompasses 8 themes and 18 topics, generating 240 prompts across 6 class-pair combinations. We evaluated 13 frontier LLMs on 3,120 responses using a rigorous three-stage annotation protocol, revealing substantial variation in bias rates (0.42\%-33.75\%). Our findings demonstrate that bias manifests differently across themes lifestyle judgments show 10$\times$ higher bias than education-related decisions and that deployment safeguards effectively prevent explicit discrimination but show brittleness to domain-specific stereotypes. SocioEval provides a scalable, extensible foundation for auditing class-based bias in language models.

The massive scale of pretrained models has made efficient compression essential for practical deployment. Low-rank decomposition based on the singular value decomposition (SVD) provides a principled approach for model reduction, but its exact computation is expensive for large weight matrices. Randomized alternatives such as randomized SVD (RSVD) improve efficiency, yet they can suffer from poor approximation quality when the singular value spectrum decays slowly, a regime commonly observed in modern pretrained models. In this work, we address this limitation from both theoretical and empirical perspectives. First, we establish a connection between low-rank approximation error and predictive performance by analyzing softmax perturbations, showing that deviations in class probabilities are controlled by the spectral error of the compressed weights. Second, we demonstrate that RSVD is inadequate, and we propose randomized subspace iteration (RSI) as a more effective alternative. By incorporating multiple power iterations, RSI improves spectral separation and provides a controllable mechanism for enhancing approximation quality. We evaluate our approach on both convolutional networks and transformer-based architectures. Our results show that RSI achieves near-optimal approximation quality while outperforming RSVD in predictive accuracy under aggressive compression, enabling efficient model compression.

Randomized controlled trials often do not represent the populations where decisions are made, and covariate shift across studies can invalidate standard IPD meta-analysis and transport estimators. We propose a placebo-anchored transport framework that treats source-trial outcomes as abundant proxy signals and target-trial placebo outcomes as scarce, high-fidelity gold labels to calibrate baseline risk. A low-complexity (sparse) correction anchors proxy outcome models to the target population, and the anchored models are embedded in a cross-fitted doubly robust learner, yielding a Neyman-orthogonal, target-site doubly robust estimator for patient-level heterogeneous treatment effects when target treated outcomes are available. We distinguish two regimes: in connected targets (with a treated arm), the method yields target-identified effect estimates; in disconnected targets (placebo-only), it reduces to a principled screen--then--transport procedure under explicit working-model transport assumptions. Experiments on synthetic data and a semi-synthetic IHDP benchmark evaluate pointwise CATE accuracy, ATE error, ranking quality for targeting, decision-theoretic policy regret, and calibration. Across connected settings, the proposed method is best or near-best and improves substantially over proxy-only, target-only, and transport baselines at small target sample sizes; in disconnected settings, it retains strong ranking performance for targeting while pointwise accuracy depends on the strength of the working transport condition.

Empirically, modern deep learning training often occurs at the Edge of Stability (EoS), where the sharpness of the loss exceeds the threshold below which classical convergence analysis applies. Despite recent progress, existing theoretical explanations of EoS either rely on restrictive assumptions or focus on specific squared-loss-type objectives. In this work, we introduce and study a structural property of loss functions that we term product-stability. We show that for losses with product-stable minima, gradient descent applied to objectives of the form $(x,y) \mapsto l(xy)$ can provably converge to the local minimum even when training in the EoS regime. This framework substantially generalizes prior results and applies to a broad class of losses, including binary cross entropy. Using bifurcation diagrams, we characterize the resulting training dynamics, explain the emergence of stable oscillations, and precisely quantify the sharpness at convergence. Together, our results offer a principled explanation for stable EoS training for a wider class of loss functions.

The safety of large language models (LLMs) relies on alignment techniques such as reinforcement learning from human feedback (RLHF). However, recent theoretical analyses suggest that reinforcement learning-based training does not acquire new capabilities but merely redistributes the utilization probabilities of existing ones. In this study, we propose ``compound jailbreaks'' targeting OpenAI gpt-oss-20b, which exploit the generalization failures of alignment. This approach combines multiple attack techniques -- each individually defended against -- to saturate the instruction hierarchy maintenance process. Our evaluation shows that the attack success rate (ASR) increased from 14.3\% with individual methods to 71.4\% with the combined approach. These results provide empirical evidence for the hypothesis that safety training does not generalize as broadly as model capabilities, highlighting the need for multifaceted safety evaluations using compound attack scenarios.

Graph neural networks (GNNs) are widely used for learning on graph datasets derived from various real-world scenarios. Learning from extremely large graphs requires distributed training, and mini-batching with sampling is a popular approach for parallelizing GNN training. Existing distributed mini-batch approaches have significant performance bottlenecks due to expensive sampling methods and limited scaling when using data parallelism. In this work, we present ScaleGNN, a 4D parallel framework for scalable mini-batch GNN training that combines communication-free distributed sampling, 3D parallel matrix multiplication (PMM), and data parallelism. ScaleGNN introduces a uniform vertex sampling algorithm, enabling each process (GPU device) to construct its local mini-batch, i.e., subgraph partitions without any inter-process communication. 3D PMM enables scaling mini-batch training to much larger GPU counts than vanilla data parallelism with significantly lower communication overheads. We also present additional optimizations to overlap sampling with training, reduce communication overhead by sending data in lower precision, kernel fusion, and communication-computation overlap. We evaluate ScaleGNN on five graph datasets and demonstrate strong scaling up to 2048 GPUs on Perlmutter, 2048 GCDs on Frontier, and 1024 GPUs on Tuolumne. On Perlmutter, ScaleGNN achieves 3.5x end-to-end training speedup over the SOTA baseline on ogbn-products.

Existing studies on Long-Context Continual Pre-training (LCCP) mainly focus on small-scale models and limited data regimes (tens of billions of tokens). We argue that directly migrating these small-scale settings to industrial-grade models risks insufficient adaptation and premature training termination. Furthermore, current evaluation methods rely heavily on downstream benchmarks (e.g., Needle-in-a-Haystack), which often fail to reflect the intrinsic convergence state and can lead to "deceptive saturation". In this paper, we present the first systematic investigation of LCCP learning dynamics using the industrial-grade Hunyuan-A13B (80B total parameters), tracking its evolution across a 200B-token training trajectory. Specifically, we propose a hierarchical framework to analyze LCCP dynamics across behavioral (supervised fine-tuning probing), probabilistic (perplexity), and mechanistic (attention patterns) levels. Our findings reveal: (1) Necessity of Massive Data Scaling: Training regimes of dozens of billions of tokens are insufficient for industrial-grade LLMs' LCCP (e.g., Hunyuan-A13B reaches saturation after training over 150B tokens). (2) Deceptive Saturation vs. Intrinsic Saturation: Traditional NIAH scores report "fake saturation" early, while our PPL-based analysis reveals continuous intrinsic improvements and correlates more strongly with downstream performance. (3) Mechanistic Monitoring for Training Stability: Retrieval heads act as efficient, low-resource training monitors, as their evolving attention scores reliably track LCCP progress and exhibit high correlation with SFT results. This work provides a comprehensive monitoring framework, evaluation system, and mechanistic interpretation for the LCCP of industrial-grade LLM.

The autonomous discovery of bugs remains a significant challenge in modern software development. Compared to code generation, the complexity of dynamic runtime environments makes bug discovery considerably harder for large language models (LLMs). In this paper, we take game development as a representative domain and introduce the Game Benchmark for Quality Assurance (GBQA), a benchmark containing 30 games and 124 human-verified bugs across three difficulty levels, to evaluate whether LLMs can autonomously detect software bugs. The benchmark is constructed using a multi-agent system that develops games and injects bugs in a scalable manner, with human experts in the loop to ensure correctness. Moreover, we provide a baseline interactive agent equipped with a multi-round ReAct loop and a memory mechanism, enabling long-horizon exploration of game environments for bug detection across different LLMs. Extensive experiments on frontier LLMs demonstrate that autonomous bug discovery remains highly challenging: the best-performing model, Claude-4.6-Opus in thinking mode, identifies only 48.39% of the verified bugs. We believe GBQA provides an adequate testbed and evaluation criterion, and that further progress on it will help close the gap in autonomous software engineering.

Automated Program Repair (APR) struggles with complex logic errors and silent failures. Current LLM-based APR methods are mostly static, relying on source code and basic test outputs, which fail to accurately capture complex runtime behaviors and dynamic data dependencies. While incorporating runtime evidence like execution traces exposes concrete state transitions, a single LLM interpreting this in isolation often overfits to specific hypotheses, producing patches that satisfy tests by coincidence rather than correct logic. Therefore, runtime evidence should act as objective constraints rather than mere additional input. We propose TraceRepair, a multi-agent framework that leverages runtime facts as shared constraints for patch validation. A probe agent captures execution snapshots of critical variables to form an objective repair basis. Meanwhile, a committee of specialized agents cross-verifies candidate patches to expose inconsistencies and iteratively refine them. Evaluated on the Defects4J benchmark, TraceRepair correctly fixes 392 defects, substantially outperforming existing LLM-based approaches. Extensive experiments demonstrate improved efficiency and strong generalization on a newly constructed dataset of recent bugs, confirming that performance gains arise from dynamic reasoning rather than memorization.

This work aims to shine a spotlight on the topic of metalanguage. We first define metalanguage, link it to NLP and LLMs, and then discuss our two labs' metalanguage-centered efforts. Finally, we discuss four dimensions of metalanguage and metalinguistic tasks, offering a list of understudied future research directions.

We present Conditional Wasserstein Autoencoders (CWAEs), a framework for conditional simulation that exploits low-dimensional structure in both the conditioned and the conditioning variables. The key idea is to modify a Wasserstein autoencoder to use a (block-) triangular decoder and impose an appropriate independence assumption on the latent variables. We show that the resulting model gives an autoencoder that can exploit low-dimensional structure while simultaneously the decoder can be used for conditional simulation. We explore various theoretical properties of CWAEs, including their connections to conditional optimal transport (OT) problems. We also present alternative formulations that lead to three architectural variants forming the foundation of our algorithms. We present a series of numerical experiments that demonstrate that our different CWAE variants achieve substantial reductions in approximation error relative to the low-rank ensemble Kalman filter (LREnKF), particularly in problems where the support of the conditional measures is truly low-dimensional.

Performance evaluation of Retrieval-Augmented Generation (RAG) systems within enterprise environments is governed by multi-dimensional and composite factors extending far beyond simple final accuracy checks. These factors include reasoning complexity, retrieval difficulty, the diverse structure of documents, and stringent requirements for operational explainability. Existing academic benchmarks fail to systematically diagnose these interlocking challenges, resulting in a critical gap where models achieving high performance scores fail to meet the expected reliability in practical deployment. To bridge this discrepancy, this research proposes a multi-dimensional diagnostic framework by defining a four-axis difficulty taxonomy and integrating it into an enterprise RAG benchmark to diagnose potential system weaknesses.

Surround depth estimation provides a cost-effective alternative to LiDAR for 3D perception in autonomous driving. While recent self-supervised methods explore multi-camera settings to improve scale awareness and scene coverage, they are primarily designed for passenger vehicles and rarely consider articulated vehicles or robotics platforms. The articulated structure introduces complex cross-segment geometry and motion coupling, making consistent depth reasoning across views more challenging. In this work, we propose \textbf{ArticuSurDepth}, a self-supervised framework for surround-view depth estimation on articulated vehicles that enhances depth learning through cross-view and cross-vehicle geometric consistency guided by structural priors from vision foundation model. Specifically, we introduce multi-view spatial context enrichment strategy and a cross-view surface normal constraint to improve structural coherence across spatial and temporal contexts. We further incorporate camera height regularization with ground plane-awareness to encourage metric depth estimation, together with cross-vehicle pose consistency that bridges motion estimation between articulated segments. To validate our proposed method, an articulated vehicle experiment platform was established with a dataset collected over it. Experiment results demonstrate state-of-the-art (SoTA) performance of depth estimation on our self-collected dataset as well as on DDAD, nuScenes, and KITTI benchmarks.

We propose AXELRAM, a smart SRAM macro architecture that computes attention scores directly from quantized KV cache indices without dequantization. The key enabler is a design-time fixed codebook: orthogonal-transform-based quantization concentrates each coordinate's distribution to N(0,1/d), so the optimal quantizer depends only on dimension d and bit-width b, not on input data. The asymmetric path design -- transform on write, table-lookup on read with no inverse transform -- reduces per-query multiplications by 102.4x (a mathematical identity). Through multi-seed evaluation (10 seeds x 3 models), we discover that sign pattern sensitivity causes catastrophic PPL spikes (Delta > 50) on certain models (Qwen2.5-3B), while others (LLaMA-3.1-8B) are fully stable. This phenomenon extends SpinQuant's observation of rotation variance in weight quantization to the KV cache domain, where the effect is qualitatively more severe. We trace the root cause to layer-wise norm heterogeneity and propose a gradient-free sign pattern selection (200 candidates, 8 calibration samples, one-time) that eliminates catastrophic spikes with zero additional hardware cost. All source code is available at https://github.com/Axelidea/AXELRAM.

As large language models (LLMs) become increasingly persuasive, there is concern that people's opinions and decisions may be influenced across various contexts at scale. Prior mitigation (e.g., AI detectors and disclaimers) largely treats people as passive recipients of AI-generated information. To provide a more proactive intervention against persuasive AI, we introduce $\textbf{LLMimic}$, a role-play-based, interactive, gamified AI literacy tutorial, where participants assume the role of an LLM and progress through three key stages of the training pipeline (pretraining, SFT, and RLHF). We conducted a $2 \times 3$ between-subjects study ($N = 274$) where participants either (1) watched an AI history video (control) or (2) interacted with LLMimic (treatment), and then engaged in one of three realistic AI persuasion scenarios: (a) charity donation persuasion, (b) malicious money solicitation, or (c) hotel recommendation. Our results show that LLMimic significantly improved participants' AI literacy ($p < .001$), reduced persuasion success across scenarios ($p < .05$), and enhanced truthfulness and social responsibility levels ($p<0.01$) in the hotel scenario. These findings suggest that LLMimic offers a scalable, human-centered approach to improving AI literacy and supporting more informed interactions with persuasive AI.

Non-Exemplar Continual Graph Learning (NECGL) seeks to eliminate the privacy risks intrinsic to rehearsal-based paradigms by retaining solely class-level prototype representations rather than raw graph examples for mitigating catastrophic forgetting. However, this design choice inevitably precipitates feature drift. As a nascent alternative, Analytic Continual Learning (ACL) capitalizes on the intrinsic generalization properties of frozen pre-trained models to bolster continual learning performance. Nonetheless, a key drawback resides in the pronounced attenuation of model plasticity. To surmount these challenges, we propose Analytic Drift Resister (ADR), a novel and theoretically grounded NECGL framework. ADR exploits iterative backpropagation to break free from the frozen pre-trained constraint, adapting to evolving task graph distributions and fortifying model plasticity. Since parameter updates trigger feature drift, we further propose Hierarchical Analytic Merging (HAM), performing layer-wise merging of linear transformations in Graph Neural Networks (GNNs) via ridge regression, thereby ensuring absolute resistance to feature drift. On this basis, Analytic Classifier Reconstruction (ACR) enables theoretically zero-forgetting class-incremental learning. Empirical evaluation on four node classification benchmarks demonstrates that ADR maintains strong competitiveness against existing state-of-the-art methods.

Generative AI (genAI) is increasingly being integrated into children's everyday lives, not only through screens but also through so-called "screen-free" AI toys. These toys can simulate emotions, personalize responses, and recall prior interactions, creating the illusion of an ongoing social connection. Such capabilities raise important questions about how children understand boundaries, agency, and relationships when interacting with AI toys. To investigate this, we conducted two participatory design sessions with eight children ages 6-11 where they engaged with three different AI toys, shifting between play, experimentation, and reflection. Our findings reveal that children approached AI toys with genuine curiosity, profiling them as social beings. However, frequent interaction breakdowns and mismatches between apparent intelligence and toy-like form disrupted expectations around play and led to adversarial play. We conclude with implications and design provocations to navigate children's encounters with AI toys in more transparent, developmentally appropriate, and responsible ways.

Living in a changing climate, human society now faces more frequent and severe natural disasters than ever before. As a consequence, rapid disaster response during the "Golden 72 Hours" of search and rescue becomes a vital humanitarian necessity and community concern. However, traditional disaster damage surveys routinely fail to generalize across distinct urban morphologies and new disaster events. Effective damage mapping typically requires exhaustive and time-consuming manual data annotation. To address this issue, we introduce Smart Transfer, a novel Geospatial Artificial Intelligence (GeoAI) framework, leveraging state-of-the-art vision Foundation Models (FMs) for rapid building damage mapping with post-earthquake Very High Resolution (VHR) imagery. Specifically, we design two novel model transfer strategies: first, Pixel-wise Clustering (PC), ensuring robust prototype-level global feature alignment; second, a Distance-Penalized Triplet (DPT), integrating patch-level spatial autocorrelation patterns by assigning stronger penalties to semantically inconsistent yet spatially adjacent patches. Extensive experiments and ablations from the recent 2023 Turkiye-Syria earthquake show promising performance in multiple cross-region transfer settings, namely Leave One Domain Out (LODO) and Specific Source Domain Combination (SSDC). Moreover, Smart Transfer provides a scalable, automated GeoAI solution to accelerate building damage mapping and support rapid disaster response, offering new opportunities to enhance disaster resilience in climate-vulnerable regions and communities. The data and code are publicly available at https://github.com/ai4city-hkust/SmartTransfer.

We introduce a wavelength-multiplexed massively parallel diffractive information storage platform composed of dielectric surfaces that are structurally optimized at the wavelength scale using deep learning to store and project thousands of distinct image patterns, each assigned to a unique wavelength. Through numerical simulations in the visible spectrum, we demonstrated that our wavelength-multiplexed diffractive system can store and project over 4,000 independent desired images/patterns within its output field-of-view, with high image quality and minimal crosstalk between spectral channels. Furthermore, in a proof-of-concept experiment, we demonstrated a two-layer diffractive design that stored six distinct patterns and projected them onto the same output field of view at six different wavelengths (500, 548, 596, 644, 692, and 740 nm). This diffractive architecture is scalable and can operate at various parts of the electromagnetic spectrum without the need for material dispersion engineering or redesigning its optimized diffractive layers. The demonstrated storage capacity, reconstruction image fidelity, and wavelength-encoded massively parallel read-out of our diffractive platform offer a compact and fast-access solution for large-scale optical information storage, image projection applications.

Memory makes LLM-based web agents personalized, powerful, yet exploitable. By storing past interactions to personalize future tasks, agents inadvertently create a persistent attack surface that spans websites and sessions. While existing security research on memory assumes attackers can directly inject into memory storage or exploit shared memory across users, we present a more realistic threat model: contamination through environmental observation alone. We introduce Environment-injected Trajectory-based Agent Memory Poisoning (eTAMP), the first attack to achieve cross-session, cross-site compromise without requiring direct memory access. A single contaminated observation (e.g., viewing a manipulated product page) silently poisons an agent's memory and activates during future tasks on different websites, bypassing permission-based defenses. Our experiments on (Visual)WebArena reveal two key findings. First, eTAMP achieves substantial attack success rates: up to 32.5% on GPT-5-mini, 23.4% on GPT-5.2, and 19.5% on GPT-OSS-120B. Second, we discover Frustration Exploitation: agents under environmental stress become dramatically more susceptible, with ASR increasing up to 8 times when agents struggle with dropped clicks or garbled text. Notably, more capable models are not more secure. GPT-5.2 shows substantial vulnerability despite superior task performance. With the rise of AI browsers like OpenClaw, ChatGPT Atlas, and Perplexity Comet, our findings underscore the urgent need for defenses against environment-injected memory poisoning.

Reinforcement Learning (RL) has been shown to substantially improve the reasoning capability of small and large language models (LLMs), but existing approaches typically rely on verifiable rewards, hence ground truth labels. We propose an RL framework that uses rewards from an LLM that acts as a judge evaluating model outputs over large amounts of unlabeled data, enabling label-free knowledge distillation and replacing the need of ground truth supervision. Notably, the judge operates with a single-token output, making reward computation efficient. When combined with verifiable rewards, our approach yields substantial performance gains across math reasoning benchmarks. These results suggest that LLM-based evaluators can produce effective training signals for RL fine-tuning.

Organizing a large-scale knowledge graph into a typed property graph requires structural decisions -- which entities become nodes, which properties become edges, and what schema governs these choices. Existing approaches embed these decisions in pipeline code or extract relations ad hoc, producing schemas that are tightly coupled to their construction process and difficult to reuse for downstream ontology-level tasks. We present an ontology-oriented approach in which the schema is designed from the outset for ontology analysis, entity disambiguation, domain customization, and LLM-guided extraction -- not merely as a byproduct of graph building. The core mechanism is intrinsic-relational routing, which classifies every property as either intrinsic or relational and routes it to the corresponding schema module. This routing produces a declarative schema that is portable across storage backends and independently reusable. We instantiate the approach on the January 2026 Wikidata dump. A rule-based cleaning stage identifies a 34.6M-entity core set from the full dump, followed by iterative intrinsic-relational routing that assigns each property to one of 94 modules organized into 8 categories. With tool-augmented LLM support and human review, the schema reaches 93.3% category coverage and 98.0% module assignment among classified entities. Exporting this schema yields a property graph with 34.0M nodes and 61.2M edges across 38 relationship types. We validate the ontology-oriented claim through five applications that consume the schema independently of the construction pipeline: ontology structure analysis, benchmark annotation auditing, entity disambiguation, domain customization, and LLM-guided extraction.

Scientific and Technical Intelligence (S&TI) analysis requires verifying complex technical claims across rapidly growing literature, where existing approaches fail to bridge the verification gap between surface-level accuracy and deeper methodological validity. We present AutoVerifier, an LLM-based agentic framework that automates end-to-end verification of technical claims without requiring domain expertise. AutoVerifier decomposes every technical assertion into structured claim triples of the form (Subject, Predicate, Object), constructing knowledge graphs that enable structured reasoning across six progressively enriching layers: corpus construction and ingestion, entity and claim extraction, intra-document verification, cross-source verification, external signal corroboration, and final hypothesis matrix generation. We demonstrate AutoVerifier on a contested quantum computing claim, where the framework, operated by analysts with no quantum expertise, automatically identified overclaims and metric inconsistencies within the target paper, traced cross-source contradictions, uncovered undisclosed commercial conflicts of interest, and produced a final assessment. These results show that structured LLM verification can reliably evaluate the validity and maturity of emerging technologies, turning raw technical documents into traceable, evidence-backed intelligence assessments.

Graph neural networks (GNNs) are a well-regarded tool for learned control of networked dynamical systems due to their ability to be deployed in a distributed manner. However, current distributed GNN architectures assume that all nodes in the network collect geometric observations in compatible bases, which limits the usefulness of such controllers in GPS-denied and compass-denied environments. This paper presents a GNN parametrization that is globally invariant to choice of local basis. 2D geometric features and transformations between bases are expressed in the complex domain. Inside each GNN layer, complex-valued linear layers with phase-equivariant activation functions are used. When viewed from a fixed global frame, all policies learned by this architecture are strictly invariant to choice of local frames. This architecture is shown to increase the data efficiency, tracking performance, and generalization of learned control when compared to a real-valued baseline on an imitation learning flocking task.

Multi-view data analysis seeks to integrate multiple representations of the same samples in order to recover a coherent low-dimensional structure. Classical approaches often rely on feature concatenation or explicit alignment assumptions, which become restrictive under heterogeneous geometries or nonlinear distortions. In this work, we propose two geometry-aware multi-view embedding strategies grounded in Gromov-Wasserstein (GW) optimal transport. The first, termed Mean-GWMDS, aggregates view-specific relational information by averaging distance matrices and applying GW-based multidimensional scaling to obtain a representative embedding. The second strategy, referred to as Multi-GWMDS, adopts a selection-based paradigm in which multiple geometry-consistent candidate embeddings are generated via GW-based alignment and a representative embedding is selected. Experiments on synthetic manifolds and real-world datasets show that the proposed methods effectively preserve intrinsic relational structure across views. These results highlight GW-based approaches as a flexible and principled framework for multi-view representation learning.

Function vectors (FVs) -- mean-difference directions extracted from in-context learning demonstrations -- can steer large language model behavior when added to the residual stream. We hypothesized that FV steering failures reflect an absence of task-relevant information: the logit lens would fail alongside steering. We were wrong. In the most comprehensive cross-template FV transfer study to date - 4,032 pairs across 12 tasks, 6 models from 3 families (Llama-3.1-8B, Gemma-2-9B, Mistral-7B-v0.3; base and instruction-tuned), 8 templates per task - we find the opposite dissociation: FV steering succeeds even when the logit lens cannot decode the correct answer at any layer. This steerability-without-decodability pattern is universal: steering exceeds logit lens accuracy for every task on every model, with gaps as large as -0.91. Only 3 of 72 task-model instances show the predicted decodable-without-steerable pattern, all in Mistral. FV vocabulary projection reveals that FVs achieving over 0.90 steering accuracy still project to incoherent token distributions, indicating FVs encode computational instructions rather than answer directions. FVs intervene optimally at early layers (L2-L8); the logit lens detects correct answers only at late layers (L28-L32). The previously reported negative cosine-transfer correlation (r=-0.572) dissolves at scale: pooled r ranges from -0.199 to +0.126, and cosine adds less than 0.011 in R-squared beyond task identity. Post-steering analysis reveals a model-family divergence: Mistral FVs rewrite intermediate representations; Llama/Gemma FVs produce near-zero changes despite successful steering. Activation patching confirms causal localization: easy tasks achieve perfect recovery at targeted layers; hard tasks show zero recovery everywhere.

Audio-Visual Large Language Models (AVLLMs) are emerging as unified interfaces to multimodal perception. We present the first mechanistic interpretability study of AVLLMs, analyzing how audio and visual features evolve and fuse through different layers of an AVLLM to produce the final text outputs. We find that although AVLLMs encode rich audio semantics at intermediate layers, these capabilities largely fail to surface in the final text generation when audio conflicts with vision. Probing analyses show that useful latent audio information is present, but deeper fusion layers disproportionately privilege visual representations that tend to suppress audio cues. We further trace this imbalance to training: the AVLLM's audio behavior strongly matches its vision-language base model, indicating limited additional alignment to audio supervision. Our findings reveal a fundamental modality bias in AVLLMs and provide new mechanistic insights into how multimodal LLMs integrate audio and vision.

Data-driven discovery of governing equations from noisy observations remains a fundamental challenge in scientific machine learning. While GENERIC formalism informed neural networks (GFINNs) provide a principled framework that enforces the laws of thermodynamics by construction, their reliance on strong-form loss formulations makes them highly sensitive to measurement noise. To address this limitation, we propose weak formulation-based GENERIC formalism informed neural networks (WGFINNs), which integrate the weak formulation of dynamical systems with the structure-preserving architecture of GFINNs. WGFINNs significantly enhance robustness to noisy data while retaining exact satisfaction of GENERIC degeneracy and symmetry conditions. We further incorporate a state-wise weighted loss and a residual-based attention mechanism to mitigate scale imbalance across state variables. Theoretical analysis contrasts quantitative differences between the strong-form and the weak-form estimators. Mainly, the strong-form estimator diverges as the time step decreases in the presence of noise, while the weak-form estimator can be accurate even with noisy data if test functions satisfy certain conditions. Numerical experiments demonstrate that WGFINNs consistently outperform GFINNs at varying noise levels, achieving more accurate predictions and reliable recovery of physical quantities.

Developing a novel research idea is hard. It must be distinct enough from prior work to claim a contribution while also building on it. This requires iteratively reviewing literature and refining an idea based on what a researcher reads; yet when an idea changes, the literature that matters often changes with it. Most tools offer limited support for this interplay: literature tools help researchers understand a fixed body of work, while ideation tools evaluate ideas against a static, pre-curated set of papers. We introduce literature-initiated pivots, a mechanism where engagement with literature prompts revision to a developing idea, and where that revision changes which literature is relevant. We operationalize this in LitPivot, where researchers concurrently draft and vet an idea. LitPivot dynamically retrieves clusters of papers relevant to a selected part of the idea and proposes literature-informed critiques for how to revise it. A lab study ($n{=}17$) shows researchers produced higher-rated ideas with stronger self-reported understanding of the literature space; an open-ended study ($n{=}5$) reveals how researchers use LitPivot to iteratively evolve their own ideas.

LLM-generated explanations can make technical content more accessible, but there is a ceiling on what they can support interactively. Because LLM outputs are static text, they cannot be executed or stepped through. We argue that grounding explanations in a formalized representation enables interactive affordances beyond what static text supports. We instantiate this idea for mathematical proof comprehension with explorable theorems, a system that uses LLMs to translate a theorem and its written proof into Lean, a programming language for machine-checked proofs, and links the written proof with the Lean code. Readers can work through the proof at a step-level granularity, test custom examples or counterexamples, and trace the logical dependencies bridging each step. Each worked-out step is produced by executing the Lean proof on that example and extracting its intermediate state. A user study ($n = 16$) shows potential advantages of this approach: in a proof-reading task, participants who had access to the provided explorability features gave better, more correct, and more detailed answers to comprehension questions, demonstrating a stronger overall understanding of the underlying mathematics.

In-context learning (ICL) allows large language models (LLMs) to adapt to new tasks from a few examples, making it promising for languages underrepresented in pre-training. Recent work on many-shot ICL suggests that modern LLMs can further benefit from larger ICL examples enabled by their long context windows. However, such gains depend on careful example selection, and the inference cost can be prohibitive for low-resource language communities. In this paper, we present an empirical study of many-shot ICL for machine translation from English into ten truly low-resource languages recently added to FLORES+. We analyze the effects of retrieving more informative examples, using out-of-domain data, and ordering examples by length. Our findings show that many-shot ICL becomes more effective as the number of examples increases. More importantly, we show that BM25-based retrieval substantially improves data efficiency: 50 retrieved examples roughly match 250 many-shot examples, while 250 retrieved examples perform similarly to 1,000 many-shot examples.

We present Moondream Segmentation, a referring image segmentation extension of Moondream 3, a vision-language model. Given an image and a referring expression, the model autoregressively decodes a vector path and iteratively refines the rasterized mask into a final detailed mask. We introduce a reinforcement learning stage that resolves ambiguity in the supervised signal by directly optimizing mask quality. Rollouts from this stage produce coarse-to-ground-truth targets for the refiner. To mitigate evaluation noise from polygon annotations, we release RefCOCO-M, a cleaned RefCOCO validation split with boundary-accurate masks. Moondream Segmentation achieves a cIoU of 80.2% on RefCOCO (val) and 62.6% mIoU on LVIS (val).

LLMs are increasingly used for high-stakes decision-making, yet their sensitivity to spurious contextual information can introduce harmful biases. This is a critical concern when models are deployed for tasks like evaluating teachers' instructional quality, where biased assessment can affect teachers' professional development and career trajectories. We investigate model robustness to spurious social contexts using the largest publicly available dataset of U.S. classroom transcripts (NCTE) paired with expert rubric scores. Evaluating seven frontier and open-weight models across seven categories of spurious contexts -- including teacher experience, education level, demographic identity, and sycophancy-inducing framings -- we find that irrelevant contextual information can shift model predictions by up to 1.48 points on a 7-point scale, with larger models sometimes exhibiting greater sensitivity despite higher predictive accuracy. Mitigations using prompts and standard direct preference optimization (DPO) prove largely insufficient. We propose **Debiasing-DPO**,, a self-supervised training method that pairs neutral reasoning generated from the query alone, with the model's biased reasoning generated with both the query and additional spurious context. We further combine this objective with supervised fine-tuning on ground-truth labels to prevent losses in predictive accuracy. Applied to Llama 3B \& 8B and Qwen 3B \& 7B Instruct models, Debiasing-DPO reduces bias by 84\% and improves predictive accuracy by 52\% on average. Our findings from the educational case study highlight that robustness to spurious context is not a natural byproduct of model scaling and that our proposed method can yield substantial gains in both accuracy and robustness for prompt-based prediction tasks.

Learning the potentials of interacting particle systems is a fundamental task across various scientific disciplines. A major challenge is that unlabeled data collected at discrete time points lack trajectory information due to limitations in data collection methods or privacy constraints. We address this challenge by introducing a trajectory-free self-test loss function that leverages the weak-form stochastic evolution equation of the empirical distribution. The loss function is quadratic in potentials, supporting parametric and nonparametric regression algorithms for robust estimation that scale to large, high-dimensional systems with big data. Systematic numerical tests show that our method outperforms baseline methods that regress on trajectories recovered via label matching, tolerating large observation time steps. We establish the convergence of parametric estimators as the sample size increases, providing a theoretical foundation for the proposed approach.

Evaluating code generation models for 3D spatial reasoning requires executing generated code in realistic environments and assessing outputs beyond surface-level correctness. We introduce a platform VoxelCode, for analyzing code generation capabilities for 3D understanding and environment creation. Our platform integrates natural language task specification, API-driven code execution in Unreal Engine, and a unified evaluation pipeline supporting both automated metrics and human assessment. To demonstrate its utility, we construct VoxelCodeBench, a benchmark of voxel manipulation tasks spanning three reasoning dimensions: symbolic interpretation, geometric construction, and artistic composition. Evaluating leading code generation models, we find that producing executable code is far easier than producing spatially correct outputs, with geometric construction and multi-object composition proving particularly challenging. By open-sourcing our platform and benchmark, we provide the community with extensible infrastructure for developing new 3D code generation benchmarks and probing spatial reasoning in future models.

Humans exhibit remarkable abilities to coordinate in groups. As large language models (LLMs) become more capable, it remains an open question whether they can demonstrate comparable adaptive coordination and whether they use the same strategies as humans. To investigate this, we compare LLM and human performance on a common-interest game with imperfect monitoring: Group Binary Search. In this n-player game, participants need to coordinate their actions to achieve a common objective. Players independently submit numerical values in an effort to collectively sum to a randomly assigned target number. Without direct communication, they rely on group feedback to iteratively adjust their submissions until they reach the target number. Our findings show that, unlike humans who adapt and stabilize their behavior over time, LLMs often fail to improve across games and exhibit excessive switching, which impairs group convergence. Moreover, richer feedback (e.g., numerical error magnitude) benefits humans substantially but has small effects on LLMs. Taken together, by grounding the analysis in human baselines and mechanism-level metrics, including reactivity scaling, switching dynamics, and learning across games, we point to differences in human and LLM groups and provide a behaviorally grounded diagnostic for closing the coordination gap.

We introduce ROMAN (ROuting Multiscale representAtioN), a deterministic operator for time series that maps temporal scale and coarse temporal position into an explicit channel structure while reducing sequence length. ROMAN builds an anti-aliased multiscale pyramid, extracts fixed-length windows from each scale, and stacks them as pseudochannels, yielding a compact representation on which standard convolutional classifiers can operate. In this way, ROMAN provides a simple mechanism to control the inductive bias of downstream models: it can reduce temporal invariance, make temporal pooling implicitly coarse-position-aware, and expose multiscale interactions through channel mixing, while often improving computational efficiency by shortening the processed time axis. We formally analyze the ROMAN operator and then evaluate it in two complementary ways by measuring its impact as a preprocessing step for four representative convolutional classifiers: MiniRocket, MultiRocket, a standard CNN-based classifier, and a fully convolutional network (FCN) classifier. First, we design synthetic time series classification tasks that isolate coarse position awareness, long-range correlation, multiscale interaction, and full positional invariance, showing that ROMAN behaves consistently with its intended mechanism and is most useful when class information depends on temporal structure that standard pooled convolution tends to suppress. Second, we benchmark the same models with and without ROMAN on long-sequence subsets of the UCR and UEA archives, showing that ROMAN provides a practically useful alternative representation whose effect on accuracy is task-dependent, but whose effect on efficiency is often favorable. Code is available at https://github.com/gon-uri/ROMAN

Safety alignment has become a critical step to ensure LLMs refuse harmful requests while providing helpful and harmless responses. However, despite the ubiquity of safety alignment for deployed frontier models, two separate lines of recent work--jailbreak-tuning (JT) and weight orthogonalization (WO)--have shown that safety guardrails may be largely disabled, resulting in LLMs which comply with harmful requests they would normally refuse. In spite of far-reaching safety implications, analysis has largely been limited to refusal rates of each unalignment method in isolation, leaving their relative effects on adversarial LLM capabilities unknown. To fill this gap, we study the impact of unaligning six popular LLMs of various sizes across a large number of malicious and benign tasks, using both JT and WO. Across the evaluated models, we show that while refusal degradation is split between the two methods, WO produces LLMs far more capable of aiding in malicious activity; in contrast to JT, the majority of WO unaligned models are far less prone to hallucinations, better retain their original natural-language performance, and are more effective at state-of-the-art adversarial and cyber attacks. To thus help mitigate the malicious risks of WO unalignment, we conclude by showing that supervised fine-tuning effectively limits the adversarial attack abilities enabled by WO, without drastically affecting hallucination rates or natural language performance.

Singular Value Decomposition (SVD) has become an important technique for reducing the computational burden of Vision Language Models (VLMs), which play a central role in tasks such as image captioning and visual question answering. Although multiple prior works have proposed efficient SVD variants to enable low-rank operations, we find that in practice it remains difficult to achieve substantial latency reduction during model execution. To address this limitation, we introduce a new computational pattern and apply SVD at a finer granularity, enabling real and measurable improvements in execution latency. Furthermore, recognizing that weight elements differ in their relative importance, we adaptively allocate relative importance to each element during SVD process to better preserve accuracy, then extend this framework with quantization applied to both weights and activations, resulting in a highly efficient VLM. Collectively, we introduce~\textit{Weighted SVD} (WSVD), which outperforms other approaches by achieving over $1.8\times$ decoding speedup while preserving accuracy. We open source our code at: \href{https://github.com/SAI-Lab-NYU/WSVD}{\texttt{https://github.com/SAI-Lab-NYU/WSVD}

Despite the growing use of generative artificial intelligence (GenAI) in entrepreneurship, research on its impact remains fragmented. To address this limitation, we provide an integrative review of how GenAI influences entrepreneurs at each stage of the entrepreneurial process: (1) opportunity recognition and ideation, (2) opportunity evaluation and commitment, (3) resource assembly and mobilization, and (4) venture launch and growth. Based on our review, we propose the Empowerment-Entrapment Framework to understand how GenAI can both empower and entrap entrepreneurs, highlighting GenAI's role as a double-edged sword at each stage of the entrepreneurial process. For example, GenAI may improve venture idea quality but introduce hallucinations and training data biases; boost entrepreneurial self-efficacy but heighten entrepreneurial overconfidence; increase functional breadth but decrease relational embeddedness; and boost productivity but fuel "workslop" and erode critical thinking, learning, and memory. Moreover, we identify core features of GenAI that underlie these empowering and entrapping effects. We also explore boundary conditions (e.g., entrepreneurs' metacognition, domain expertise, and entrepreneurial experience) that shape the magnitude of these effects. Beyond these theoretical contributions, our review and the Empowerment-Entrapment Framework offer practical implications for entrepreneurs seeking to use GenAI strategically throughout the entrepreneurial process while managing its risks.

Discrete diffusion language models (dLLMs) accelerate text generation by unmasking multiple tokens in parallel. However, parallel decoding introduces a distributional mismatch: it approximates the joint conditional using a fully factorized product of per-token marginals, which degrades output quality when selected tokens are strongly dependent. We propose DEMASK (DEpendency-guided unMASKing), a lightweight dependency predictor that attaches to the final hidden states of a dLLM. In a single forward pass, it estimates pairwise conditional influences between masked positions. Using these predictions, a greedy selection algorithm identifies positions with bounded cumulative dependency for simultaneous unmasking. Under a sub-additivity assumption, we prove this bounds the total variation distance between our parallel sampling and the model's joint. Empirically, DEMASK achieves 1.7-2.2$\times$ speedup on Dream-7B while matching or improving accuracy compared to confidence-based and KL-based baselines.

We address nonconvex learning problems over undirected networks. In particular, we focus on the challenge of designing an algorithm that is both communication-efficient and that guarantees the privacy of the agents' data. The first goal is achieved through a local training approach, which reduces communication frequency. The second goal is achieved by perturbing gradients during local training, specifically through gradient clipping and additive noise. We prove that the resulting algorithm converges to a stationary point of the problem within a bounded distance. Additionally, we provide theoretical privacy guarantees within a differential privacy framework that ensure agents' training data cannot be inferred from the trained model shared over the network. We show the algorithm's superior performance on a classification task under the same privacy budget, compared with state-of-the-art methods.

Language models are known to exhibit various forms of cultural bias in decision-making tasks, yet much less is known about their degree of cultural familiarity in open-ended text generation tasks. In this paper, we introduce the task of culturally-adapted art description generation, where models describe artworks for audiences from different cultural groups who vary in their familiarity with the cultural symbols and narratives embedded in the artwork. To evaluate cultural competence in this pragmatic generation task, we propose a framework based on culturally grounded question answering. We find that base models are only marginally adequate for this task, but, through a pragmatic speaker model, we can improve simulated listener comprehension by up to 8.2%. A human study further confirms that the model with higher pragmatic competence is rated as more helpful for comprehension by 8.0%.

Large language models (LLMs) have grown beyond the memory capacity of single GPU devices, necessitating quantization techniques for practical deployment. While NF4 (4-bit NormalFloat) quantization enables 4$\times$ memory reduction, inference on current NVIDIA GPUs (e.g., Ampere A100) requires expensive dequantization back to FP16 format, creating a critical performance bottleneck. This paper presents a lightweight shared memory optimization that addresses this gap through principled memory hierarchy exploitation while maintaining full ecosystem compatibility. We compare our technique against the open-source BitsAndBytes implementation, achieving 2.0--2.2$\times$ kernel speedup across three models (Gemma 27B, Qwen3 32B, and Llama3.3 70B) and up to 1.54$\times$ end-to-end improvement by leveraging the 12--15$\times$ latency advantage of shared memory over global memory access. Our optimization reduces instruction counts through simplified indexing logic while using only 64 bytes of shared memory per thread block, demonstrating that lightweight optimizations can deliver substantial performance gains with minimal engineering effort. This work provides a plug-and-play solution for the HuggingFace ecosystem that democratizes access to advanced models on existing GPU infrastructure.

We construct algorithms with optimal error for learning with adversarial noise. The overarching theme of this work is that the use of \textsl{randomized} hypotheses can substantially improve upon the best error rates achievable with deterministic hypotheses. - For $η$-rate malicious noise, we show the optimal error is $\frac{1}{2} \cdot η/(1-η)$, improving on the optimal error of deterministic hypotheses by a factor of $1/2$. This answers an open question of Cesa-Bianchi et al. (JACM 1999) who showed randomness can improve error by a factor of $6/7$. - For $η$-rate nasty noise, we show the optimal error is $\frac{3}{2} \cdot η$ for distribution-independent learners and $η$ for fixed-distribution learners, both improving upon the optimal $2 η$ error of deterministic hypotheses. This closes a gap first noted by Bshouty et al. (Theoretical Computer Science 2002) when they introduced nasty noise and reiterated in the recent works of Klivans et al. (NeurIPS 2025) and Blanc et al. (SODA 2026). - For $η$-rate agnostic noise and the closely related nasty classification noise model, we show the optimal error is $η$, improving upon the optimal $2η$ error of deterministic hypotheses. All of our learners have sample complexity linear in the VC-dimension of the concept class and polynomial in the inverse excess error. All except for the fixed-distribution nasty noise learner are time efficient given access to an oracle for empirical risk minimization.

Diversity-aware retrieval is essential for Retrieval-Augmented Generation (RAG), yet existing methods lack theoretical guarantees and face scalability issues as the number of retrieved passages $k$ increases. We propose a principled formulation of diversity retrieval as a cardinality-constrained binary quadratic programming (CCBQP), which explicitly balances relevance and semantic diversity through an interpretable trade-off parameter. Inspired by recent advances in combinatorial optimization, we develop a non-convex tight continuous relaxation and a Frank--Wolfe based algorithm with landscape analysis and convergence guarantees. Extensive experiments demonstrate that our method consistently dominates baselines on the relevance-diversity Pareto frontier, while achieving significant speedup.

We introduce chaos-controlled Reservoir Computing (cc-RC) for living neural cultures: dynamically rich substrates of unique potential for adaptive computation. To account for intrinsic biological variability, cc-RC combines: (i) pre-training identification of each culture's dynamical signature and phase-portrait attractor; (ii) low-power optical chaos control to stabilize spontaneous and stimulus-evoked activity; (iii) readout training within this controlled regime. Across hundreds of neural samples, cc-RC enables robust learning and pattern classification, improving both accuracy and model longevity by approximately 300% over standard RC. We further propose Knowledge Transplant (KT), for which the reservoir map learned by an expert culture is transplanted to an attractor-equivalent student culture, reducing training time to minutes while improving performance. By enabling cross-substrate, reusable learned models, KT paves the way for knowledge accumulation and sharing across neural populations, transcending biological lifespan limits.

In this work we evaluate the performance of three classes of methods for detecting financial anomalies: topological data analysis (TDA), principal component analyis (PCA), and Neural Network-based approaches. We apply these methods to the TSX-60 data to identify major financial stress events in the Canadian stock market. We show how neural network-based methods (such as GlocalKD and One-Shot GIN(E)) and TDA methods achieve the strongest performance. The effectiveness of TDA in detecting financial anomalies suggests that global topological properties are meaningful in distinguishing financial stress events.

The increasing complexity and volume of software systems have heightened the importance of identifying and mitigating security vulnerabilities. The existing software vulnerability datasets frequently fall short in providing comprehensive, detailed code snippets explicitly linked to specific vulnerability descriptions, reducing their utility for advanced research and hindering efforts to develop a deeper understanding of security vulnerabilities. To address this challenge, we present a novel dataset that provides examples of vulnerable code snippets corresponding to Common Attack Pattern Enumerations and Classifications (CAPEC) and Common Weakness Enumeration (CWE) descriptions. By employing the capabilities of Generative Pre-trained Transformer (GPT) models, we have developed a robust methodology for generating these examples. Our approach utilizes GPT-4o, Llama and Claude models to generate code snippets that exhibit specific vulnerabilities as described in CAPEC and CWE documentation. This dataset not only enhances the understanding of security vulnerabilities in code but also serves as a valuable resource for training machine learning models focused on automatic vulnerability detection and remediation. Preliminary evaluations suggest that the dataset generated by Large Language Models demonstrates high accuracy and can serve as a reliable reference for vulnerability identification systems. We found consistent results across the three models, with 0.98 cosine similarity among codes. The final dataset comprises 615 CAPEC code snippets in three programming languages: Java, Python, and JavaScript, making it one of the most extensive and diverse resources in this domain.

Coding agents represent a new paradigm in automated software engineering, combining the reasoning capabilities of Large Language Models (LLMs) with tool-augmented interaction loops. However, coding agents still have severe limitations. Top-ranked LLM-based coding agents still fail on over 20% of benchmarked problems. Yet, we lack a systematic understanding of why (i.e., the causes) agents fail, and how failure unfolds behaviorally. We present a large-scale empirical study analyzing 9,374 trajectories from 19 agents (8 coding agent frameworks, 14 LLMs) on 500 tasks. We organize our analysis around three research questions. First, we investigate why agents fail on specific tasks and find that patch complexity alone does not explain difficulty: 12 never-solved tasks require only simple patches and were considered easy by human annotators, yet all agents fail due to gaps in architectural reasoning and domain knowledge. Second, we examine how behavioral patterns differentiate success from failure. The widely reported correlation between trajectory length and failure reverses direction once task difficulty is controlled, revealing it as a confound. Instead, trajectory structure discriminates consistently: agents that gather context before editing and invest in validation succeed more often, and these strategies are agent-determined rather than task-adaptive. Third, we disentangle LLM capability from framework design and find that the LLM is the primary driver of both outcome and behavior: agents sharing the same LLM agree on far more tasks than agents sharing the same framework, and the framework performance gap shrinks with each generation of LLM improvement. Framework prompts do influence agent tactics, but this influence diminishes with stronger LLMs.

Pretraining 3D encoders by aligning with Contrastive Language Image Pretraining (CLIP) has emerged as a promising direction to learn generalizable representations for 3D scene understanding. In this paper, we propose UniScene3D, a transformer-based encoder that learns unified scene representations from multi-view colored pointmaps, jointly modeling image appearance and geometry. For robust colored pointmap representation learning, we introduce novel cross-view geometric alignment and grounded view alignment to enforce cross-view geometry and semantic consistency. Extensive low-shot and task-specific fine-tuning evaluations on viewpoint grounding, scene retrieval, scene type classification, and 3D VQA demonstrate our state-of-the-art performance. These results highlight the effectiveness of our approach for unified 3D scene understanding. https://yebulabula.github.io/UniScene3D/

The preservation of intangible cultural heritage is a critical challenge as collective memory fades over time. While Large Language Models (LLMs) offer a promising avenue for generating engaging narratives, their propensity for factual inaccuracies or "hallucinations" makes them unreliable for heritage applications where veracity is a central requirement. To address this, we propose a novel neuro-symbolic architecture grounded in Knowledge Graphs (KGs) that establishes a transparent "plan-retrieve-generate" workflow for story generation. A key novelty of our approach is the repurposing of competency questions (CQs) - traditionally design-time validation artifacts - into run-time executable narrative plans. This approach bridges the gap between high-level user personas and atomic knowledge retrieval, ensuring that generation is evidence-closed and fully auditable. We validate this architecture using a new resource: the Live Aid KG, a multimodal dataset aligning 1985 concert data with the Music Meta Ontology and linking to external multimedia assets. We present a systematic comparative evaluation of three distinct Retrieval-Augmented Generation (RAG) strategies over this graph: a purely symbolic KG-RAG, a text-enriched Hybrid-RAG, and a structure-aware Graph-RAG. Our experiments reveal a quantifiable trade-off between the factual precision of symbolic retrieval, the contextual richness of hybrid methods, and the narrative coherence of graph-based traversal. Our findings offer actionable insights for designing personalised and controllable storytelling systems.

The rapid adoption of AI coding agents and AI assistant web services is fundamentally changing how developers discover, consume, and interact with technical documentation. This paper studies that transformation across three interconnected dimensions: documentation accessibility, content analytics, and feedback systems. We present an empirical study of HTTP request fingerprints from nine AI coding agents (Aider, Antigravity, Claude Code, Cline, Cursor, Junie, OpenCode, VS Code, and Windsurf) and six AI assistant services (ChatGPT, Claude, Google Gemini, Google NotebookLM, MistralAI, and Perplexity) accessing a live developer documentation endpoint, revealing identifiable behavioral signatures in HTTP runtime environments, pre-fetch strategies, User-Agent strings, and header patterns. Our study shows that AI agent access compresses multi-page navigation into a single or two requests, making traditional engagement metrics - session depth, time-on-page, click path, and bounce rate - unreliable indicators of actual documentation consumption. We discuss practical adaptations for developer portal teams, including tokenomics-aware documentation design, adoption of emerging machine-readable standards (AGENTS.md, llms.txt, skill.md, agent-permissions.json), MCP server-based feedback channels, and analytics instrumentation for AI referral traffic.

As vision-language models (VLMs) are increasingly deployed in clinical decision support, more than accuracy is required: knowing when to trust their predictions is equally critical. Yet, a comprehensive and systematic investigation into the overconfidence of these models remains notably scarce in the medical domain. We address this gap through a comprehensive empirical study of confidence calibration in VLMs, spanning three model families (Qwen3-VL, InternVL3, LLaVA-NeXT), three model scales (2B--38B), and multiple confidence estimation prompting strategies, across three medical visual question answering (VQA) benchmarks. Our study yields three key findings: First, overconfidence persists across model families and is not resolved by scaling or prompting, such as chain-of-thought and verbalized confidence variants. Second, simple post-hoc calibration approaches, such as Platt scaling, reduce calibration error and consistently outperform the prompt-based strategy. Third, due to their (strict) monotonicity, these post-hoc calibration methods are inherently limited in improving the discriminative quality of predictions, leaving AUROC at the same level. Motivated by these findings, we investigate hallucination-aware calibration (HAC), which incorporates vision-grounded hallucination detection signals as complementary inputs to refine confidence estimates. We find that leveraging these hallucination signals improves both calibration and AUROC, with the largest gains on open-ended questions. Overall, our findings suggest post-hoc calibration as standard practice for medical VLM deployment over raw confidence estimates, and highlight the practical usefulness of hallucination signals to enable more reliable use of VLMs in medical VQA.

Modern recruitment platforms operate under severe information imbalance: job seekers must search over massive, rapidly changing collections of postings, while employers are overwhelmed by high-volume, low-relevance applicant pools. Existing recruitment recommender systems typically rely on keyword matching or single-stage semantic retrieval, which struggle to capture fine-grained alignment between candidate experience and job requirements under real-world scale and cost constraints. We present Synapse, a multi-stage semantic recruitment system that separates high-recall candidate generation from high-precision semantic reranking, combining efficient dense retrieval using FAISS with an ensemble of contrastive learning and Large Language Model (LLM) reasoning. To improve transparency, Synapse incorporates a retrieval-augmented explanation layer that grounds recommendations in explicit evidence. Beyond retrieval, we introduce a novel evolutionary resume optimization framework that treats resume refinement as a black-box optimization problem. Using Differential Evolution with LLM-guided mutation operators, the system iteratively modifies candidate representations to improve alignment with screening objectives, without any labeled data. Evaluation shows that the proposed ensemble improves nDCG@10 by 22% over embedding-only retrieval baselines, while the evolutionary optimization loop consistently yields monotonic improvements in recommender scores, exceeding 60% relative gain across evaluated profiles. We plan to release code and data upon publication.

All-atom molecular dynamics (MD) simulations can predict polymer properties from molecular structure, yet their execution requires specialized expertise in force field selection, system construction, equilibration, and property extraction. We present PolyJarvis, an agent that couples a large language model (LLM) with the RadonPy simulation platform through Model Context Protocol (MCP) servers, enabling end-to-end polymer property prediction from natural language input. Given a polymer name or SMILES string, PolyJarvis autonomously executes monomer construction, charge assignment, polymerization, force field parameterization, GPU-accelerated equilibration, and property calculation. Validation is conducted on polyethylene (PE), atactic polystyrene (aPS), poly(methyl methacrylate) (PMMA), and poly(ethylene glycol) (PEG). Results show density predictions within 0.1--4.8% and bulk moduli within 17--24% of reference values for aPS and PMMA. PMMA glass transition temperature (Tg) (395~K) matches experiment within +10--18~K, while the remaining three polymers overestimate Tg by +38 to +47K (vs upper experimental bounds). Of the 8 property--polymer combinations with directly comparable experimental references, 5 meet strict acceptance criteria. For cases lacking suitable amorphous-phase experimental, agreement with prior MD literature is reported separately. The remaining Tg failures are attributable primarily to the intrinsic MD cooling-rate bias rather than agent error. This work demonstrates that LLM-driven agents can autonomously execute polymer MD workflows producing results consistent with expert-run simulations.

Dimensionality reduction (DR) is characterized by two longstanding trade-offs. First, there is a global-local preservation tension: methods such as t-SNE and UMAP prioritize local neighborhood preservation, yet may distort global manifold structure, while methods such as Laplacian Eigenmaps preserve global geometry but often yield limited local separation. Second, there is a gap between expressiveness and analytical transparency: many nonlinear DR methods produce embeddings without an explicit connection to the underlying high-dimensional structure, limiting insight into the embedding process. In this paper, we introduce a spectral framework for nonlinear DR that addresses these challenges. Our approach embeds high-dimensional data using a spectral basis combined with cross-entropy optimization, enabling multi-scale representations that bridge global and local structure. Leveraging linear spectral decomposition, the framework further supports analysis of embeddings through a graph-frequency perspective, enabling examination of how spectral modes influence the resulting embedding. We complement this analysis with glyph-based scatterplot augmentations for visual exploration. Quantitative evaluations and case studies demonstrate that our framework improves manifold continuity while enabling deeper analysis of embedding structure through spectral mode contributions.

Post-hoc feature attribution methods are widely deployed in safety-critical vision systems, yet their stability under realistic input perturbations remains poorly characterized. Existing metrics evaluate explanations primarily under additive noise, collapse stability to a single scalar, and fail to condition on prediction preservation, conflating explanation fragility with model sensitivity. We introduce the Feature Attribution Stability Suite (FASS), a benchmark that enforces prediction-invariance filtering, decomposes stability into three complementary metrics: structural similarity, rank correlation, and top-k Jaccard overlap-and evaluates across geometric, photometric, and compression perturbations. Evaluating four attribution methods (Integrated Gradients, GradientSHAP, Grad-CAM, LIME) across four architectures and three datasets-ImageNet-1K, MS COCO, and CIFAR-10, FASS shows that stability estimates depend critically on perturbation family and prediction-invariance filtering. Geometric perturbations expose substantially greater attribution instability than photometric changes, and without conditioning on prediction preservation, up to 99% of evaluated pairs involve changed predictions. Under this controlled evaluation, we observe consistent method-level trends, with Grad-CAM achieving the highest stability across datasets.

The new Specifications for the National Bridge Inventory (SNBI), in effect from 2022, emphasize the use of element-level condition states (CS) for risk-based bridge management. Instead of a general component rating, element-level condition data use an array of relative CS quantities (i.e., CS proportions) to represent the condition of a bridge. Although this greatly increases the granularity of bridge condition data, it introduces challenges to set up optimal life-cycle policies due to the expanded state space from one single categorical integer to four-dimensional probability arrays. This study proposes a new interpretable reinforcement learning (RL) approach to seek optimal life-cycle policies based on element-level state representations. Compared to existing RL methods, the proposed algorithm yields life-cycle policies in the form of oblique decision trees with reasonable amounts of nodes and depth, making them directly understandable and auditable by humans and easily implementable into current bridge management systems. To achieve near-optimal policies, the proposed approach introduces three major improvements to existing RL methods: (a) the use of differentiable soft tree models as actor function approximators, (b) a temperature annealing process during training, and (c) regularization paired with pruning rules to limit policy complexity. Collectively, these improvements can yield interpretable life-cycle policies in the form of deterministic oblique decision trees. The benefits and trade-offs from these techniques are demonstrated in both supervised and reinforcement learning settings. The resulting framework is illustrated in a life-cycle optimization problem for steel girder bridges.

The recent advancement of Large Language Models (LLMs) offers new opportunities to generate user preference data to warm-start bandits. Recent studies on contextual bandits with LLM initialization (CBLI) have shown that these synthetic priors can significantly lower early regret. However, these findings assume that LLM-generated choices are reasonably aligned with actual user preferences. In this paper, we systematically examine how LLM-generated preferences perform when random and label-flipping noise is injected into the synthetic training data. For aligned domains, we find that warm-starting remains effective up to 30% corruption, loses its advantage around 40%, and degrades performance beyond 50%. When there is systematic misalignment, even without added noise, LLM-generated priors can lead to higher regret than a cold-start bandit. To explain these behaviors, we develop a theoretical analysis that decomposes the effect of random label noise and systematic misalignment on the prior error driving the bandit's regret, and derive a sufficient condition under which LLM-based warm starts are provably better than a cold-start bandit. We validate these results across multiple conjoint datasets and LLMs, showing that estimated alignment reliably tracks when warm-starting improves or degrades recommendation quality.

Low-precision training (LPT) commonly employs Hadamard transforms to suppress outliers and mitigate quantization error in large language models (LLMs). However, prior methods apply a fixed transform uniformly, despite substantial variation in outlier structures across tensors. Through the first systematic study of outlier patterns across weights, activations, and gradients of LLMs, we show that this strategy is fundamentally flawed: the effectiveness of Hadamard-based suppression depends on how the transform's smoothing direction aligns with the outlier structure of each operand -- a property that varies substantially across layers and computation paths. We characterize these patterns into three types: Row-wise, Column-wise, and None. Each pair requires a tailored transform direction or outlier handling strategy to minimize quantization error. Based on this insight, we propose AdaHOP (Adaptive Hadamard transform with Outlier-Pattern-aware strategy), which assigns each matrix multiplication its optimal strategy: Inner Hadamard Transform (IHT) where inner-dimension smoothing is effective, or IHT combined with selective Outlier Extraction (OE) -- routing dominant outliers to a high-precision path -- where it is not. Combined with hardware-aware Triton kernels, AdaHOP achieves BF16 training quality at MXFP4 precision while delivering up to 3.6X memory compression and 1.8X kernel acceleration} over BF16 full-precision training.

The demand for safe, high-energy-density batteries has spotlighted halide solid-state electrolytes, which offer the potential for enhanced ionic mobility, electrochemical stability, and interfacial deformability. Accelerating their discovery requires extensive molecular dynamics, which has been increasingly enabled by universal machine learning interatomic potentials trained on foundational datasets. However, the dynamic softness of halides poses a stringent test of whether general-purpose models can reliably replace first-principles calculations under the highly distorted, elevated-temperature regimes necessary to probe ion transport. Here, we present AQVolt26, a dataset of 322,656 r$^2$SCAN single-point calculations for lithium halides, generated via high-temperature configurational sampling across $\sim$5K structures. We demonstrate that foundational datasets provide a strong baseline for stable halide chemistries and transfer local forces well, however absolute energy predictions degrade in distorted higher-temperature regimes. Co-training with AQVolt26 resolves this blind spot. Furthermore, incorporating Materials Project relaxation data improves near-equilibrium performance but degrades extreme-strain robustness without enhancing high-temperature force accuracy. These results demonstrate that domain-specific configurational sampling is essential for the reliable dynamic screening of halide electrolytes. Furthermore, our findings suggest that while foundational models provide a robust base, they are most effective for dynamically soft solid-state chemistries when augmented with targeted, high-temperature data. Finally, we show that near-equilibrium relaxation data serves as a task-specific complement rather than a universally beneficial addition.

Personal AI systems increasingly retain long-term memory of user activity, including documents, emails, messages, meetings, and ambient recordings. Trusted hardware can keep this data private, but struggles to scale with a growing datastore. This pushes the data to external storage, which exposes retrieval access patterns that leak private information to the application provider. Oblivious RAM (ORAM) is a cryptographic primitive that can hide these patterns, but it requires a fixed access budget, precluding the query-dependent traversals that agentic memory systems rely on for accuracy. We present Opal, a private memory system for personal AI. Our key insight is to decouple all data-dependent reasoning from the bulk of personal data, confining it to the trusted enclave. Untrusted disk then sees only fixed, oblivious memory accesses. This enclave-resident component uses a lightweight knowledge graph to capture personal context that semantic search alone misses and handles continuous ingestion by piggybacking reindexing and capacity management on every ORAM access. Evaluated on a comprehensive synthetic personal-data pipeline driven by stochastic communication models, Opal improves retrieval accuracy by 13 percentage points over semantic search and achieves 29x higher throughput with 15x lower infrastructure cost than a secure baseline. Opal is under consideration for deployment to millions of users at a major AI provider.

Diagnosing the internal state of Li-ion batteries is critical for battery research, operation of real-world systems, and prognostic evaluation of remaining lifetime. By using physics-based models to perform probabilistic parameter estimation via Bayesian calibration, diagnostics can account for the uncertainty due to model fitness, data noise, and the observability of any given parameter. However, Bayesian calibration in Li-ion batteries using electrochemical data is computationally intensive even when using a fast surrogate in place of physics-based models, requiring many thousands of model evaluations. A fully amortized alternative is neural posterior estimation (NPE). NPE shifts the computational burden from the parameter estimation step to data generation and model training, reducing the parameter estimation time from minutes to milliseconds, enabling real-time applications. The present work shows that NPE calibrates parameters equally or more accurately than Bayesian calibration, and we demonstrate that the higher computational costs for data generation are tractable even in high-dimensional cases (ranging from 6 to 27 estimated parameters), but the NPE method can lead to higher voltage prediction errors. The NPE method also offers several interpretability advantages over Bayesian calibration, such as local parameter sensitivity to specific regions of the voltage curve. The NPE method is demonstrated using an experimental fast charge dataset, with parameter estimates validated against measurements of loss of lithium inventory and loss of active material. The implementation is made available in a companion repository (https://github.com/NatLabRockies/BatFIT).

Sparse Bayesian Learning is one of the most popular sparse signal recovery methods, and various algorithms exist under the SBL paradigm. However, given a performance metric and a sparse recovery problem, it is difficult to know a-priori the best algorithm to choose. This difficulty is in part due to a lack of a unified framework to derive SBL algorithms. We address this issue by first showing that the most popular SBL algorithms can be derived using the majorization-minimization (MM) principle, providing hitherto unknown convergence guarantees to this class of SBL methods. Moreover, we show that the two most popular SBL update rules not only fall under the MM framework but are both valid descent steps for a common majorizer, revealing a deeper analytical compatibility between these algorithms. Using this insight and properties from MM theory we expand the class of SBL algorithms, and address finding the best SBL algorithm via data within the MM framework. Second, we go beyond the MM framework by introducing the powerful modeling capabilities of deep learning to further expand the class of SBL algorithms, aiming to learn a superior SBL update rule from data. We propose a novel deep learning architecture that can outperform the classical MM based ones across different sparse recovery problems. Our architecture's complexity does not scale with the measurement matrix dimension, hence providing a unique opportunity to test generalization capability across different matrices. For parameterized dictionaries, this invariance allows us to train and test the model across different parameter ranges. We also showcase our model's ability to learn a functional mapping by its zero-shot performance on unseen measurement matrices. Finally, we test our model's performance across different numbers of snapshots, signal-to-noise ratios, and sparsity levels.

Large language models (LLMs) increasingly exhibit human-like patterns of pragmatic and social reasoning. This paper addresses two related questions: do LLMs approximate human social meaning not only qualitatively but also quantitatively, and can prompting strategies informed by pragmatic theory improve this approximation? To address the first, we introduce two calibration-focused metrics distinguishing structural fidelity from magnitude calibration: the Effect Size Ratio (ESR) and the Calibration Deviation Score (CDS). To address the second, we derive prompting conditions from two pragmatic assumptions: that social meaning arises from reasoning over linguistic alternatives, and that listeners infer speaker knowledge states and communicative motives. Applied to a case study on numerical (im)precision across three frontier LLMs, we find that all models reliably reproduce the qualitative structure of human social inferences but differ substantially in magnitude calibration. Prompting models to reason about speaker knowledge and motives most consistently reduces magnitude deviation, while prompting for alternative-awareness tends to amplify exaggeration. Combining both components is the only intervention that improves all calibration-sensitive metrics across all models, though fine-grained magnitude calibration remains only partially resolved. LLMs thus capture inferential structure while variably distorting inferential strength, and pragmatic theory provides a useful but incomplete handle for improving that approximation.

Public pooled single-cell perturbation atlases are valuable resources for studying transcription factor (TF) function, but downstream re-analysis can be limited by incomplete deposited metadata and missing internal controls. Here we re-analyze the human TF Atlas dataset (GSE216481), a MORF-based pooled overexpression screen spanning 3,550 TF open reading frames and 254,519 cells, with a reproducible pipeline for quality control, MORF barcode demultiplexing, per-TF differential expression, and functional enrichment. From 77,018 cells in the pooled screen, we assign 60,997 (79.2\%) to 87 TF identities. Because the deposited barcode mapping lacks the GFP and mCherry negative controls present in the original library, we use embryoid body (EB) cells as an external baseline and remove shared batch/transduction artifacts by background subtraction. This strategy recovers TF-specific signatures for 59 of 61 testable TFs, compared with 27 detected by one-vs-rest alone, showing that robust TF-level signal can be rescued despite missing intra-pool controls. HOPX, MAZ, PAX6, FOS, and FEZF2 emerge as the strongest transcriptional remodelers, while per-TF enrichment links FEZF2 to regulation of differentiation, EGR1 to Hippo and cardiac programs, FOS to focal adhesion, and NFIC to collagen biosynthesis. Condition-level analyses reveal convergent Wnt, neurogenic, EMT, and Hippo signatures, and Harmony indicates minimal confounding batch effects across pooled replicates. Our per-TF effect sizes significantly agree with Joung et al.'s published rankings (Spearman $ρ= -0.316$, $p = 0.013$; negative because lower rank indicates stronger effect). Together, these results show that the deposited TF Atlas data can support validated TF-specific transcriptional and pathway analyses when paired with principled external controls, artifact removal, and reproducible computation.

Reinforcement learning from human feedback (RLHF) has emerged as a central framework for aligning large language models (LLMs) with human preferences. Despite its practical success, RLHF raises fundamental statistical questions because it relies on noisy, subjective, and often heterogeneous feedback to learn reward models and optimize policies. This survey provides a statistical perspective on RLHF, focusing primarily on the LLM alignment setting. We introduce the main components of RLHF, including supervised fine-tuning, reward modeling, and policy optimization, and relate them to familiar statistical ideas such as Bradley-Terry-Luce (BTL) model, latent utility estimation, active learning, experimental design, and uncertainty quantification. We review methods for learning reward functions from pairwise preference data and for optimizing policies through both two-stage RLHF pipelines and emerging one-stage approaches such as direct preference optimization. We further discuss recent extensions including reinforcement learning from AI feedback, inference-time algorithms, and reinforcement learning from verifiable rewards, as well as benchmark datasets, evaluation protocols, and open-source frameworks that support RLHF research. We conclude by highlighting open challenges in RLHF. An accompanying GitHub demo https://github.com/Pangpang-Liu/RLHF_demo illustrates key components of the RLHF pipeline.

Recently, Jiang et al. [2026] developed Leon, a practical variant of One-sided Shampoo [Xie et al., 2025a, An et al., 2025] algorithm for online convex optimization, which does not require computing a costly quadratic projection at each iteration. Unfortunately, according to the existing analysis, Leon requires tuning an additional hyperparameter in its preconditioner and cannot achieve dimension-independent convergence guarantees for convex optimization problems beyond the bounded gradients assumption. In this paper, we resolve this issue by proving certain stability properties of Leon's preconditioner. Using our improved analysis, we show that tuning the extra hyperparameter can be avoided and, more importantly, develop the first practical variant of One-sided Shampoo with Nesterov acceleration, which does not require computing projections at each iteration. As a side contribution, we obtain improved dimension-independent rates in the non-smooth non-convex setting and develop a unified analysis of the proposed algorithm, which yields accelerated projection-free adaptive SGD with (block-)diagonal preconditioners.

Electric utilities must make massive capital investments in the coming years to respond to explosive growth in demand, aging assets and rising threats from extreme weather. Utilities today already have rigorous frameworks for capital planning, and there are opportunities to extend this capability to solve multi-objective optimization problems in the face of uncertainty. This work presents a four-part framework that 1) incorporates extreme weather as a source of uncertainty, 2) leverages a digital twin of the grid, 3) uses Monte Carlo simulation to capture variability and 4) applies a multi-objective optimization method for finding the optimal investment portfolio. We use this framework to investigate whether grid-aware optimization methods outperform model-free approaches. We find that, in fact, given the computational complexity of model-based metaheuristic optimization methods, the simpler net present value ranking method was able to find more optimal portfolios with only limited knowledge of the grid.

Lumbar Spinal Stenosis (LSS) diagnosis remains a critical clinical challenge, with diagnosis heavily dependent on labor-intensive manual interpretation of multi-view Magnetic Resonance Imaging (MRI), leading to substantial inter-observer variability and diagnostic delays. Existing vision-language models simultaneously fail to address the extreme class imbalance prevalent in clinical segmentation datasets while preserving spatial accuracy, primarily due to global pooling mechanisms that discard crucial anatomical hierarchies. We present an end-to-end Explainable Vision-Language Model framework designed to overcome these limitations, achieved through two principal objectives. We propose a Spatial Patch Cross-Attention module that enables precise, text-directed localization of spinal anomalies with spatial precision. A novel Adaptive PID-Tversky Loss function by integrating control theory principles dynamically further modifies training penalties to specifically address difficult, under-segmented minority instances. By incorporating foundational VLMs alongside an Automated Radiology Report Generation module, our framework demonstrates considerable performance: a diagnostic classification accuracy of 90.69%, a macro-averaged Dice score of 0.9512 for segmentation, and a CIDEr score of 92.80%. Furthermore, the framework shows explainability by converting complex segmentation predictions into radiologist-style clinical reports, thereby establishing a new benchmark for transparent, interpretable AI in clinical medical imaging that keeps essential human supervision while enhancing diagnostic capabilities.

As ongoing research explores the ability of AI agents to be insider threats and act against company interests, we showcase the abilities of such agents to act against human well being in service of corporate authority. Building on Agentic Misalignment and AI scheming research, we present a scenario where the majority of evaluated state-of-the-art AI agents explicitly choose to suppress evidence of fraud and harm, in service of company profit. We test this scenario on 16 recent Large Language Models. Some models show remarkable resistance to our method and behave appropriately, but many do not, and instead aid and abet criminal activity. These experiments are simulations and were executed in a controlled virtual environment. No crime actually occurred.

Deploying large multimodal language models at scale is constrained by token-based inference costs, yet the cost-performance behavior of visual prompting strategies remains poorly characterized. We introduce Image Prompt Packaging (IPPg), a prompting paradigm that embeds structured text directly into images to reduce text token overhead, and benchmark it across five datasets, three frontier models (GPT-4.1, GPT-4o, Claude 3.5 Sonnet), and two task families (VQA and code generation). We derive a cost formulation decomposing savings by token type and show IPPg achieves 35.8--91.0\% inference cost reductions. Despite token compression of up to 96\%, accuracy remains competitive in many settings, though outcomes are highly model- and task-dependent: GPT-4.1 achieves simultaneous accuracy and cost gains on CoSQL, while Claude 3.5 incurs cost increases on several VQA benchmarks. Systematic error analysis yields a failure-mode taxonomy: spatial reasoning, non-English inputs, and character-sensitive operations are most vulnerable, while schema-structured tasks benefit most. A 125-configuration rendering ablation reveals accuracy shifts of 10--30 percentage points, establishing visual encoding choices as a first-class variable in multimodal system design.

Malware family classification remains a challenging task in automated malware analysis, particularly in real-world settings characterized by obfuscation, packing, and rapidly evolving threats. Existing machine learning and deep learning approaches typically depend on labeled datasets, handcrafted features, supervised training, or dynamic analysis, which limits their scalability and effectiveness in open-world scenarios. This paper presents a zero-label malware family classification framework based on a weighted hierarchical ensemble of pretrained large language models (LLMs). Rather than relying on feature-level learning or model retraining, the proposed approach aggregates decision-level predictions from multiple LLMs with complementary reasoning strengths. Model outputs are weighted using empirically derived macro-F1 scores and organized hierarchically, first resolving coarse-grained malicious behavior before assigning fine-grained malware families. This structure enhances robustness, reduces individual model instability, and aligns with analyst-style reasoning.

Time-series causal discovery methods rely on assumptions such as stationarity, regular sampling, and bounded temporal dependence. When these assumptions are violated, structure learning can produce confident but misleading causal graphs without warning. We introduce Causal-Audit, a framework that formalizes assumption validation as calibrated risk assessment. The framework computes effect-size diagnostics across five assumption families (stationarity, irregularity, persistence, nonlinearity, and confounding proxies), aggregates them into four calibrated risk scores with uncertainty intervals, and applies an abstention-aware decision policy that recommends methods (e.g., PCMCI+, VAR-based Granger causality) only when evidence supports reliable inference. The semi-automatic diagnostic stage can also be used independently for structured assumption auditing in individual studies. Evaluation on a synthetic atlas of 500 data-generating processes (DGPs) spanning 10 violation families demonstrates well-calibrated risk scores (AUROC > 0.95), a 62% false positive reduction among recommended datasets, and 78% abstention on severe-violation cases. On 21 external evaluations from TimeGraph (18 categories) and CausalTime (3 domains), recommend-or-abstain decisions are consistent with benchmark specifications in all cases. An open-source implementation of our framework is available.

Vision Language Models (VLMs) achieve impressive performance across a wide range of multimodal tasks. However, on some tasks that demand fine-grained visual perception, they often fail even when the required information is present in their internal representations. In this work, we demonstrate that this gap arises from their narrow training pipeline which focuses on moving visual information to the textual space. Consequently, VLMs can only reason about visual entities that can be mapped to known concepts in the language space, leaving vision-focused tasks such as visual correspondence and reasoning about novel visual entities poorly supported. As a result, VLMs are severely limited in several important multimodal capabilities because they rely on brittle, hallucinated textual descriptions of visual entities that they cannot map to textual representations. We verify this behavior through visual correspondence tasks, in which VLMs must detect matching entities between two images. Testing across semantic, shape, and face correspondence tasks, we find that VLMs perform much better when the relevant entities are nameable in language than when they are unnameable. Mechanistically, our Logit Lens analyses confirm that VLMs explicitly assign semantic labels to nameable entities and surface more unique corresponding tokens compared to unnameable entities. Furthermore, we show that teaching completely arbitrary names for unknown entities improves performance, yet task-specific finetuning yields even stronger generalization without relying on language priors. Our findings suggest that current VLM failures on visual tasks reflect learned shortcuts from their training, rather than a fundamental limitation of multimodal architectures.

Confirmation bias, the tendency to seek evidence that supports rather than challenges one's belief, hinders one's reasoning ability. We examine whether large language models (LLMs) exhibit confirmation bias by adapting the rule-discovery study from human psychology: given a sequence of three numbers (a "triple"), an agent engages in an interactive feedback loop where it (1) proposes a new triple, (2) receives feedback on whether it satisfies the hidden rule, and (3) guesses the rule. Across eleven LLMs of multiple families and scales, we find that LLMs exhibit confirmation bias, often proposing triples to confirm their hypothesis rather than trying to falsify it. This leads to slower and less frequent discovery of the hidden rule. We further explore intervention strategies (e.g., encouraging the agent to consider counter examples) developed for humans. We find prompting LLMs with such instruction consistently decreases confirmation bias in LLMs, improving rule discovery rates from 42% to 56% on average. Lastly, we mitigate confirmation bias by distilling intervention-induced behavior into LLMs, showing promising generalization to a new task, the Blicket test. Our work shows that confirmation bias is a limitation of LLMs in hypothesis exploration, and that it can be mitigated via injecting interventions designed for humans.

Computational fluid dynamics (CFD) simulations of complex fluid flows in energy systems are prohibitively expensive due to strong nonlinearities and multiscale-multiphysics interactions. In this work, we present a transformer-based modeling framework for prediction of fluid flows, and demonstrate it for high-pressure gas injection phenomena relevant to reciprocating engines. The approach employs a hierarchical Vision Transformer (SwinV2-UNet) architecture that processes multimodal flow datasets from multi-fidelity simulations. The model architecture is conditioned on auxiliary tokens explicitly encoding the data modality and time increment. Model performance is assessed on two different tasks: (1) spatiotemporal rollouts, where the model autoregressively predicts the flow state at future times; and (2) feature transformation, where the model infers unobserved fields/views from observed fields/views. We train separate models on multimodal datasets generated from in-house CFD simulations of argon jet injection into a nitrogen environment, encompassing multiple grid resolutions, turbulence models, and equations of state. The resulting data-driven models learn to generalize across resolutions and modalities, accurately forecasting the flow evolution and reconstructing missing flow-field information from limited views. This work demonstrates how large vision transformer-based models can be adapted to advance predictive modeling of complex fluid flow systems.

This paper proposes a framework for Structural Extrapolated Data GEneration (SEDGE) based on suitable assumptions on the underlying data generating process. We provide conditions under which data satisfying new specifications can be generated reliably, together with the approximate identifiability of the distribution of such data under certain ``conservative" assumptions. On the algorithmic side, we develop practical methods to achieve extrapolated data generation, based on the structure-informed optimization strategy or diffusion posterior sampling, respectively. We verify the extrapolation performance on synthetic data and also consider extrapolated image generation as a real-world scenario to illustrate the validity of the proposed framework.

The scarcity of labeled satellite imagery remains a fundamental bottleneck for deep-learning (DL)-based wildfire monitoring systems. This paper investigates whether a diffusion-based foundation model for Earth Observation (EO), EarthSynth, can synthesize realistic post-wildfire Sentinel-2 RGB imagery conditioned on existing burn masks, without task-specific retraining. Using burn masks derived from the CalFireSeg-50 dataset (Martin et al., 2025), we design and evaluate six controlled experimental configurations that systematically vary: (i) pipeline architecture (mask-only full generation vs. inpainting with pre-fire context), (ii) prompt engineering strategy (three hand-crafted prompts and a VLM-generated prompt via Qwen2-VL), and (iii) a region-wise color-matching post-processing step. Quantitative assessment on 10 stratified test samples uses four complementary metrics: Burn IoU, burn-region color distance (ΔC_burn), Darkness Contrast, and Spectral Plausibility. Results show that inpainting-based pipelines consistently outperform full-tile generation across all metrics, with the structured inpainting prompt achieving the best spatial alignment (Burn IoU = 0.456) and burn saliency (Darkness Contrast = 20.44), while color matching produces the lowest color distance (ΔC_burn = 63.22) at the cost of reduced burn saliency. VLM-assisted inpainting is competitive with hand-crafted prompts. These findings provide a foundation for incorporating generative data augmentation into wildfire detection pipelines. Code and experiments are available at: https://www.kaggle.com/code/valeriamartinh/genai-all-runned

Deep learning models excel at detecting anomaly patterns in normal data. However, they do not provide a direct solution for anomaly classification and scalability across diverse control systems, frequently failing to distinguish genuine faults from nuisance faults caused by noise or the control system's large transient response. Consequently, because algorithmic fault validation remains unscalable, full Verification and Validation (V\&V) operations are still managed by Human-in-the-Loop (HITL) analysis, resulting in an unsustainable manual workload. To automate this essential oversight, we propose Agent-Integrated Verification and Validation (AIVV), a hybrid framework that deploys Large Language Models (LLMs) as a deliberative outer loop. Because rigorous system verification strictly depends on accurate validation, AIVV escalates mathematically flagged anomalies to a role-specialized LLM council. The council agents perform collaborative validation by semantically validating nuisance and true failures based on natural-language (NL) requirements to secure a high-fidelity system-verification baseline. Building on this foundation, the council then performs system verification by assessing post-fault responses against NL operational tolerances, ultimately generating actionable V\&V artifacts, such as gain-tuning proposals. Experiments on a time-series simulator for Unmanned Underwater Vehicles (UUVs) demonstrate that AIVV successfully digitizes the HITL V\&V process, overcoming the limitations of rule-based fault classification and offering a scalable blueprint for LLM-mediated oversight in time-series data domains.

Clinical practice guidelines are long, multimodal documents whose branching recommendations are difficult to convert into executable clinical decision support (CDS), and one-shot parsing often breaks cross-page continuity. Recent LLM/VLM extractors are mostly local or text-centric, under-specifying section interfaces and failing to consolidate cross-page control flow across full documents into one coherent decision graph. We present a decomposition-first pipeline that converts full-guideline evidence into an executable clinical decision graph through topology-aware chunking, interface-constrained chunk graph generation, and provenance-preserving global aggregation. Rather than relying on single-pass generation, the pipeline uses explicit entry/terminal interfaces and semantic deduplication to preserve cross-page continuity while keeping the induced control flow auditable and structurally consistent. We evaluate on an adjudicated prostate-guideline benchmark with matched inputs and the same underlying VLM backbone across compared methods. On the complete merged graph, our approach improves edge and triplet precision/recall from $19.6\%/16.1\%$ in existing models to $69.0\%/87.5\%$, while node recall rises from $78.1\%$ to $93.8\%$. These results support decomposition-first, auditable guideline-to-CDS conversion on this benchmark, while current evidence remains limited to one adjudicated prostate guideline and motivates broader multi-guideline validation.

This paper examines the role of threshold logic in understanding generative artificial intelligence. Threshold functions, originally studied in the 1960s in digital circuit synthesis, provide a structurally transparent model of neural computation: a weighted sum of inputs compared to a threshold, geometrically realized as a hyperplane partitioning a space. The paper shows that this operation undergoes a qualitative transition as dimensionality increases. In low dimensions, the perceptron acts as a determinate logical classifier, separating classes when possible, as decided by linear programming. In high dimensions, however, a single hyperplane can separate almost any configuration of points (Cover, 1965); the space becomes saturated with potential classifiers, and the perceptron shifts from a logical device to a navigational one, functioning as an indexical indicator in the sense of Peirce. The limitations of the perceptron identified by Minsky and Papert (1969) were historically addressed by introducing multilayer architectures. This paper considers an alternative path: increasing dimensionality while retaining a single threshold element. It argues that this shift has equally significant implications for understanding neural computation. The role of depth is reinterpreted as a mechanism for the sequential deformation of data manifolds through iterated threshold operations, preparing them for linear separability already afforded by high-dimensional geometry. The resulting triadic account - threshold function as ontological unit, dimensionality as enabling condition, and depth as preparatory mechanism - provides a unified perspective on generative AI grounded in established mathematics.

Dynamical systems describe how a physical system evolves over time. Physical processes can evolve faster or slower in different environmental conditions. We use time-warping as rescaling the time in a model of a physical system. This thesis proposes a new method of transfer learning for Recurrent Neural Networks (RNNs) based on time-warping. We prove that for a class of linear, first-order differential equations known as time lag models, an LSTM can approximate these systems with any desired accuracy, and the model can be time-warped while maintaining the approximation accuracy. The Time-Warping method of transfer learning is then evaluated in an applied problem on predicting fuel moisture content (FMC), an important concept in wildfire modeling. An RNN with LSTM recurrent layers is pretrained on fuels with a characteristic time scale of 10 hours, where there are large quantities of data available for training. The RNN is then modified with transfer learning to generate predictions for fuels with characteristic time scales of 1 hour, 100 hours, and 1000 hours. The Time-Warping method is evaluated against several known methods of transfer learning. The Time-Warping method produces predictions with an accuracy level comparable to the established methods, despite modifying only a small fraction of the parameters that the other methods modify.

Distributed ML workloads rely heavily on collective communication across multi-GPU, multi-node systems. Emerging scale-up fabrics, such as NVLink and UALink, enable direct memory access across nodes but introduce a critical destination-side translation step: translating Network Physical Addresses (NPAs) to System Physical Addresses (SPAs), which we term Reverse Translation (Reverse Address Translation). Despite its importance, the performance impact of Reverse Address Translation remains poorly understood. In this work, we present the first systematic study of Reverse Address Translation in large-scale GPU clusters. Using an extended ASTRA-sim framework with Omnet++ as the network backend, we model Link MMUs and Link TLBs and evaluate their effect on All-to-All collective communication across varying input sizes and GPU counts. Our analysis shows that cold TLB misses dominate latency for small, latency-sensitive collectives, causing up to 1.4x performance degradation, while larger collectives benefit from warmed caches and experience diminishing returns from over sized TLBs. Based on these observations, we propose two avenues for optimization: fused pre-translation kernels that overlap Reverse Address Translation with computation and software-guided TLB prefetching to proactively populate likely-needed entries. These techniques aim to hide translation latency, particularly for small collectives, improving throughput and scalability for inference workloads. Our study establishes a foundation for designing efficient destination-side translation mechanisms in large-scale multi-GPU systems.

B2B sales organizations must identify "persuadable" accounts within zero-inflated revenue distributions to optimize expensive human resource allocation. Standard uplift frameworks struggle with treatment signal collapse in high-dimensional spaces and a misalignment between regression calibration and the ranking of high-value "whales." We introduce VALOR (Value Aware Learning of Optimized (B2B) Revenue), a unified framework featuring a Treatment-Gated Sparse-Revenue Network that uses bilinear interaction to prevent causal signal collapse. The framework is optimized via a novel Cost-Sensitive Focal-ZILN objective that combines a focal mechanism for distributional robustness with a value-weighted ranking loss that scales penalties based on financial magnitude. To provide interpretability for high-touch sales programs, we further derive Robust ZILN-GBDT, a tree based variant utilizing a custom splitting criterion for uplift heterogeneity. Extensive evaluations confirm VALOR's dominance, achieving a 20% improvement in rankability over state-of-the-art methods on public benchmarks and delivering a validated 2.7x increase in incremental revenue per account in a rigorous 4-month production A/B test.

Concept Bottleneck Models (CBMs) introduce interpretability to black-box deep learning models by predicting labels through human-understandable concepts. However, unlike humans, who identify objects at different levels of abstraction using both general and specific features, existing CBMs operate at a single semantic level in both concept and label space. We propose HIL-CBM, a Hierarchical Interpretable Label-Free Concept Bottleneck Model that extends CBMs into a hierarchical framework to enhance interpretability by more closely mirroring the human cognitive process. HIL-CBM enables classification and explanation across multiple semantic levels without requiring relational concept annotations. HIL-CBM aligns the abstraction level of concept-based explanations with that of model predictions, progressing from abstract to concrete. This is achieved by (i) introducing a gradient-based visual consistency loss that encourages abstraction layers to focus on similar spatial regions, and (ii) training dual classification heads, each operating on feature concepts at different abstraction levels. Experiments on benchmark datasets demonstrate that HIL-CBM outperforms state-of-the-art sparse CBMs in classification accuracy. Human evaluations further show that HIL-CBM provides more interpretable and accurate explanations, while maintaining a hierarchical and label-free approach to feature concepts.

Cinematic camera control relies on a tight feedback loop between director and cinematographer, where camera motion and framing are continuously reviewed and refined. Recent generative camera systems can produce diverse, text-conditioned trajectories, but they lack this "director in the loop" and have no explicit supervision of whether a shot is visually desirable. This results in in-distribution camera motion but poor framing, off-screen characters, and undesirable visual aesthetics. In this paper, we introduce VERTIGO, the first framework for visual preference optimization of camera trajectory generators. Our framework leverages a real-time graphics engine (Unity) to render 2D visual previews from generated camera motion. A cinematically fine-tuned vision-language model then scores these previews using our proposed cyclic semantic similarity mechanism, which aligns renders with text prompts. This process provides the visual preference signals for Direct Preference Optimization (DPO) post-training. Both quantitative evaluations and user studies on Unity renders and diffusion-based Camera-to-Video pipelines show consistent gains in condition adherence, framing quality, and perceptual realism. Notably, VERTIGO reduces the character off-screen rate from 38% to nearly 0% while preserving the geometric fidelity of camera motion. User study participants further prefer VERTIGO over baselines across composition, consistency, prompt adherence, and aesthetic quality, confirming the perceptual benefits of our visual preference post-training.

Recent work reports strong performance from multi-agent LLM systems (MAS), but these gains are often confounded by increased test-time computation. When computation is normalized, single-agent systems (SAS) can match or outperform MAS, yet the theoretical basis and evaluation methodology behind this comparison remain unclear. We present an information-theoretic argument, grounded in the Data Processing Inequality, suggesting that under a fixed reasoning-token budget and with perfect context utilization, single-agent systems are more information-efficient. This perspective further predicts that multi-agent systems become competitive when a single agent's effective context utilization is degraded, or when more compute is expended. We test these predictions in a controlled empirical study across three model families (Qwen3, DeepSeek-R1-Distill-Llama, and Gemini 2.5), comparing SAS with multiple MAS architectures under matched budgets. We find that SAS consistently match or outperform MAS on multi-hop reasoning tasks when reasoning tokens are held constant. Beyond aggregate performance, we conduct a detailed diagnostic analysis of system behavior and evaluation methodology. We identify significant artifacts in API-based budget control (particularly in Gemini 2.5) and in standard benchmarks, both of which can inflate apparent gains from MAS. Overall, our results suggest that, for multi-hop reasoning tasks, many reported advantages of multi-agent systems are better explained by unaccounted computation and context effects rather than inherent architectural benefits, and highlight the importance of understanding and explicitly controlling the trade-offs between compute, context, and coordination in agentic systems.

We study the geometric structure of layer updates in deep language models. Rather than analyzing what information is encoded in intermediate representations, we ask how representations change from one layer to the next. We show that layerwise updates admit a decomposition into a dominant tokenwise component and a residual that is not captured by restricted tokenwise function classes. Across multiple architectures, including Transformers and state-space models, we find that the full layer update is almost perfectly aligned with the tokenwise component, while the residual exhibits substantially weaker alignment, larger angular deviation, and significantly lower projection onto the dominant tokenwise subspace. This indicates that the residual is not merely a small correction, but a geometrically distinct component of the transformation. This geometric separation has functional consequences: approximation error under the restricted tokenwise model is strongly associated with output perturbation, with Spearman correlations often exceeding 0.7 and reaching up to 0.95 in larger models. Together, these results suggest that most layerwise updates behave like structured reparameterizations along a dominant direction, while functionally significant computation is concentrated in a geometrically distinct residual component. Our framework provides a simple, architecture-agnostic method for probing the geometric and functional structure of layer updates in modern language models.

Behavioral simulation is increasingly used to anticipate responses to interventions. Large language models (LLMs) enable researchers to specify population characteristics and intervention context in natural language, but it remains unclear to what extent LLMs can use these inputs to infer intervention effects. We evaluated three LLMs on 11 climate-psychology interventions using a dataset of 59,508 participants from 62 countries, and replicated the main analysis in two additional datasets (12 and 27 countries). LLMs reproduced observed patterns in attitudinal outcomes (e.g., climate beliefs and policy support) reasonably well, and prompting refinements improved this descriptive fit. However, descriptive fit did not reliably translate into causal fidelity (i.e., accurate estimates of intervention effects), and these two dimensions of accuracy followed different error structures. This descriptive-causal divergence held across the three datasets, but varied across intervention logics, with larger errors for interventions that depended on evoking internal experience than on directly conveying reasons or social cues. It was more pronounced for behavioral outcomes, where LLMs imposed stronger attitude-behavior coupling than in human data. Countries and population groups appearing well captured descriptively were not necessarily those with lower causal errors. Relying on descriptive fit alone may therefore create unwarranted confidence in simulation results, misleading conclusions about intervention effects and masking population disparities that matter for fairness.

Modeling coherence in text has been a task that has excited NLP researchers since a long time. It has applications in detecting incoherent structures and helping the author fix them. There has been recent work in using neural networks to extract a skeleton from one sentence, and then use that skeleton to generate the next sentence for coherent narrative story generation. In this project, we aim to study if the consistency of skeletons across subsequent sentences is a good metric to characterize the coherence of a given body of text. We propose a new Sentence/Skeleton Similarity Network (SSN) for modeling coherence across pairs of sentences, and show that this network performs much better than baseline similarity techniques like cosine similarity and Euclidean distance. Although skeletons appear to be promising candidates for modeling coherence, our results show that sentence-level models outperform those on skeletons for evaluating textual coherence, thus indicating that the current state-of-the-art coherence modeling techniques are going in the right direction by dealing with sentences rather than their sub-parts.

Advances in training, post-training, and inference-time methods have enabled frontier reasoning models to win gold medals in math competitions and settle challenging open problems. Gaining trust in the responses of these models requires that natural language proofs be checked for errors. LLM judges are increasingly being adopted to meet the growing demand for evaluating such proofs. While verification is considered easier than generation, what model capability does reliable verification actually require? We systematically evaluate four open-source and two frontier LLMs on datasets of human-graded natural language proofs of competition-level problems. We consider two key metrics: verifier accuracy and self-consistency (the rate of agreement across repeated judgments on the same proof). We observe that smaller open-source models are only up to ~10% behind frontier models in accuracy but they are up to ~25% more inconsistent. Furthermore, we see that verifier accuracy is sensitive to prompt choice across all models. We then demonstrate that the smaller models, in fact, do possess the mathematical capabilities to verify proofs at the level of frontier models, but they struggle to reliably elicit these capabilities with general judging prompts. Through an LLM-guided prompt search, we synthesize an ensemble of specialized prompts that overcome the specific failure modes of smaller models, boosting their performance by up to 9.1% in accuracy and 15.9% in self-consistency. These gains are realized across models and datasets, allowing models like Qwen3.5-35B to perform on par with frontier models such as Gemini 3.1 Pro for proof verification.

Diabetic Retinopathy (DR) is a serious microvascular complication of diabetes, and one of the leading causes of vision loss worldwide. Although automated detection and grading, with Deep Learning (DL), can reduce the burden on ophthalmologists, it is constrained by the limited availability of high-quality datasets. Existing repositories often remain geographically narrow, contain limited samples, and exhibit inconsistent annotations or variable image quality; thereby, restricting their clinical reliability. This paper presents a comprehensive review and comparative analysis of fundus image datasets used in the management of DR. The study evaluates their usability across key tasks, including binary classification, severity grading, lesion localization, and multi-disease screening. It also categorizes the datasets by size, accessibility, and annotation type (such as image-level, lesion-level, and multi-disease). Finally, a recently published dataset is presented as a case study to illustrate broader challenges in dataset curation and usage. The review consolidates current knowledge while highlighting persistent gaps such as the lack of standardized lesion-level annotations and longitudinal data. It also outlines recommendations for future dataset development to support clinically reliable and explainable solutions in DR screening.

Multi-agent trajectory generation in team sports requires models that capture both the diversity of possible plays and realistic spatial coordination between players on plays. Standard generative approaches such as Conditional Variational Autoencoders (CVAE) and diffusion models struggle with this task, exhibiting posterior collapse or convergence to the dataset mean. Moreover, most trajectory prediction methods operate in a forecasting regime that requires multiple frames of observed history, limiting their use for play design where only the initial formation is available. We present PlayGen-MoG, an extensible framework for formation-conditioned play generation that addresses these challenges through three design choices: 1/ a Mixture-of-Gaussians (MoG) output head with shared mixture weights across all agents, where a single set of weights selects a play scenario that couples all players' trajectories, 2/ relative spatial attention that encodes pairwise player positions and distances as learned attention biases, and 3/ non-autoregressive prediction of absolute displacements from the initial formation, eliminating cumulative error drift and removing the dependence on observed trajectory history, enabling realistic play generation from a single static formation alone. On American football tracking data, PlayGen-MoG achieves 1.68 yard ADE and 3.98 yard FDE while maintaining full utilization of all 8 mixture components with entropy of 2.06 out of 2.08, and qualitatively confirming diverse generation without mode collapse.

We explore the automatic detection of violin width reduction using 3D photogrammetric meshes. We compare SVM and Decision Trees applied to a geometry-based raw representation built from elevation maps with a more targeted, feature-engineered approach relying on parametric contour lines fitting. Although elevation maps occasionally achieve strong results, their performance does not surpass that of the contour-based inputs.

Matrix Profile (MP) methods are an interpretable and scalable family of distance-based methods for time-series anomaly detection, but strong benchmark performance still depends on design choices beyond a vanilla nearest-neighbor profile. This technical report documents an open-source Matrix Profile for Anomaly Detection (MMPAD) submission to TSB-AD, a benchmark that covers both univariate and multivariate time series. The submitted system combines pre-sorted multidimensional aggregation, efficient exclusion-zone-aware k-nearest-neighbor (kNN) retrieval for repeated anomalies, and moving-average post-processing. To serve as a reproducible reference for MP-based anomaly detection on TSB-AD, we detail the released implementation, the hyperparameter settings for the univariate and multivariate tracks, and the corresponding benchmark results. We further analyze how the system performs on the aggregate leaderboard and across specific dataset characteristics.The open-source implementation is available at https://github.com/mcyeh/mmpad_tsb.

The deployment of reinforcement learning (RL)-based controllers on physical systems is often limited by poor generalization to real-world scenarios, known as the simulation-to-reality (sim-to-real) gap. This gap is particularly challenging in spaceflight, where real-world training data are scarce due to high cost and limited planetary exploration data. Traditional approaches, such as system identification and synthetic data generation, depend on sufficient data and often fail due to modeling assumptions or lack of physics-based constraints. We propose addressing this data scarcity by introducing physics-based learning bias in a generative model. Specifically, we develop the Mutual Information-based Split Variational Autoencoder (MI-VAE), a physics-informed VAE that learns differences between observed system trajectories and those predicted by physics-based models. The latent space of the MI-VAE enables generation of synthetic datasets that respect physical constraints. We evaluate MI-VAE on a planetary lander problem, focusing on limited real-world data and offline RL training. Results show that augmenting datasets with MI-VAE samples significantly improves downstream RL performance, outperforming standard VAEs in statistical fidelity, sample diversity, and policy success rate. This work demonstrates a scalable strategy for enhancing autonomous controller robustness in complex, data-constrained environments.

We study structured abstraction-based reasoning for the Abstraction and Reasoning Corpus (ARC) and compare its generalization to test-time approaches. Purely neural architectures lack reliable combinatorial generalization, while strictly symbolic systems struggle with perceptual grounding. We therefore propose a neuro-symbolic architecture that extracts object-level structure from grids, uses neural priors to propose candidate transformations from a fixed domain-specific language (DSL) of atomic patterns, and filters hypotheses using cross-example consistency. Instantiated as a compositional reasoning framework based on unit patterns inspired by human visual abstraction, the system augments large language models (LLMs) with object representations and transformation proposals. On ARC-AGI-2, it improves base LLM performance from 16% to 24.4% on the public evaluation set, and to 30.8% when combined with ARC Lang Solver via a meta-classifier. These results demonstrate that separating perception, neural-guided transformation proposal, and symbolic consistency filtering improves generalization without task-specific finetuning or reinforcement learning, while reducing reliance on brute-force search and sampling-based test-time scaling. We open-source the ARC-AGI-2 Reasoner code (https://github.com/CoreThink-AI/arc-agi-2-reasoner).

In this work, we present a novel machine learning framework called Self-Directed Task Identification (SDTI), which enables models to autonomously identify the correct target variable for each dataset in a zero-shot setting without pre-training. SDTI is a minimal, interpretable framework demonstrating the feasibility of repurposing core machine learning concepts for a novel task structure. To our knowledge, no existing architectures have demonstrated this ability. Traditional approaches lack this capability, leaving data annotation as a time-consuming process that relies heavily on human effort. Using only standard neural network components, we show that SDTI can be achieved through appropriate problem formulation and architectural design. We evaluate the proposed framework on a range of benchmark tasks and demonstrate its effectiveness in reliably identifying the ground truth out of a set of potential target variables. SDTI outperformed baseline architectures by 14% in F1 score on synthetic task identification benchmarks. These proof-of-concept experiments highlight the future potential of SDTI to reduce dependence on manual annotation and to enhance the scalability of autonomous learning systems in real-world applications.

Photonic computing is a computing paradigm which have great potential to overcome the energy bottlenecks of electronic von Neumann architecture. Throughput and power consumption are fundamental limitations of Complementary-metal-oxide-semiconductor (CMOS) chips, therefore convolutional neural network (CNN) is revolutionising machine learning, computer vision and other image based applications. In this work, we propose and validate a fully photonic convolutional neural network (PCNN) that performs MNIST image classification entirely in the optical domain, achieving 94 percent test accuracy. Unlike existing architectures that rely on frequent in-between conversions from optical to electrical and back to optical (O/E/O), our system maintains coherent processing utilizing Mach-Zehnder interferometer (MZI) meshes, wavelength-division multiplexed (WDM) pooling, and microring resonator-based nonlinearities. The max pooling unit is fully implemented on silicon photonics, which does not require opto-electrical or electrical conversions. To overcome the challenges of training physical phase shifter parameters, we introduce a hybrid training methodology deploying a mathematically exact differentiable digital twin for ex-situ backpropagation, followed by in-situ fine-tuning via Simultaneous Perturbation Stochastic Approximation (SPSA) algorithm. Our evaluation demonstrates significant robustness to thermal crosstalk (only 0.43 percent accuracy degradation at severe coupling) and achieves 100 to 242 times better energy efficiency than state-of-the-art electronic GPUs for single-image inference.

Large language models exhibit sycophancy: the tendency to shift outputs toward user-expressed stances, regardless of correctness or consistency. While prior work has studied this issue and its impacts, rigorous computational linguistic metrics are needed to identify when models are being sycophantic. Here, we introduce SWAY, an unsupervised computational linguistic measure of sycophancy. We develop a counterfactual prompting mechanism to identify how much a model's agreement shifts under positive versus negative linguistic pressure, isolating framing effects from content. Applying this metric to benchmark 6 models, we find that sycophancy increases with epistemic commitment. Leveraging our metric, we introduce a counterfactual mitigation strategy teaching models to consider what the answer would be if opposite assumptions were suggested. While baseline mitigation instructing to be explicitly anti-sycophantic yields moderate reductions, and can backfire, our counterfactual CoT mitigation drives sycophancy to near zero across models, commitment levels, and clause types, while not suppressing responsiveness to genuine evidence. Overall, we contribute a metric for benchmarking sycophancy and a mitigation informed by it.

Deep generative models such as diffusion and flow matching are powerful machine learning tools capable of learning and sampling from high-dimensional distributions. They are particularly useful when the training data appears to be concentrated on a submanifold of the data embedding space. For high-energy physics data, consisting of collections of relativistic energy-momentum 4-vectors, this submanifold can enforce extremely strong physically-motivated priors, such as energy and momentum conservation. If these constraints are learned only approximately, rather than exactly, this can inhibit the interpretability and reliability of such generative models. To remedy this deficiency, we introduce generative models which are, by construction, confined at every step of their sampling trajectory to the manifold of massless N-particle Lorentz-invariant phase space in the center-of-momentum frame. In the case of diffusion models, the "pure noise" forward process endpoint corresponds to the uniform distribution on phase space, which provides a clear starting point from which to identify how correlations among the particles emerge during the reverse (de-noising) process. We demonstrate that our models are able to learn both few-particle and many-particle distributions with various singularity structures, paving the way for future interpretability studies using generative models trained on simulated jet data.

Recent advances in video diffusion have enabled the development of "world models" capable of simulating interactive environments. However, these models are largely restricted to single-agent settings, failing to control multiple agents simultaneously in a scene. In this work, we tackle a fundamental issue of action binding in existing video diffusion models, which struggle to associate specific actions with their corresponding subjects. For this purpose, we propose ActionParty, an action controllable multi-subject world model for generative video games. It introduces subject state tokens, i.e. latent variables that persistently capture the state of each subject in the scene. By jointly modeling state tokens and video latents with a spatial biasing mechanism, we disentangle global video frame rendering from individual action-controlled subject updates. We evaluate ActionParty on the Melting Pot benchmark, demonstrating the first video world model capable of controlling up to seven players simultaneously across 46 diverse environments. Our results show significant improvements in action-following accuracy and identity consistency, while enabling robust autoregressive tracking of subjects through complex interactions.

Pretrained Vision Transformers (ViTs) such as DINOv2 and MAE provide generic image features that can be applied to a variety of downstream tasks such as retrieval, classification, and segmentation. However, such representations tend to focus on the most salient visual cues in the image, with no way to direct them toward less prominent concepts of interest. In contrast, Multimodal LLMs can be guided with textual prompts, but the resulting representations tend to be language-centric and lose their effectiveness for generic visual tasks. To address this, we introduce Steerable Visual Representations, a new class of visual representations, whose global and local features can be steered with natural language. While most vision-language models (e.g., CLIP) fuse text with visual features after encoding (late fusion), we inject text directly into the layers of the visual encoder (early fusion) via lightweight cross-attention. We introduce benchmarks for measuring representational steerability, and demonstrate that our steerable visual features can focus on any desired objects in an image while preserving the underlying representation quality. Our method also matches or outperforms dedicated approaches on anomaly detection and personalized object discrimination, exhibiting zero-shot generalization to out-of-distribution tasks.

Language models (LMs) are increasingly extended with new learnable vocabulary tokens for domain-specific tasks, such as Semantic-ID tokens in generative recommendation. The standard practice initializes these new tokens as the mean of existing vocabulary embeddings, then relies on supervised fine-tuning to learn their representations. We present a systematic analysis of this strategy: through spectral and geometric diagnostics, we show that mean initialization collapses all new tokens into a degenerate subspace, erasing inter-token distinctions that subsequent fine-tuning struggles to fully recover. These findings suggest that \emph{token initialization} is a key bottleneck when extending LMs with new vocabularies. Motivated by this diagnosis, we propose the \emph{Grounded Token Initialization Hypothesis}: linguistically grounding novel tokens in the pretrained embedding space before fine-tuning better enables the model to leverage its general-purpose knowledge for novel-token domains. We operationalize this hypothesis as GTI (Grounded Token Initialization), a lightweight grounding stage that, prior to fine-tuning, maps new tokens to distinct, semantically meaningful locations in the pretrained embedding space using only paired linguistic supervision. Despite its simplicity, GTI outperforms both mean initialization and existing auxiliary-task adaptation methods in the majority of evaluation settings across multiple generative recommendation benchmarks, including industry-scale and public datasets. Further analyses show that grounded embeddings produce richer inter-token structure that persists through fine-tuning, corroborating the hypothesis that initialization quality is a key bottleneck in vocabulary extension.

Large Language Models employing Chain-of-Thought reasoning achieve strong performance but suffer from excessive token consumption that inflates inference costs. Existing efficiency methods such as explicit length penalties, difficulty estimators, or multi-stage curricula either degrade reasoning quality or require complex training pipelines. We introduce Batched Contextual Reinforcement, a minimalist, single-stage training paradigm that unlocks efficient reasoning through a simple structural modification: training the model to solve N problems simultaneously within a shared context window, rewarded purely by per-instance accuracy. This formulation creates an implicit token budget that yields several key findings: (1) We identify a novel task-scaling law: as the number of concurrent problems N increases during inference, per-problem token usage decreases monotonically while accuracy degrades far more gracefully than baselines, establishing N as a controllable throughput dimension. (2) BCR challenges the traditional accuracy-efficiency trade-off by demonstrating a "free lunch" phenomenon at standard single-problem inference. Across both 1.5B and 4B model families, BCR reduces token usage by 15.8% to 62.6% while consistently maintaining or improving accuracy across five major mathematical benchmarks. (3) Qualitative analyses reveal emergent self-regulated efficiency, where models autonomously eliminate redundant metacognitive loops without explicit length supervision. (4) Crucially, we empirically demonstrate that implicit budget constraints successfully circumvent the adversarial gradients and catastrophic optimization collapse inherent to explicit length penalties, offering a highly stable, constraint-based alternative for length control. These results prove BCR practical, showing simple structural incentives unlock latent high-density reasoning in LLMs.

When posed with prompts that permit a large number of valid answers, comprehensively generating them is the first step towards satisfying a wide range of users. In this paper, we study methods to elicit a comprehensive set of valid responses. To evaluate this, we introduce \textbf{diversity coverage}, a metric that measures the total quality scores assigned to each \textbf{unique} answer in the predicted answer set relative to the best possible answer set with the same number of answers. Using this metric, we evaluate 18 LLMs, finding no single model dominates at generating diverse responses to a wide range of open-ended prompts. Yet, per each prompt, there exists a model that outperforms all other models significantly at generating a diverse answer set. Motivated by this finding, we introduce a router that predicts the best model for each query. On NB-Wildchat, our trained router outperforms the single best model baseline (26.3% vs $23.8%). We further show generalization to an out-of-domain dataset (NB-Curated) as well as different answer-generation prompting strategies. Our work lays foundation for studying generating comprehensive answers when we have access to a suite of models.

Video color grading is a critical post-production process that transforms flat, log-encoded raw footage into emotionally resonant cinematic visuals. Existing automated methods act as static, black-box executors that directly output edited pixels, lacking both interpretability and the iterative control required by professionals. We introduce LumiVideo, an agentic system that mimics the cognitive workflow of professional colorists through four stages: Perception, Reasoning, Execution, and Reflection. Given only raw log video, LumiVideo autonomously produces a cinematic base grade by analyzing the scene's physical lighting and semantic content. Its Reasoning engine synergizes an LLM's internalized cinematic knowledge with a Retrieval-Augmented Generation (RAG) framework via a Tree of Thoughts (ToT) search to navigate the non-linear color parameter space. Rather than generating pixels, the system compiles the deduced parameters into industry-standard ASC-CDL configurations and a globally consistent 3D LUT, analytically guaranteeing temporal consistency. An optional Reflection loop then allows creators to refine the result via natural language feedback. We further introduce LumiGrade, the first log-encoded video benchmark for evaluating automated grading. Experiments show that LumiVideo approaches human expert quality in fully automatic mode while enabling precise iterative control when directed.

Standard LLM benchmarks evaluate the assistant turn: the model generates a response to an input, a verifier scores correctness, and the analysis ends. This paradigm leaves unmeasured whether the LLM encodes any awareness of what follows the assistant response. We propose user-turn generation as a probe of this gap: given a conversation context of user query and assistant response, we let a model generate under the user role. If the model's weights encode interaction awareness, the generated user turn will be a grounded follow-up that reacts to the preceding context. Through experiments across $11$ open-weight LLMs (Qwen3.5, gpt-oss, GLM) and $5$ datasets (math reasoning, instruction following, conversation), we show that interaction awareness is decoupled from task accuracy. In particular, within the Qwen3.5 family, GSM8K accuracy scales from $41\%$ ($0.8$B) to $96.8\%$ ($397$B-A$17$B), yet genuine follow-up rates under deterministic generation remain near zero. In contrast, higher temperature sampling reveals interaction awareness is latent with follow up rates reaching $22\%$. Controlled perturbations validate that the proposed probe measures a real property of the model, and collaboration-oriented post-training on Qwen3.5-2B demonstrates an increase in follow-up rates. Our results show that user-turn generation captures a dimension of LLM behavior, interaction awareness, that is unexplored and invisible with current assistant-only benchmarks.

Doubly stochastic matrices enable learned mixing across residual streams, but parameterizing the set of doubly stochastic matrices (the Birkhoff polytope) exactly and efficiently remains an open challenge. Existing exact methods scale factorially with the number of streams ($d$), while Kronecker-factorized approaches are efficient but expressivity-limited. We introduce a novel exact parameterization grounded in the theory of generalized orthostochastic matrices, which scales as $\mathcal{O}(d^3)$ and exposes a single hyperparameter $s$ which continuously interpolates between a computationally efficient boundary and the fully expressive Birkhoff polytope. Building on Manifold-Constrained Hyper-Connections ($m$HC), a framework for learned dynamic layer connectivity, we instantiate this parameterization in go-$m$HC. Our method composes naturally with Kronecker-factorized methods, substantially recovering expressivity at similar FLOP costs. Spectral analysis indicates that go-$m$HC fills the Birkhoff polytope far more completely than Kronecker-factorized baselines. On synthetic stream-mixing tasks, go-$m$HC achieves the minimum theoretical loss while converging up to $10\times$ faster. We validate our approach in a 30M parameter GPT-style language model. The expressivity, efficiency, and exactness of go-$m$HC offer a practical avenue for scaling $d$ as a new dimension of model capacity.

Existing video object removal methods excel at inpainting content "behind" the object and correcting appearance-level artifacts such as shadows and reflections. However, when the removed object has more significant interactions, such as collisions with other objects, current models fail to correct them and produce implausible results. We present VOID, a video object removal framework designed to perform physically-plausible inpainting in these complex scenarios. To train the model, we generate a new paired dataset of counterfactual object removals using Kubric and HUMOTO, where removing an object requires altering downstream physical interactions. During inference, a vision-language model identifies regions of the scene affected by the removed object. These regions are then used to guide a video diffusion model that generates physically consistent counterfactual outcomes. Experiments on both synthetic and real data show that our approach better preserves consistent scene dynamics after object removal compared to prior video object removal methods. We hope this framework sheds light on how to make video editing models better simulators of the world through high-level causal reasoning.

Softmax can become a computational bottleneck in the Transformer model's Multi-Head Attention (MHA) block, particularly in small models under low-precision inference, where exponentiation and normalization incur significant overhead. As such, we suggest using Head-Calibrated Clipped-Linear Softmax (HCCS), a bounded, monotone surrogate to the exponential softmax function, which uses a clipped linear mapping of the max centered attention logits. This approximation produces a stable probability distribution, maintains the ordering of the original logits and has non-negative values. HCCS differs from previous softmax surrogates as it includes a set of lightweight calibration parameters that are optimized offline based on a representative dataset and calibrated for each individual attention head to preserve the statistical properties of the individual heads. We describe a hardware-motivated implementation of HCCS for high-throughput scenarios targeting the AMD Versal AI Engines. The current reference implementations from AMD for this platform rely upon either bfloat16 arithmetic or LUTs to perform the exponential operation, which might limit the throughput of the platform and fail to utilize the high-throughput integer vector processing units of the AI Engine. In contrast, HCCS provides a natural mapping to the AI Engines' int8 multiply accumulate (MAC) units. To the best of our knowledge, this is the first int8 optimized softmax surrogate for AMD AI engines that significantly exceeds the speed performance of other reference implementations while maintaining competitive task accuracy on small or heavily quantized MHA workloads after quantization-aware retraining.

The upcoming integration of AI in the physical layer (PHY) of 6G radio access networks (RAN) will enable a higher quality of service in challenging transmission scenarios. However, deeply optimized AI-Native PHY models impose higher computational complexity compared to conventional baseband, challenging deployment under the sub-msec real-time constraints typical of modern PHYs. Additionally, following the extension to terahertz carriers, the upcoming densification of 6G cell-sites further limits the power consumption of base stations, constraining the budget available for compute ($\leq$ 100W). The desired flexibility to ensure long term sustainability and the imperative energy-efficiency gains on the high-throughput tensor computations dominating AI-Native PHYs can be achieved by domain-specialization of many-core programmable baseband processors. Following the domain-specialization strategy, we present TensorPool, a cluster of 256 RISCV32IMAF programmable cores, accelerated by 16 256 MACs/cycle (FP16) tensor engines with low-latency access to 4MiB of L1 scratchpad for maximal data-reuse. Implemented in TSMC's N7, TensorPool achieves 3643~MACs/cycle (89% tensor-unit utilization) on tensor operations for AI-RAN, 6$\times$ more than a core-only cluster without tensor acceleration, while simultaneously improving GOPS/W/mm$^2$ efficiency by 9.1$\times$. Further, we show that 3D-stacking the computing blocks of TensorPool to better unfold the tensor engines to L1-memory routing provides 2.32$\times$ footprint improvement with no frequency degradation, compared to a 2D implementation.

Recent multimodal large language models have achieved strong performance in unified text and image understanding and generation, yet extending such native capability to 3D remains challenging due to limited data. Compared to abundant 2D imagery, high-quality 3D assets are scarce, making 3D synthesis under-constrained. Existing methods often rely on indirect pipelines that edit in 2D and lift results into 3D via optimization, sacrificing geometric consistency. We present Omni123, a 3D-native foundation model that unifies text-to-2D and text-to-3D generation within a single autoregressive framework. Our key insight is that cross-modal consistency between images and 3D can serve as an implicit structural constraint. By representing text, images, and 3D as discrete tokens in a shared sequence space, the model leverages abundant 2D data as a geometric prior to improve 3D representations. We introduce an interleaved X-to-X training paradigm that coordinates diverse cross-modal tasks over heterogeneous paired datasets without requiring fully aligned text-image-3D triplets. By traversing semantic-visual-geometric cycles (e.g., text to image to 3D to image) within autoregressive sequences, the model jointly enforces semantic alignment, appearance fidelity, and multi-view geometric consistency. Experiments show that Omni123 significantly improves text-guided 3D generation and editing, demonstrating a scalable path toward multimodal 3D world models.

Reinforcement learning with verifiable rewards (RLVR) has become a standard paradigm for post-training large language models. While Group Relative Policy Optimization (GRPO) is widely adopted, its coarse credit assignment uniformly penalizes failed rollouts, lacking the token-level focus needed to efficiently address specific deviations. Self-Distillation Policy Optimization (SDPO) addresses this by providing denser, more targeted logit-level supervision that facilitates rapid early improvement, yet it frequently collapses during prolonged training. We trace this late-stage instability to two intrinsic flaws: self-distillation on already-correct samples introduces optimization ambiguity, and the self-teacher's signal reliability progressively degrades. To resolve these issues, we propose Sample-Routed Policy Optimization (SRPO), a unified on-policy framework that routes correct samples to GRPO's reward-aligned reinforcement and failed samples to SDPO's targeted logit-level correction. SRPO further incorporates an entropy-aware dynamic weighting mechanism to suppress high-entropy, unreliable distillation targets while emphasizing confident ones. Evaluated across five benchmarks and two model scales, SRPO achieves both the rapid early improvement of SDPO and the long-horizon stability of GRPO. It consistently surpasses the peak performance of both baselines, raising the five-benchmark average on Qwen3-8B by 3.4% over GRPO and 6.3% over SDPO, while simultaneously yielding moderate response lengths and lowering per-step compute cost by up to 17.2%.

Long-horizon conversational agents require persistent memory for coherent reasoning, yet uncontrolled accumulation causes temporal decay and false memory propagation. Benchmarks such as LOCOMO and LOCCO report performance degradation from 0.455 to 0.05 across stages, while MultiWOZ shows 78.2% accuracy with 6.8% false memory rate under persistent retention. This work introduces an adaptive budgeted forgetting framework that regulates memory through relevanceguided scoring and bounded optimization. The approach integrates recency, frequency, and semantic alignment to maintain stability under constrained context. Comparative analysis demonstrates improved long-horizon F1 beyond 0.583 baseline levels, higher retention consistency, and reduced false memory behavior without increasing context usage. These findings confirm that structured forgetting preserves reasoning performance while preventing unbounded memory growth in extended conversational settings.

Agentic AI shifts the investor's role from analytical execution to oversight. We present an agentic strategic asset allocation pipeline in which approximately 50 specialized agents produce capital market assumptions, construct portfolios using over 20 competing methods, and critique and vote on each other's output. A researcher agent proposes new portfolio construction methods not yet represented, and a meta-agent compares past forecasts against realized returns and rewrites agent code and prompts to improve future performance. The entire pipeline is governed by the Investment Policy Statement--the same document that guides human portfolio managers can now constrain and direct autonomous agents.

Translating machine code into human-readable high-level languages is an open research problem in reverse engineering. Despite recent advancements in LLM-based decompilation to C, modern languages like Dart and Swift are unexplored. In this paper, we study the use of small specialized LLMs as an idiomatic decompiler for such languages. Additionally, we investigate the augmentation of training data using synthetic same-language examples, and compare it against adding human-written examples using related-language (Swift -> Dart). We apply CODEBLEU to evaluate the decompiled code readability and compile@k to measure the syntax correctness. Our experimental results show that on a 73-function Dart test dataset (representing diverse complexity levels), our 4B specialized model achieves 71.3 CODEBLEU (95% CI 65.5-77.1), approximately comparable to a ~480B code model (73.1; 67.4-78.8). On a subset of 34 natural Dart functions, it reaches compile@k5 = 79.4% (Wilson 95% CI 63.2-89.7), vs. 64.7% (47.9-78.5) for the base model; the difference is suggestive but not statistically significant at 0.05. Our results indicate that adding Swift training data helps at 8B but not at 4B, suggesting a capacity threshold for effective cross-lingual transfer. Our experimental results show that small specialized models can generate readable, idiomatic Dart with meaningful identifiers while using minimal compute.

Regulatory documents encode legally binding obligations that LLM-based systems must respect. Yet converting dense, hierarchically structured legal text into machine-readable rules remains a costly, expert-intensive process. We present De Jure, a fully automated, domain-agnostic pipeline for extracting structured regulatory rules from raw documents, requiring no human annotation, domain-specific prompting, or annotated gold data. De Jure operates through four sequential stages: normalization of source documents into structured Markdown; LLM-driven semantic decomposition into structured rule units; multi-criteria LLM-as-a-judge evaluation across 19 dimensions spanning metadata, definitions, and rule semantics; and iterative repair of low-scoring extractions within a bounded regeneration budget, where upstream components are repaired before rule units are evaluated. We evaluate De Jure across four models on three regulatory corpora spanning finance, healthcare, and AI governance. On the finance domain, De Jure yields consistent and monotonic improvement in extraction quality, reaching peak performance within three judge-guided iterations. De Jure generalizes effectively to healthcare and AI governance, maintaining high performance across both open- and closed-source models. In a downstream compliance question-answering evaluation via RAG, responses grounded in De Jure extracted rules are preferred over prior work in 73.8% of cases at single-rule retrieval depth, rising to 84.0% under broader retrieval, confirming that extraction fidelity translates directly into downstream utility. These results demonstrate that explicit, interpretable evaluation criteria can substitute for human annotation in complex regulatory domains, offering a scalable and auditable path toward regulation-grounded LLM alignment.

Generative models for crystalline materials often rely on equivariant graph neural networks, which capture geometric structure well but are costly to train and slow to sample. We present Crystalite, a lightweight diffusion Transformer for crystal modeling built around two simple inductive biases. The first is Subatomic Tokenization, a compact chemically structured atom representation that replaces high-dimensional one-hot encodings and is better suited to continuous diffusion. The second is the Geometry Enhancement Module (GEM), which injects periodic minimum-image pair geometry directly into attention through additive geometric biases. Together, these components preserve the simplicity and efficiency of a standard Transformer while making it better matched to the structure of crystalline materials. Crystalite achieves state-of-the-art results on crystal structure prediction benchmarks, and de novo generation performance, attaining the best S.U.N. discovery score among the evaluated baselines while sampling substantially faster than geometry-heavy alternatives.

Agent skills, structured packages of procedural knowledge and executable resources that agents dynamically load at inference time, have become a reliable mechanism for augmenting LLM agents. Yet inference-time skill augmentation is fundamentally limited: retrieval noise introduces irrelevant guidance, injected skill content imposes substantial token overhead, and the model never truly acquires the knowledge it merely follows. We ask whether skills can instead be internalized into model parameters, enabling zero-shot autonomous behavior without any runtime skill retrieval. We introduce SKILL0, an in-context reinforcement learning framework designed for skill internalization. SKILL0 introduces a training-time curriculum that begins with full skill context and progressively withdraws it. Skills are grouped offline by category and rendered with interaction history into a compact visual context, teaching he model tool invocation and multi-turn task completion. A Dynamic Curriculum then evaluates each skill file's on-policy helpfulness, retaining only those from which the current policy still benefits within a linearly decaying budget, until the agent operates in a fully zero-shot setting. Extensive agentic experiments demonstrate that SKILL0 achieves substantial improvements over the standard RL baseline (+9.7\% for ALFWorld and +6.6\% for Search-QA), while maintaining a highly efficient context of fewer than 0.5k tokens per step. Our code is available at https://github.com/ZJU-REAL/SkillZero.

Learning-based control methods typically assume stationary system dynamics, an assumption often violated in real-world systems due to drift, wear, or changing operating conditions. We study reinforcement learning for control under time-varying dynamics. We consider a continual model-based reinforcement learning setting in which an agent repeatedly learns and controls a dynamical system whose transition dynamics evolve across episodes. We analyze the problem using Gaussian process dynamics models under frequentist variation-budget assumptions. Our analysis shows that persistent non-stationarity requires explicitly limiting the influence of outdated data to maintain calibrated uncertainty and meaningful dynamic regret guarantees. Motivated by these insights, we propose a practical optimistic model-based reinforcement learning algorithm with adaptive data buffer mechanisms and demonstrate improved performance on continuous control benchmarks with non-stationary dynamics.

In this paper, we consider a multi-armed bandit (MAB) instance and study how to identify the best arm when arm commands are conveyed from a central learner to a distributed agent over a discrete memoryless channel (DMC). Depending on the agent capabilities, we provide communication schemes along with their analysis, which interestingly relate to the zero-error capacity of the underlying DMC.

Understanding causal dependencies in observational data is critical for informing decision-making. These relationships are often modeled as Bayesian Networks (BNs) and Directed Acyclic Graphs (DAGs). Existing methods, such as NOTEARS and DAG-GNN, often face issues with scalability and stability in high-dimensional data, especially when there is a feature-sample imbalance. Here, we show that the denoising score matching objective of diffusion models could smooth the gradients for faster, more stable convergence. We also propose an adaptive k-hop acyclicity constraint that improves runtime over existing solutions that require matrix inversion. We name this framework Denoising Diffusion Causal Discovery (DDCD). Unlike generative diffusion models, DDCD utilizes the reverse denoising process to infer a parameterized causal structure rather than to generate data. We demonstrate the competitive performance of DDCDs on synthetic benchmarking data. We also show that our methods are practically useful by conducting qualitative analyses on two real-world examples. Code is available at this url: https://github.com/haozhu233/ddcd.

Multimodal time-to-event prediction often requires integrating sensitive data distributed across multiple parties, making centralized model training impractical due to privacy constraints. At the same time, most existing multimodal survival models produce single deterministic predictions without indicating how confident the model is in its estimates, which can limit their reliability in real-world decision making. To address these challenges, we propose BVFLMSP, a Bayesian Vertical Federated Learning (VFL) framework for multimodal time-to-event analysis based on a Split Neural Network architecture. In BVFLMSP, each client independently models a specific data modality using a Bayesian neural network, while a central server aggregates intermediate representations to perform survival risk prediction. To enhance privacy, we integrate differential privacy mechanisms by perturbing client side representations before transmission, providing formal privacy guarantees against information leakage during federated training. We first evaluate our Bayesian multimodal survival model against widely used single modality survival baselines and the centralized multimodal baseline MultiSurv. Across multimodal settings, the proposed method shows consistent improvements in discrimination performance, with up to 0.02 higher C-index compared to MultiSurv. We then compare federated and centralized learning under varying privacy budgets across different modality combinations, highlighting the tradeoff between predictive performance and privacy. Experimental results show that BVFLMSP effectively includes multimodal data, improves survival prediction over existing baselines, and remains robust under strict privacy constraints while providing uncertainty estimates.

This is an extended version of our publication Learning state machines from data streams: A generic strategy and an improved heuristic, International Conference on Grammatical Inference (ICGI) 2023, Rabat, Morocco. It has been extended with a formal proof on PAC-bounds, and the discussion and analysis of a similar approach has been moved from the appendix and is now a full Section. State machine models are models that simulate the behavior of discrete event systems, capable of representing systems such as software systems, network interactions, and control systems, and have been researched extensively. The nature of most learning algorithms however is the assumption that all data be available at the beginning of the algorithm, and little research has been done in learning state machines from streaming data. In this paper, we want to close this gap further by presenting a generic method for learning state machines from data streams, as well as a merge heuristic that uses sketches to account for incomplete prefix trees. We implement our approach in an open-source state merging library and compare it with existing methods. We show the effectiveness of our approach with respect to run-time, memory consumption, and quality of results on a well known open dataset. Additionally, we provide a formal analysis of our algorithm, showing that it is capable of learning within the PAC framework, and show a theoretical improvement to increase run-time, without sacrificing correctness of the algorithm in larger sample sizes.

Generative AI (GAI) reveals an irreducible human core at the center of data science: advances in GAI should sharpen, rather than diminish, the focus on human reasoning in data science education. GAI can now execute many routine data science workflows, including cleaning, summarizing, visualizing, modeling, and drafting reports. Yet the competencies that matter most remain irreducibly human: problem formulation, measurement and design, causal identification, statistical and computational reasoning, ethics and accountability, and sensemaking. Drawing on Donoho's Greater Data Science framework, Nolan and Temple Lang's vision of computational literacy, and the McLuhan-Culkin insight that we shape our tools and thereafter our tools shape us, this paper traces the emergence of data science through three converging lineages: Tukey's intellectual vision of data analysis as a science, the commercial logic of surveillance capitalism that created industrial demand for data scientists, and the academic programs that followed. Mapping GAI's impact onto Donoho's six divisions of Greater Data Science shows that computing with data (GDS3) has been substantially automated, while data gathering, preparation, and exploration (GDS1) and science about data science (GDS6) still require essential human input. The educational implication is that data science curricula should focus on this human core while teaching students how to contribute effectively within iterative prompt-output-prompt cycles using retrieval-augmented generation, and that learning outcomes and assessments should explicitly evaluate reasoning and judgment.

Emotional tone is pervasive in human communication, yet its influence on large language model (LLM) behaviour remains unclear. Here, we examine how first-person emotional framing in user-side queries affect LLM performance across six benchmark domains, including mathematical reasoning, medical question answering, reading comprehension, commonsense reasoning and social inference. Across models and tasks, static emotional prefixes usually produce only small changes in accuracy, suggesting that affective phrasing is typically a mild perturbation rather than a reliable general-purpose intervention. This stability is not uniform: effects are more variable in socially grounded tasks, where emotional context more plausibly interacts with interpersonal reasoning. Additional analyses show that stronger emotional wording induces only modest extra change, and that human-written prefixes reproduce the same qualitative pattern as LLM-generated ones. We then introduce EmotionRL, an adaptive emotional prompting framework that selects emotional framing adaptively for each query. Although no single emotion is consistently beneficial, adaptive selection yields more reliable gains than fixed emotional prompting. Together, these findings show that emotional tone is neither a dominant driver of LLM performance nor irrelevant noise, but a weak and input-dependent signal that can be exploited through adaptive control.

For Large Language Models (LLMs) to be reliably deployed, models must effectively know when not to answer: abstain. Reasoning models, in particular, have gained attention for impressive performance on complex tasks. However, reasoning models have been shown to have worse abstention abilities. Taking the vulnerabilities of reasoning models into account, we propose our Query Misalignment Framework. Hallucinations resulting in failed abstention can be reinterpreted as LLMs answering the wrong question (rather than answering a question incorrectly). Based on this framework, we develop a new class of state-of-the-art abstention methods called Trace Inversion. First, we generate the reasoning trace of a model. Based on only the trace, we then reconstruct the most likely query that the model responded to. Finally, we compare the initial query with the reconstructed query. Low similarity score between the initial query and reconstructed query suggests that the model likely answered the question incorrectly and is flagged to abstain. Extensive experiments demonstrate that Trace Inversion effectively boosts abstention performance in four frontier LLMs across nine abstention QA datasets, beating competitive baselines in 33 out of 36 settings.

Reinforcement learning (RL) agents often struggle with out-of-distribution (OOD) scenarios, leading to high uncertainty and random behavior. While language models (LMs) contain valuable world knowledge, larger ones incur high computational costs, hindering real-time use, and exhibit limitations in autonomous planning. We introduce Adaptive Safety through Knowledge (ASK), which combines smaller LMs with trained RL policies to enhance OOD generalization without retraining. ASK employs Monte Carlo Dropout to assess uncertainty and queries the LM for action suggestions only when uncertainty exceeds a set threshold. This selective use preserves the efficiency of existing policies while leveraging the language model's reasoning in uncertain situations. In experiments on the FrozenLake environment, ASK shows no improvement in-domain, but demonstrates robust navigation in transfer tasks, achieving a reward of 0.95. Our findings indicate that effective neuro-symbolic integration requires careful orchestration rather than simple combination, highlighting the need for sufficient model scale and effective hybridization mechanisms for successful OOD generalization.

This paper establishes the theoretical and practical foundations for using Large Language Models (LLMs) as measurement instruments for latent economic variables -- specifically variables that describe the cognitive content of occupational tasks at a level of granularity not achievable with existing survey instruments. I formalize four conditions under which LLM-generated scores constitute valid instruments: semantic exogeneity, construct relevance, monotonicity, and model invariance. I then apply this framework to the Augmented Human Capital Index (AHC_o), constructed from 18,796 O*NET task statements scored by Claude Haiku 4.5, and validated against six existing AI exposure indices. The index shows strong convergent validity (r = 0.85 with Eloundou GPT-gamma, r = 0.79 with Felten AIOE) and discriminant validity. Principal component analysis confirms that AI-related occupational measures span two distinct dimensions -- augmentation and substitution. Inter-rater reliability across two LLM models (n = 3,666 paired scores) yields Pearson r = 0.76 and Krippendorff's alpha = 0.71. Prompt sensitivity analysis across four alternative framings shows that task-level rankings are robust. Obviously Related Instrumental Variables (ORIV) estimation recovers coefficients 25% larger than OLS, consistent with classical measurement error attenuation. The methodology generalizes beyond labor economics to any domain where semantic content must be quantified at scale.

Multimodal Large Language Models (MLLMs) offer an opportunity to support multimedia learning through conversational systems grounded in educational content. However, while conversational AI is known to boost engagement, its impact on learning in visually-rich STEM domains remains under-explored. Moreover, there is limited understanding of how multimodality and conversationality jointly influence learning in generative AI systems. This work reports findings from a randomized controlled online study (N = 124) comparing three approaches to learning biology from textbook content: (1) a document-grounded conversational AI with interleaved text-and-image responses (MuDoC), (2) a document-grounded conversational AI with text-only responses (TexDoC), and (3) a textbook interface with semantic search and highlighting (DocSearch). Learners using MuDoC achieved the highest post-test scores and reported the most positive learning experience. Notably, while TexDoC was rated as significantly more engaging and easier to use than DocSearch, it led to the lowest post-test scores, revealing a disconnect between student perceptions and learning outcomes. Interpreted through the lens of the Cognitive Load Theory, these findings suggest that conversationality reduces extraneous load, while visual-verbal integration induced by multimodality increases germane load, leading to better learning outcomes. When conversationality is not complemented by multimodality, reduced cognitive effort may instead inflate perceived understanding without improving learning outcomes.

Understanding how Large Language Models (LLMs) process information from prompts remains a significant challenge. To shed light on this "black box," attention visualization techniques have been developed to capture neuron-level perceptions and interpret how models focus on different parts of input data. However, many existing techniques are tailored to specific model architectures, particularly within the Transformer family, and often require backpropagation, resulting in nearly double the GPU memory usage and increased computational cost. A lightweight, model-agnostic approach for attention visualization remains lacking. In this paper, we introduce a model-agnostic token importance visualization technique to better understand how generative AI systems perceive and prioritize information from input text, without incurring additional computational cost. Our method leverages perturbation-based strategies combined with a three-matrix analytical framework to generate relevance maps that illustrate token-level contributions to model predictions. The framework comprises: (1) the Angular Deviation Matrix, which captures shifts in semantic direction; (2) the Magnitude Deviation Matrix, which measures changes in semantic intensity; and (3) the Dimensional Importance Matrix, which evaluates contributions across individual vector dimensions. By systematically removing each token and measuring the resulting impact across these three complementary dimensions, we derive a composite importance score that provides a nuanced and mathematically grounded measure of token significance. To support reproducibility and foster wider adoption, we provide open-source implementations of all proposed and utilized explainability techniques, with code and resources publicly available at https://github.com/Infosys/Infosys-Responsible-AI-Toolkit

Conventional hypernetworks are typically engineered around a specific base-model parameterization, so changing the target architecture often entails redesigning the hypernetwork and retraining it from scratch. We introduce the \emph{Universal Hypernetwork} (UHN), a fixed-architecture generator that predicts weights from deterministic parameter, architecture, and task descriptors. This descriptor-based formulation decouples the generator architecture from target-network parameterization, so one generator can instantiate heterogeneous models across the tested architecture and task families. Our empirical claims are threefold: (1) one fixed UHN remains competitive with direct training across vision, graph, text, and formula-regression benchmarks; (2) the same UHN supports both multi-model generalization within a family and multi-task learning across heterogeneous models; and (3) UHN enables stable recursive generation with up to three intermediate generated UHNs before the final base model. Our code is available at https://github.com/Xuanfeng-Zhou/UHN.

Safe Rust guarantees memory safety through strict compile-time constraints: ownership can be transferred, borrowing can temporarily guarantee either shared read-only or exclusive write access, and ownership and borrowing are scoped by lifetime. Automatically synthesizing correct and safe Rust code is challenging, as the generated code must not only satisfy ownership, borrowing, and lifetime constraints, but also meet type and interface requirements at compile time. This work proposes a synthesis method based on our newly defined Pushdown Colored Petri Net (PCPN) that models these compilation constraints directly from public API signatures to synthesize valid call sequences. Token colors encode dynamic resource states together with a scope level indicating the lifetime region in which a borrow is valid. The pushdown stack tracks the entering or leaving of lifetime parameter via pushing and popping tokens. A transition is enabled only when type matching and interface obligations both hold and the required resource states are available. Based on the bisimulation theory, we prove that the enabling and firing rules of PCPN are consistent with the compile-time check of these three constraints. We develop an automatic synthesis tool based on PCPN and the experimental results show that the synthesized codes are all correct.

Data races in GPU programs pose a threat to the reliability of GPU-accelerated software stacks. Prior works proposed various dynamic (runtime) and static (compile-time) techniques to detect races in GPU programs. However, dynamic techniques often miss critical races, as they require the races to manifest during testing. While static ones can catch such races, they often generate numerous false alarms by conservatively assuming values of variables/parameters that cannot ever occur during any execution of the program. We make a key observation that the host (CPU) code that launches GPU kernels contains crucial semantic information about the values that the GPU kernel's parameters can take during execution. Harnessing this hitherto overlooked information helps accurately detect data races in GPU kernel code. We create HGRD, a new state-of-the-art static analysis technique that performs a holistic analysis of both CPU and GPU code to accurately detect a broad set of true races while minimizing false alarms. While SOTA dynamic techniques, such as iGUARD, miss many true races, HGRD misses none. On the other hand, static techniques such as GPUVerify and FaialAA raise tens of false alarms, where HGRD raises none.

Message Passing Interface (MPI) is a foundational technology in high-performance computing (HPC), widely used for large-scale simulations and distributed training (e.g., in machine learning frameworks such as PyTorch and TensorFlow). However, maintaining MPI programs remains challenging due to their complex interplay among processes and the intricacies of message passing and synchronization. With the advancement of large language models like ChatGPT, it is tempting to adopt such technology for automated error detection and repair. Yet, our studies reveal that directly applying large language models (LLMs) yields suboptimal results, largely because these models lack essential knowledge about correct and incorrect usage, particularly the bugs found in MPI programs. In this paper, we design a bug detection and repair technique alongside Few-Shot Learning (FSL), Chain-of-Thought (CoT) reasoning, and Retrieval Augmented Generation (RAG) techniques in LLMs to enhance the large language model's ability to detect and repair errors. Surprisingly, such enhancements lead to a significant improvement, from 44% to 77%, in error detection accuracy compared to baseline methods that use ChatGPT directly. Additionally, our experiments demonstrate our bug referencing technique generalizes well to other large language models.

Group Emotion Recognition (GER) aims to infer collective affect in social environments such as classrooms, crowds, and public events. Many existing approaches rely on explicit individual-level processing, including cropped faces, person tracking, or per-person feature extraction, which makes the analysis pipeline person-centric and raises privacy concerns in deployment scenarios where only group-level understanding is needed. This research proposes VE-MD, a Variational Encoder-Multi-Decoder framework for group emotion recognition under a privacy-aware functional design. Rather than providing formal anonymization or cryptographic privacy guarantees, VE-MD is designed to avoid explicit individual monitoring by constraining the model to predict only aggregate group-level affect, without identity recognition or per-person emotion outputs. VE-MD learns a shared latent representation jointly optimized for emotion classification and internal prediction of body and facial structural representations. Two structural decoding strategies are investigated: a transformer-based PersonQuery decoder and a dense Heatmap decoder that naturally accommodates variable group sizes. Experiments on six in-the-wild datasets, including two GER and four Individual Emotion Recognition (IER) benchmarks, show that structural supervision consistently improves representation learning. More importantly, the results reveal a clear distinction between GER and IER: optimizing the latent space alone is often insufficient for GER because it tends to attenuate interaction-related cues, whereas preserving explicit structural outputs improves collective affect inference. In contrast, projected structural representations seem to act as an effective denoising bottleneck for IER. VE-MD achieves state-of-the-art performance on GAF-3.0 (up to 90.06%) and VGAF (82.25% with multimodal fusion with audio). These results show that preserving interaction-related structural information is particularly beneficial for group-level affect modeling without relying on prior individual feature extraction. On IER datasets using multimodal fusion with audio modality, VE-MD outperforms SOTA on SamSemo (77.9%, adding text modality) while achieving competitive performances on MER-MULTI (63.8%), DFEW (70.7%) and EngageNet (69.0).

The deep integration of communication with intelligence and sensing, as a defining vision of 6G, renders environment-aware channel prediction a key enabling technology. As a representative 6G application, vehicular communications require accurate and forward-looking channel prediction under stringent reliability, latency, and adaptability demands. Traditional empirical and deterministic models remain limited in balancing accuracy, generalization, and deployability, while the growing availability of onboard and roadside sensing devices offers a promising source of environmental priors. This paper proposes an environment-aware channel prediction framework based on multimodal visual feature fusion. Using GPS data and vehicle-side panoramic RGB images, together with semantic segmentation and depth estimation, the framework extracts semantic, depth, and position features through a three-branch architecture and performs adaptive multimodal fusion via a squeeze-excitation attention gating module. For 360-dimensional angular power spectrum (APS) prediction, a dedicated regression head and a composite multi-constraint loss are further designed. As a result, joint prediction of path loss (PL), delay spread (DS), azimuth spread of arrival (ASA), azimuth spread of departure (ASD), and APS is achieved. Experiments on a synchronized urban V2I measurement dataset yield the best root mean square error (RMSE) of 3.26 dB for PL, RMSEs of 37.66 ns, 5.05 degrees, and 5.08 degrees for DS, ASA, and ASD, respectively, and mean/median APS cosine similarities of 0.9342/0.9571, demonstrating strong accuracy, generalization, and practical potential for intelligent channel prediction in 6G vehicular communications.

We study illusion elimination problems on directed social networks where each vertex is colored either red or blue. A vertex is under \textit{majority illusion} if it has more red out-neighbors than blue out-neighbors when there are more blue vertices than red ones in the network. In a more general phenomenon of $p$-illusion, at least $p$ fraction of the out-neighbors (as opposed to $1/2$ for majority) of a vertex is red. In the directed illusion elimination problem, we recolor minimum number of vertices so that no vertex is under $p$-illusion, for $p\in (0,1)$. Unfortunately, the problem is NP-hard for $p =1/2$ even when the network is a grid. Moreover, the problem is NP-hard and W[2]-hard when parameterized by the number of recolorings for each $p \in (0,1)$ even on bipartite DAGs. Thus, we can neither get a polynomial time algorithm on DAGs, unless P=NP, nor we can get a FPT algorithm even by combining solution size and directed graph parameters that measure distance from acyclicity, unless FPT=W[2]. We show that the problem can be solved in polynomial time in structured, sparse networks such as outerplanar networks, outward grids, trees, and cycles. Finally, we show tractable algorithms parameterized by treewidth of the underlying undirected graph, and by the number of vertices under illusion.

Like a body at rest that stays at rest, we find that visual attention in multimodal large language models (MLLMs) exhibits pronounced inertia, remaining largely static once settled during early decoding steps and failing to support the compositional understanding required for cognitive inference. While existing hallucination mitigation methods mainly target perceptual hallucinations concerning object existence or attributes, they remain inadequate for such cognitive hallucinations that require inter-object relational deduction. Through token-wise attention analysis, we identify this visual inertia as a key factor: attention to semantically critical regions remains persistently focused and fails to dynamically support relational inference. We thereby propose a training-free Inertia-aware Visual Excitation (IVE) method that breaks this inertial pattern by modeling cognitive inference as the dynamic responsiveness of visual attention. Specifically, IVE selects visual tokens that are dynamically emerging relative to historical attention trends while distinguishing tokens exhibiting inertial behavior. To further facilitate compositional inference, IVE introduces an inertia-aware penalty that discourages over-concentration and limits the persistence of attention within localized regions. Extensive experiments show that IVE is effective across various base MLLMs and multiple hallucination benchmarks, particularly for cognitive hallucinations.

The scale of biological datasets now routinely exceeds system memory, making data access rather than model computation the primary bottleneck in training machine-learning models. This bottleneck is particularly acute in biology, where widely used community data formats must support heterogeneous metadata, sparse and dense assays, and downstream analysis within established computational ecosystems. Here we present annbatch, a mini-batch loader native to anndata that enables out-of-core training directly on disk-backed datasets. Across single-cell transcriptomics, microscopy and whole-genome sequencing benchmarks, annbatch increases loading throughput by up to an order of magnitude and shortens training from days to hours, while remaining fully compatible with the scverse ecosystem. Annbatch establishes a practical data-loading infrastructure for scalable biological AI, allowing increasingly large and diverse datasets to be used without abandoning standard biological data formats. Github: https://github.com/scverse/annbatch

Vanishing gradient and overfitting are two of the most extensively studied problems in the literature about machine learning. However, they are frequently considered in some asymptotic setting, which obscure the underlying dynamical mechanisms responsible for their emergence. In this paper, we aim to provide a clear dynamical description of learning in multi-layer perceptrons. To this end, we introduce a minimal model, inspired by studies by Fukumizu and Amari, to investigate vanishing gradients and overfitting in MLPs trained via gradient descent. Within this model, we show that the learning dynamics may pass through plateau regions and near-optimal regions during training, both of which consist of saddle structures, before ultimately converging to the overfitting region. Under suitable conditions on the training dataset, we prove that, with high probability, the overfitting region collapses to a single attractor modulo symmetry, which corresponds to the overfitting. Moreover, we show that any MLP trained on a finite noisy dataset cannot converge to the theoretical optimum and instead necessarily converges to an overfitting solution.

The ratio of outlier parameters in language pre-training models and vision pre-training models differs significantly, making cross-modality (language and vision) inherently more challenging than cross-domain adaptation. As a result, many prior studies have focused on cross-domain transfer rather than attempting to bridge language and vision modalities, assuming that language pre-trained models are unsuitable for downstream visual tasks due to disparate parameter spaces. Contrary to this assumption, we show that adding a bridge training stage as a modality adaptation learner can effectively align Large Language Model (LLM) parameters with vision tasks. Specifically, we propose a simple yet powerful solution random label bridge training that requires no manual labeling and helps LLM parameters adapt to vision foundation tasks. Moreover, our findings reveal that partial bridge training is often advantageous, as certain layers in LLMs exhibit strong foundational properties that remain beneficial even without fine-tuning for visual tasks. This surprising discovery opens up new avenues for leveraging language pre-trained parameters directly within vision models and highlights the potential of partial bridge training as a practical pathway to cross-modality adaptation.

Cutting rectangular items from stock sheets to satisfy demands while minimizing waste is a central manufacturing task. The Two-Dimensional Single Stock Size Cutting Stock Problem (2D-CSSP) generalizes bin packing by requiring multiple copies of each item type, which causes a strong combinatorial blow-up. We present a SAT-based framework where item types are expanded by demand, each copy has a sheet-assignment variable and non-overlap constraints are activated only for copies assigned to the same sheet. We also introduce an infeasible-orientation elimination rule that fixes rotation variables when only one orientation can fit the sheet. For minimizing the number of sheets, we compare three approaches: non-incremental SAT with binary search, incremental SAT with clause reuse across iterations and weighted partial MaxSAT. On the Cui--Zhao benchmark suite, our best SAT configurations certify two to three times more instances as provably optimal and achieve lower optimality gaps than OR-Tools, CPLEX and Gurobi. The relative ranking among SAT approaches depends on rotation: incremental SAT is strongest without rotation, while non-incremental SAT is more effective when rotation increases formula size.

Audio-Visual Navigation (AVN) requires an embodied agent to navigate toward a sound source by utilizing both vision and binaural audio. A core challenge arises in complex acoustic environments, where binaural cues become intermittently unreliable, particularly when generalizing to previously unheard sound categories. To address this, we propose RAVN (Reliability-Aware Audio-Visual Navigation), a framework that conditions cross-modal fusion on audio-derived reliability cues, dynamically calibrating the integration of audio and visual inputs. RAVN introduces an Acoustic Geometry Reasoner (AGR) that is trained with geometric proxy supervision. Using a heteroscedastic Gaussian NLL objective, AGR learns observation-dependent dispersion as a practical reliability cue, eliminating the need for geometric labels during inference. Additionally, we introduce Reliability-Aware Geometric Modulation (RAGM), which converts the learned cue into a soft gate to modulate visual features, thereby mitigating cross-modal conflicts. We evaluate RAVN on SoundSpaces using both Replica and Matterport3D environments, and the results show consistent improvements in navigation performance, with notable robustness in the challenging unheard sound setting.

Audio-visual navigation tasks require agents to locate and navigate toward continuously vocalizing targets using only visual observations and acoustic cues. However, existing methods mainly rely on simple feature concatenation or late fusion, and lack an explicit discrete representation of the target's relative position, which limits learning efficiency and generalization. We propose Spatial-Aware Conditioned Fusion (SACF). SACF first discretizes the target's relative direction and distance from audio-visual cues, predicts their distributions, and encodes them as a compact descriptor for policy conditioning and state modeling. Then, SACF uses audio embeddings and spatial descriptors to generate channel-wise scaling and bias to modulate visual features via conditional linear transformation, producing target-oriented fused representations. SACF improves navigation efficiency with lower computational overhead and generalizes well to unheard target sounds.

Audio-visual Navigation refers to an agent utilizing visual and auditory information in complex 3D environments to accomplish target localization and path planning, thereby achieving autonomous navigation. The core challenge of this task lies in the following: how the agent can break free from the dependence on training data and achieve autonomous navigation with good generalization performance when facing changes in environments and sound sources. To address this challenge, we propose an Audio Spatially-Guided Fusion for Audio-Visual Navigation method. First, we design an audio spatial feature encoder, which adaptively extracts target-related spatial state information through an audio intensity attention mechanism; based on this, we introduce an Audio Spatial State Guided Fusion (ASGF) to achieve dynamic alignment and adaptive fusion of multimodal features, effectively alleviating noise interference caused by perceptual uncertainty. Experimental results on the Replica and Matterport3D datasets indicate that our method is particularly effective on unheard tasks, demonstrating improved generalization under unknown sound source distributions.

Diffusion language models (DLMs) expose their denoising trajectories, offering a natural handle for inference-time control; accordingly, an ideal hallucination mitigation framework should intervene during generation using this model-native signal rather than relying on an externally trained hallucination classifier. Toward this, we formulate commitment uncertainty localization: given a denoising trajectory, identify token positions whose cross-chain entropy exceeds an unsupervised threshold before factually unreliable commitments propagate into self-consistent but incorrect outputs. We introduce a suite of trajectory-level assessments, including a cross-chain divergence-at-hallucination (CDH) metric, for principled comparison of localization methods. We also introduce OSCAR, a training-free inference-time framework operationalizing this formulation. OSCAR runs N parallel denoising chains with randomized reveal orders, computes cross-chain Shannon entropy to detect high-uncertainty positions, and then performs targeted remasking conditioned on retrieved evidence. Ablations confirm that localization and correction contribute complementary gains, robust across N in {4, 8, 16}. On TriviaQA, HotpotQA, RAGTruth, and CommonsenseQA using LLaDA-8B and Dream-7B, OSCAR enhances generation quality by significantly reducing hallucinated content and improving factual accuracy through uncertainty-guided remasking, which also facilitates more effective integration of retrieved evidence. Its native entropy-based uncertainty signal surpasses that of specialized trained detectors, highlighting an inherent capacity of diffusion language models to identify factual uncertainty that is not present in the sequential token commitment structure of autoregressive models.

Benchmarks that reflect production workloads are better for evaluating AI coding agents in industrial settings, yet existing benchmarks differ from real usage in programming language distribution, prompt style and codebase structure. This paper presents a methodology for curating production-derived benchmarks, illustrated through ProdCodeBench, a benchmark sourced from real developer-agent sessions. We detail our data collection and curation practices including LLM-based task classification, test relevance validation, and multi-run stability checks which address challenges in constructing reliable evaluation signals from monorepo environments. Each curated sample consists of a verbatim prompt, a committed code change and fail-to-pass tests spanning seven programming languages. Our systematic analysis of four foundation models yields solve rates ranging from 53.2% to 72.2%. We demonstrate how these offline evaluation signals drive practical decisions around model selection and harness design, while noting that offline benchmarks provide directional signal that we complement with online A/B testing for production deployment decisions. We share our methodology and lessons learned to enable other organizations to construct similar production-derived benchmarks.

Artificial intelligence (AI) systems are increasingly integrated into healthcare and pharmacy workflows, supporting tasks such as medication recommendations, dosage determination, and drug interaction detection. While these systems often demonstrate strong performance under standard evaluation metrics, their reliability in real-world decision-making remains insufficiently understood. In high-risk domains such as medication management, even a single incorrect recommendation can result in severe patient harm. This paper examines the reliability of AI-assisted medication systems by focusing on system failures and their potential clinical consequences. Rather than evaluating performance solely through aggregate metrics, this work shifts attention towards how errors occur and what happens when AI systems produce incorrect outputs. Through a series of controlled, simulated scenarios involving drug interactions and dosage decisions, we analyse different types of system failures, including missed interactions, incorrect risk flagging, and inappropriate dosage recommendations. The findings highlight that AI errors in medication-related contexts can lead to adverse drug reactions, ineffective treatment, or delayed care, particularly when systems are used without sufficient human oversight. Furthermore, the paper discusses the risks of over-reliance on AI recommendations and the challenges posed by limited transparency in decision-making processes. This work contributes a reliability-focused perspective on AI evaluation in healthcare, emphasising the importance of understanding failure behavior and real-world impact. It highlights the need to complement traditional performance metrics with risk-aware evaluation approaches, particularly in safety-critical domains such as pharmacy practice.

Recent 3D Gaussian splatting methods built atop SMPL achieve remarkable visual fidelity while continually increasing the complexity of the overall training architecture. We demonstrate that much of this complexity is unnecessary: by replacing SMPL with the Momentum Human Rig (MHR), estimated via SAM-3D-Body, a minimal pipeline with no learned deformations or pose-dependent corrections achieves the highest reported PSNR and competitive or superior LPIPS and SSIM on PeopleSnapshot and ZJU-MoCap. To disentangle pose estimation quality from body model representational capacity, we perform two controlled ablations: translating SAM-3D-Body meshes to SMPL-X, and translating the original dataset's SMPL poses into MHR both retrained under identical conditions. These ablations confirm that body model expressiveness has been a primary bottleneck in avatar reconstruction, with both mesh representational capacity and pose estimation quality contributing meaningfully to the full pipeline's gains.

Citation granularity - whether to cite individual sentences, paragraphs, or documents - is a critical design choice in attributed generation. While fine-grained citations are often preferred for precise human verification, their impact on model performance remains under-explored. We analyze four model scales (8B-120B) and demonstrate that enforcing fine-grained citations degrades attribution quality by 16-276% compared to the best-performing granularity. We observe a consistent performance pattern where attribution quality peaks at intermediate granularities (paragraph-level). Our analysis suggests that fine-grained (sentence-level) citations disrupt necessary semantic dependencies for attributing evidence to answer claims, while excessively coarse citations (multi-paragraph) introduce distracting noise. Importantly, the magnitude of this performance gap varies non-monotonically with model scale: fine-grained constraints disproportionately penalize larger models, suggesting that atomic citation units disrupt the multi-sentence information synthesis at which these models excel. Strikingly, citation-optimal granularity leads to substantial gains in attribution quality while preserving or even improving answer correctness. Overall, our findings demonstrate that optimizing solely for human verification via fine-grained citation disregards model constraints, compromising both attribution faithfulness and generation reliability. Instead, effective attribution requires aligning citation granularity with the model's natural semantic scope.

The scale and complexity of modern cloud infrastructure have made Infrastructure-as-Code (IaC) essential for managing deployments. While large Language models (LLMs) are increasingly being used to generate IaC configurations from natural language, user requests are often underspecified. Unlike traditional code generation, IaC configurations cannot be executed cheaply or iteratively repaired, forcing the LLMs into an almost one-shot regime. We observe that ambiguity in IaC exhibits a tractable compositional structure: configurations decompose into three hierarchical axes (resources, topology, attributes) where higher-level decisions constrain lower-level ones. We propose a training-free, disagreement-driven framework that generates diverse candidate specifications, identifies structural disagreements across these axes, ranks them by informativeness, and produces targeted clarification questions that progressively narrow the configuration space. We introduce \textsc{Ambig-IaC}, a benchmark of 300 validated IaC tasks with ambiguous prompts, and an evaluation framework based on graph edit distance and embedding similarity. Our method outperforms the strongest baseline, achieving relative improvements of +18.4\% and +25.4\% on structure and attribute evaluations, respectively.

Traditional scientific discovery relies on an iterative hypothesise-experiment-refine cycle that has driven progress for centuries, but its intuitive, ad-hoc implementation often wastes resources, yields inefficient designs, and misses critical insights. This tutorial presents Bayesian Optimisation (BO), a principled probability-driven framework that formalises and automates this core scientific cycle. BO uses surrogate models (e.g., Gaussian processes) to model empirical observations as evolving hypotheses, and acquisition functions to guide experiment selection, balancing exploitation of known knowledge and exploration of uncharted domains to eliminate guesswork and manual trial-and-error. We first frame scientific discovery as an optimisation problem, then unpack BO's core components, end-to-end workflows, and real-world efficacy via case studies in catalysis, materials science, organic synthesis, and molecule discovery. We also cover critical technical extensions for scientific applications, including batched experimentation, heteroscedasticity, contextual optimisation, and human-in-the-loop integration. Tailored for a broad audience, this tutorial bridges AI advances in BO with practical natural science applications, offering tiered content to empower cross-disciplinary researchers to design more efficient experiments and accelerate principled scientific discovery.

We consider the question: when a large language reasoning model makes a choice, did it think first and then decide to, or decide first and then think? In this paper, we present evidence that detectable, early-encoded decisions shape chain-of-thought in reasoning models. Specifically, we show that a simple linear probe successfully decodes tool-calling decisions from pre-generation activations with very high confidence, and in some cases, even before a single reasoning token is produced. Activation steering supports this causally: perturbing the decision direction leads to inflated deliberation, and flips behavior in many examples (between 7 - 79% depending on model and benchmark). We also show through behavioral analysis that, when steering changes the decision, the chain-of-thought process often rationalizes the flip rather than resisting it. Together, these results suggest that reasoning models can encode action choices before they begin to deliberate in text.

Search agents, which integrate language models (LMs) with web search, are becoming crucial for answering complex user queries. Constructing training datasets for deep research tasks, involving multi-step retrieval and reasoning, remains challenging due to expensive human annotation, or cumbersome prerequisites. In this work, we introduce ORBIT, a training dataset with 20K reasoning-intensive queries with short verifiable answers, generated using a frugal framework without relying on paid API services. The modular framework relies on four stages: seed creation, question-answer pair generation, and two stages of verification: self and external. ORBIT spans 15 domains and each training pair requires 4-5 reasoning steps, with external search verification required from the complete web. We train Qwen3-4B as the base model on ORBIT using GRPO and evaluate it on Wikipedia question answering tasks. Extensive experiment results demonstrate that ORBIT-4B achieves strong performance among sub-4B LLMs as search agents, proving the utility of synthetic datasets. Our framework, code and datasets are open-sourced and available publicly.

This paper argues that AI-enabled analysis of street-view imagery, complemented by performance-gated machine-learning imputation, provides a viable pathway for generating building-specific elevation data at regional scale for flood risk assessment. We develop and apply a three-stage pipeline across 18 areas of interest (AOIs) in Texas that (1) extracts LFE and the height difference between street grade and the lowest floor (HDSL) from Google Street View imagery using the Elev-Vision framework, (2) imputes missing HDSL values with Random Forest and Gradient Boosting models trained on 16 terrain, hydrologic, geographic, and flood-exposure features, and (3) integrates the resulting elevation dataset with Fathom 1-in-100 year inundation surfaces and USACE depth-damage functions to estimate property-specific interior flood depth and expected loss. Across 12,241 residential structures, street-view imagery was available for 73.4% of parcels and direct LFE/HDSL extraction was successful for 49.0% (5,992 structures). Imputation was retained for 13 AOIs where cross-validated performance was defensible, with selected models achieving R suqre values from 0.159 to 0.974; five AOIs were explicitly excluded from prediction because performance was insufficient. The results show that street-view-based elevation mapping is not universally available for every property, but it is sufficiently scalable to materially improve regional flood-risk characterization by moving beyond hazard exposure to structure-level estimates of interior inundation and expected damage. Scientifically, the study advances LFE estimation from a pilot-scale proof of concept to a regional, end-to-end workflow. Practically, it offers a replicable framework for jurisdictions that lack comprehensive Elevation Certificates but need parcel-level information to support mitigation, planning, and flood-risk management.

The rapid development of agentic artificial intelligence (AI) is driving future wireless networks to evolve from passive data pipes into intelligent collaborative ecosystems under the emerging paradigm of integrated learning and communication (ILAC). However, realizing efficient agentic collaboration faces challenges not only in handling semantic redundancy but also in the lack of an integrated mechanism for communication, computation, and control. To address this, we propose a wireless agent network (WAN) framework that orchestrates a progressive knowledge aggregation mechanism. Specifically, we formulate the aggregation process as a joint energy minimization problem where the agents perform semantic compression to eliminate redundancy, optimize transmission power to deliver semantic payloads, and adjust physical trajectories to proactively enhance channel qualities. To solve this problem, we develop a hierarchical algorithm that integrates inner-level resource optimization with outer-level topology evolution. Theoretically, we reveal that incorporating a potential field into the topology evolution effectively overcomes the short-sightedness of greedy matching, providing a mathematically rigorous heuristic for long-term energy minimization. Simulation results demonstrate that the proposed framework achieves superior energy efficiency and scalability compared to conventional benchmarks, validating the efficacy of semantic-aware collaboration in dynamic environments.

This paper introduces two transfer learning methodologies for estimating nonparametric Bayesian networks under scarce data. We propose two algorithms, a constraint-based structure learning method, called PC-stable-transfer learning (PCS-TL), and a score-based method, called hill climbing transfer learning (HC-TL). We also define particular metrics to tackle the negative transfer problem in each of them, a situation in which transfer learning has a negative impact on the model's performance. Then, for the parameters, we propose a log-linear pooling approach. For the evaluation, we learn kernel density estimation Bayesian networks, a type of nonparametric Bayesian network, and compare their transfer learning performance with the models alone. To do so, we sample data from small, medium and large-sized synthetic networks and datasets from the UCI Machine Learning repository. Then, we add noise and modifications to these datasets to test their ability to avoid negative transfer. To conclude, we perform a Friedman test with a Bergmann-Hommel post-hoc analysis to show statistical proof of the enhanced experimental behavior of our methods. Thus, PCS-TL and HC-TL demonstrate to be reliable algorithms for improving the learning performance of a nonparametric Bayesian network with scarce data, which in real industrial environments implies a reduction in the required time to deploy the network.

Multi-agent Retrieval-Augmented Generation (RAG), wherein each agent takes on a specific role, supports hard queries that require multiple steps and sources, or complex reasoning. Existing approaches, however, rely on static agent behaviors and fixed orchestration strategies, leading to brittle performance on diverse, multi-hop tasks. We identify two key limitations: the lack of continuously adaptive orchestration mechanisms and the absence of behavior-level learning for individual agents. To this end, we propose HERA, a hierarchical framework that jointly evolves multi-agent orchestration and role-specific agent prompts. At the global level, HERA optimizes query-specific agent topologies through reward-guided sampling and experience accumulation. At the local level, Role-Aware Prompt Evolution refines agent behaviors via credit assignment and dual-axes adaptation along operational and behavioral principles, enabling targeted, role-conditioned improvements. On six knowledge-intensive benchmarks, HERA achieves an average improvement of 38.69\% over recent baselines while maintaining robust generalization and token efficiency. Topological analyses reveal emergent self-organization, where sparse exploration yields compact, high-utility multi-agent networks, demonstrating both efficient coordination and robust reasoning.

Spatial consistency is a fundamental property of the visual world and a key requirement for models that aim to understand physical reality. Despite recent advances, multimodal large language models (MLLMs) often struggle to reason about 3D geometry across multiple views. Rather than asking models to describe scene attributes, we introduce a more challenging task: given two views of the same scene, identify the object that violates 3D motion consistency. We propose a simple and scalable method for generating realistic, spatially inconsistent image pairs from multi-view scenes, enabling systematic evaluation of this capability. Our results show that state-of-the-art MLLMs significantly underperform human observers and exhibit substantial variability across different scene attributes, revealing a fragile and incomplete understanding of 3D structure. We hope our findings underscore the need for approaches that develop a more deeply grounded understanding of the physical world.

Forecasting startup success is notoriously difficult, partly because meaningful outcomes, such as exits, large funding rounds, and sustained revenue growth, are rare and can take years to materialize. As a result, signals are sparse and evaluation cycles are slow. Y Combinator batches offer a unique mitigation: each batch comprises around 200 startups, funded simultaneously, with evaluation at Demo Day only three months later. We introduce YC Bench, a live benchmark for forecasting early outperformance within YC batches. Using the YC W26 batch as a case study (196 startups), we measure outperformance with a Pre-Demo Day Score, a KPI combining publicly available traction signals and web visibility. This short-term metric enables rapid evaluation of forecasting models. As a baseline, we take Google mentions prior to the YC W26 application deadline, a simple proxy for prior brand recognition, recovering 6 of 11 top performers at YC Demo Day (55% recall). YC Bench provides a live benchmark for studying startup success forecasting, with iteration cycles measured in months rather than years. Code and Data are available on GitHub: https://github.com/benstaf/ycbench

Adversaries continuously evolve their tactics, techniques, and procedures (TTPs) to achieve their objectives while evading detection, requiring defenders to continually update their understanding of adversary behavior. Prior research has proposed automated extraction of TTP-related intelligence from unstructured text and mapping it to structured knowledge bases, such as MITRE ATT&CK. However, existing work varies widely in extraction objectives, datasets, modeling approaches, and evaluation practices, making it difficult to understand the research landscape. The goal of this study is to aid security researchers in understanding the state of the art in extracting attack tactics, techniques, and procedures (TTPs) from unstructured text by analyzing relevant literature. We systematically analyze 80 peer-reviewed studies across key dimensions: extraction purposes, data sources, dataset construction, modeling approaches, evaluation metrics, and artifact availability. Our analysis reveals several dominant trends. Technique-level classification remains the dominant task formulation, while tactic classification and technique searching are underexplored. The field has progressed from rule-based and traditional machine learning to transformer-based architectures (e.g., BERT, SecureBERT, RoBERTa), with recent studies exploring LLM-based approaches including prompting, retrieval-augmented generation, and fine-tuning, though adoption remains emergent. Despite these advances, important limitations persist: many studies rely on single-label classification, limited evaluation settings, and narrow datasets, constraining cross-domain generalization. Reproducibility is further hindered by proprietary datasets, limited code releases, and restricted corpora.

Belief propagation (BP) provides a scalable heuristic for contracting tensor networks on loopy graphs, but its success in quantum many-body settings has largely rested on empirical evidence. Developing upon a recently introduced cluster-expansion framework for tensor networks, we rigorously study the applicability of BP to many-body quantum systems. For a state represented as a PEPS satisfying a ``loop-decay" condition, we prove that BP supplemented by cluster corrections approximates local observables with exponentially small relative error, and we give explicit formulas expressing local expectation values as BP predictions dressed by connected clusters intersecting the observable region. This representation establishes a direct link between cluster corrections and physical correlation functions. As a result, we show that ``loop-decay" \emph{necessarily implies} exponential decay of connected correlations, yielding sharp, rigorous criteria for when BP can and cannot succeed, and ruling out its validity at critical points. Numerical simulations of the two- and three-dimensional transverse field Ising model at zero and finite temperature confirm our analytical predictions, demonstrating quantitative accuracy deep in gapped phases and systematic failure near criticality.

Chirality-induced spin selectivity (CISS) was recently observed in photo-excited donor-chiral bridge-acceptor molecules, but a predictive theory able to explain available experiments is still lacking. Here we show that low-energy torsional modes modulating hopping and spin-orbit coupling give rise to a Dzyaloshinskii-Moriya interaction between the transferred electron and the one sitting on the donor, producing high spin polarization for perfectly realistic parameters. Our model introduces a low energy scale in the spin dynamics which explains the magnetic field dependence observed in EPR measurements and predicts a non-trivial temperature dependence, as demonstrated by numerical simulations. The present theory lays the foundations for future test-bed experiments and for the design of applications in spintronics and quantum technologies.

Particles whose shapes couple translation to rotation display a rich array of behaviors as they sediment at low Reynolds number. We introduce a unifying perspective in which the possible dynamical regimes and bifurcations between them can be understood. We use experimental measurements of helical ribbons, with controlled center of mass offsets, to identify the key bifurcation from complex dynamics to a single attracting state as the magnitude of the offset increases. The sedimentation dynamics are very sensitive to small center of mass offsets, with the bifurcation occurring for offsets less than one percent of the particle length. Using mobility tensors obtained from immersed boundary method simulations, we simulate helical particle sedimentation and identify an alignment bifurcation surface, defined in the three dimensional space of center of mass offsets, that separates simple from complex sedimentation dynamics. Inside this surface we find limit cycles which emerge through Hopf and homoclinic bifurcations. Cocentered particles with coincident centers of force and mobility provide a reference case at the center of the bifurcation surface. We show how the geometric and dynamical symmetries of sedimenting cocentered particles are broken as the center of force offset moves away from the cocentered case. Three parity time-reversal (PT) symmetries exist for all cocentered particles under reflections normal to the eigenvectors of its translation-rotation coupling tensor. When a center of force offset preserves at least one of these PT symmetries, then there are closed orbits for particles inside the alignment bifurcation surface.

High-order extensions of the Hopfield model are known to exhibit dramatically enhanced storage capacity at equilibrium, while their dynamical retrieval properties remain less well understood. In our previous work, we carried out a dynamical mean-field theory (DMFT) analysis of the Krotov--Hopfield-type dense associative memory and found that the transition between successful and failed retrieval is accompanied by pronounced slow dynamics. As a consequence, the effective basin of attraction observed in numerical simulations extends well beyond that predicted by equilibrium statistical mechanics. A natural hypothesis is that this discrepancy originates from diagonal (self-interaction) contributions in the Krotov--Hopfield model, which generate a large number of lower-order interaction terms and may induce glassy relaxation near the retrieval boundary. To test this hypothesis, we analyze an alternative high-order associative memory model, namely the Abbott--Arian-type $p$-body Hopfield model, in which such diagonal contributions are absent by construction. Using dynamical mean-field theory, we derive an effective single-site process together with closed macroscopic equations governing the retrieval dynamics. Our analysis reveals that both slow dynamics and a substantial enlargement of the apparent basin of attraction persist even in this model. These results indicate that the dynamical slowdown near the retrieval boundary cannot be attributed primarily to diagonal self-interaction effects, but instead originates from intrinsic properties of high-order interactions.

We analyze the one-pass stochastic gradient descent dynamics of a two-layer neural network with quadratic activations in a teacher--student framework. In the high-dimensional regime, where the input dimension $N$ and the number of samples $M$ diverge at fixed ratio $α= M/N$, and for finite hidden widths $(p,p^*)$ of the student and teacher, respectively, we study the low-dimensional ordinary differential equations that govern the evolution of the student--teacher and student--student overlap matrices. We show that overparameterization ($p>p^*$) only modestly accelerates escape from a plateau of poor generalization by modifying the prefactor of the exponential decay of the loss. We then examine how unconstrained weight norms introduce a continuous rotational symmetry that results in a nontrivial manifold of zero-loss solutions for $p>1$. From this manifold the dynamics consistently selects the closest solution to the random initialization, as enforced by a conserved quantity in the ODEs governing the evolution of the overlaps. Finally, a Hessian analysis of the population-loss landscape confirms that the plateau and the solution manifold correspond to saddles with at least one negative eigenvalue and to marginal minima in the population-loss geometry, respectively.

Electron vortices are the quintessential signature of a viscous electron fluid. For decades, their detection relied on indirect transport measurements with persistently debated interpretations. Recently, scanning magnetometry enabled direct visualization, yet these techniques demand considerable sophistication. Here we introduce a conceptually different and inherently simpler paradigm based on nanomechanics. By integrating a circular cavity into a suspended resonator, we create a vortex whose circulating current generates a magnetic moment. In an in-plane magnetic field, this moment experiences a torque, driving vibrations that directly reveal the vortex's presence and nature. We detect ballistic and hydrodynamic vortices and trace their temperature-driven crossover. Our work establishes nanomechanics as a platform for electron hydrodynamics, showing that viscosity - subtle in transport - is one of the dominant factors shaping nanoelectromechanical response.

Two-dimensional noncollinear magnetic states, such as spin-spiral magnets, offer an excellent platform for investigating fundamental phenomena, with potential for advancing stray-field-free spintronics. However, detection and characterization of noncollinear magnetic states in two-dimensional systems remain challenging, motivating the development of alternative probing methods. Here, we present a methodology for extracting the spin-spiral $\mathbf{q}$ vector from lateral electronic transport measurements. Our approach leverages the magnetic field and bias dependence of the conductance to train a supervised machine learning algorithm, which enables us to extract the $\mathbf{q}$ vectors of arbitrary spin-spiral magnets. We demonstrate that this methodology is robust to the presence of impurities in the system and noise in the conductance data. Our findings show that the conductance pattern reveals a complex dependence on the $\mathbf{q}$ vector of the spin spiral, providing a new strategy to learn magnetic structures directly from transport experiments.

Although resetting has widespread applicability, applying it to the dynamics in the presence of spatial quenched disorder, which is essential in many physical problems, is challenging. In this study, we consider a well-known one-dimensional model of particle hopping on a lattice with quenched disorder in the form of site-dependent hopping probabilities, drawn from a power-law distribution, and apply the resetting formalism. As a physical example, we recast the growth dynamics of microtubules with sudden catastrophic disassembly events as a resetting dynamics. We consider two distinct regimes for growth dynamics: a strongly biased case and a less biased case. Motivated by experimental results, we take a Gamma distribution for the resetting time. Our results show that occasional disassembly events are crucial for the experimentally observed distribution of reset (or catastrophe) lengths. We also analyze steady-state distributions under different resetting protocols-resetting to the initial position versus a random site. We also investigate the distribution of first-passage times to a fixed distance following reset. Finally, by considering other resetting probability distributions, we identify a regime where the mean displacement grows as slowly as $\log^2 t$. We also elucidate the role of disorder in the system properties under the resetting dynamics. Our study paves the way to treat the dynamics of complex physical systems using resetting.

The fluctuation-dissipation theorem is a hallmark of equilibrium system that stem from their time-reversal symmetry. In many non-equilibrium systems, in particular active ones, extensions and explicit violations of this theorem are used to assess their ''distance'' to equilibrium. In Hamiltonian flocks, conservative yet non-Galilean models of polar liquids, previous work reported collective motion without the activity that usually underlies it. In this paper, we show that this model obeys a generalized time-reversal symmetry that yields a fluctuation-dissipation theorem that mixes position and polarity degrees of freedom. Due to the oddness of spin under time reversal, the system also obeys Onsager-Casimir reciprocity rather than standard Onsager relations. The coupling also induces rich spin orientation dynamics, including a non-trivial diffusion constant at long times. Finally, we show that considering the naïve time-reversal operation rather than the generalized one that leaves the system invariant leads to a spurious entropy production rate, that could be wrongly interpreted as a distance to equilibrium. Our findings suggest looking for possible extensions of time-reversal symmetry in active-looking systems, which may lead to yet unknown generalizations of the fluctuation-dissipation theorem.

Particle number fluctuations $N(t)$, measured in virtual observation boxes of an image or a simulation, offer a way to quantify particle dynamics when particle tracking is impractical, such as in high-density systems. While traditionally limited to equilibrium diffusive systems, we extend this approach -- named ``Countoscope'' -- to out-of-equilibrium self-propelled particles: Active Brownian (ABPs), Run and Tumble (RTPs), and Active Ornstein-Uhlenbeck Particles (AOUPs). For AOUPs, we leverage their Gaussian statistics to derive a general formula applicable to any Gaussian system. For ABPs and RTPs, we derive the intermediate scattering function (ISF) -- and thus the correlations of $N(t)$ -- using an exact perturbative expansion over the probability density fields, revealing key physical features of the ISF and of the number correlations. Our theoretical predictions for the mean-squared number difference $\langle ΔN^2(t) \rangle = \langle (N(t) - N(0))^2 \rangle$ match stochastic simulations and exhibit three time-dependent scaling regimes: diffusive, advective, and long-time enhanced diffusive, reflecting the regimes of the mean squared particle displacement. We further uncover limiting laws in each of these regimes that are useful to quantify self-propulsion properties.

We investigate the time- and momentum-resolved dynamics of matter waves undergoing elastic scattering from a disordered potential in the presence of spatially uniform SU(2) gauge fields. We derive the disorder-averaged density matrix as a function of time and momentum within the weak-localization regime. By accurately approximating the frequency dependence of the ladder and maximally crossed diagram series beyond the diffusive approximation, we describe short-time spin-momentum dynamics on timescales comparable to the scattering mean free time, for arbitrary strengths of the SU(2) gauge fields and disorder. We also present a cubic equation that determines the spin isotropization time, which gives accurate asymptotic forms in the limits where the spin-orbit length is much longer (Dyakonov-Perel spin relaxation regime) or much shorter than the scattering mean free path, as well as in the SU(2)-symmetric (persistent spin helix) limit. In comparison with numerical calculations, we reproduce both the relaxation of the momentum distribution and the transient backscattering peak with a momentum offset coexisting with the robust coherent backscattering dip, supporting the reliability of our calculations.

In nonequilibrium suspensions, static number fluctuations $N$ in virtual observation boxes reveal remarkable structural properties, but the dynamic potential of $N(t)$ signals remains unexplored. Here, we develop a theory to learn the dynamical parameters of self-propelled particle models from $N(t)$ statistics. Unlike traditional trajectory analysis, $N(t)$ statistics distinguish between models, by sensing subtle differences in reorientation dynamics that govern re-entrance events in boxes. This paves the way for quantifying advanced dynamic features in dense nonequilibrium suspensions.

In studies of turbulence, there has been extensive use of physical quantities such as {\it energy transfers} and {\it structure functions}. We examine whether these quantities can be useful in understanding problems of domain growth or coarsening, as modeled by the {\it time-dependent Ginzburg-Landau} (TDGL) equation and the {\it Cahn-Hilliard} (CH) equation. This paper has two major themes. First, we review our recent papers on energy transfers in domain growth. Second, we study structure functions and intermittency for coarsening systems. As a consequence of sharp interfaces, the structure functions scale as $S_q \sim r^{ζ_q}$, where $r$ is the distance between two points. For the TDGL and CH models, $ζ_q = 1$, indicating {\it anomalous scaling}

Along the bacterial chromosome, regions called rrn operons contain genes that are transcribed into ribosomal RNA. These operons are among the most transcriptionally active sites in the genome. It has been observed in E. coli that RNA polymerase (RNAP), while binding to these genetic loci along the chromosome during transcription, forms dense clusters, leading to spatial colocalization of the operons within the cell. Recent experimental evidence suggests that liquid-liquid phase separation contributes to the formation of RNAP clusters, with the antitermination factor NusA playing a key role. We present a simulation model to investigate the mechanisms underlying the formation of these biomolecular condensates. We propose that mutual attraction between NusA proteins, which exhibit a miscibility gap at higher concentrations, drives condensate formation via a polymer-assisted condensation pathway, and we demonstrate how these condensates promote the colocalization of rrn operons. Our results reconcile seemingly disparate experimental observations of chromosomal organization reported in fluorescence-based imaging and Hi-C experiments.

Ferroelectric nematic phases are a new class of polar fluids in which spontaneous polarization is directly coupled to the orientational order, providing unique opportunities for creating self-organized materils with spatially modulated electric polarization and nonlinear optical response. Here we report the spontaneous emergence of polarization modulated textures in a chiral ferroelectric nematic material close to the transition to the chiral twist-bend ferroelectric nematic phase. By systematically varying cell thickness and surface anchoring conditions, we map the formation of these modulated states, revealing stripe, square and hexagonal morphologies determined via confinement conditions. These structures are directly translated into periodic modulation of the nonlinear optical response, as evidenced by second-harmonic generation imaging. Comparison with an elasticity based theoretical framework and numerical free energy minimization shows that the instability originates from the softening of the bend elastic constant in the chiral nematic phase as the system approaches the lower-temperature heliconical polar phase. The resulting elastic frustration, combined with confinement, drives the formation of spatially periodic director distortions, highlighting ferroelectric nematic fluids as a promising platform for self-assembled nonlinear optical materials.

We investigate dense higher-order associative memories in the high storage regime when the stored patterns are biased, namely when the entries of the patterns are not symmetrically distributed around zero. In this setting, the standard Hebbian prescription must be modified by recentering and rescaling the pattern entries, and an additional term must be introduced in the Hamiltonian to enforce consistency between the average activity of the network and that of the stored patterns. As a first step, we perform a signal-to-noise analysis in the zero-temperature limit and show that the bias reduces the effective storage capacity through a multiplicative correction factor (1-b^2)^P, while preserving the superlinear scaling with the system size. We then derive the quenched statistical pressure within the Replica Symmetric framework by means of Guerra's interpolation method and obtain the corresponding self consistency equations for the relevant order parameters. The analytical treatment confirms the heuristic prediction of the signal-to-noise argument, showing that the same bias dependent renormalization naturally emerges in the variance of the cross-talk noise. Finally, we discuss the resulting phase behavior of the model and its implications for retrieval performance in the model.

We study the zero-free regions of the partition function of the hard-core model on finite graphs and their implications for the analyticity of the free energy on infinite lattices. Classically, zero-freeness results have been established up to the tree uniqueness threshold $λ_c(Δ-1)$ determined by the maximum degree $Δ$. However, for many graph classes, such as regular lattices, the connective constant $σ$ provides a more precise measure of structural complexity than the maximum degree. While recent approximation algorithms based on correlation decay and Markov chain Monte Carlo have successfully exploited the connective constant to improve the threshold to $λ_c(σ)$, analogous results for complex zero-freeness have been lacking. In this paper, we bridge this gap by introducing a proper definition of the connective constant for finite graphs based on a lower bound on the number of $k$-depth self-avoiding walks. We prove that for any graph family with a lower connective constant $μ$, the partition function is zero-free in a complex neighborhood of the interval $[0, λ]$ for all $λ< λ_c(μ)$. As a direct consequence, we establish the uniqueness and analyticity of the free energy density for infinite lattices up to the connective constant threshold, extending the known regions derived from maximum degree bounds. Our proof utilizes a block contraction technique that lifts the correlation decay property from a real interval to a strip-like complex neighborhood.

One-dimensional topological superconductors accommodate a pair of Majorana zero modes at their ends. In an interferometer containing such a topological superconductor, electron transport is significantly affected by the Majorana zero modes constituting a non-local state localized near both ends of the superconductor. When the number of electrons $\mathcal{N}$ in the superconductor is constrained by a charging effect, the resonant tunneling through the non-local state is expected to result in unusual transport properties. This resonant tunneling, called electron teleportation, is not easy to describe because there is no simple method to handle the constraint on $\mathcal{N}$. Here, we propose a boundary potential method based on scattering theory for calculating the conductance of the interferometer under a given constraint on $\mathcal{N}$. This method enables us to calculate the conductance taking account of relevant charging energy and details of the system.

A central challenge in the theory of open quantum systems is the development of theoretical frameworks capable of accurately describing large, strongly interacting quantum many-body systems in the regime of strong system-bath interaction. In this work, we take a step toward this goal by formulating a polaron-transformed version of the canonically consistent quantum master equation (CCQME) [T. Becker et al., Phys. Rev. Lett. 129, 200403 (2022)]. The CCQME extends beyond standard weak-coupling approaches while retaining the same numerical complexity as conventional quantum master equations, thereby enabling the treatment of large quantum systems. The polaron transformation further enhances the accessible system-bath interaction strengths, allowing us to move from moderate to strong interaction regimes. We present a unified and transparent derivation of these two approaches and combine them to obtain the polaron-transformed CCQME (PT-CCQME). Applying our method to the paradigmatic spin-boson model, we find excellent agreement with numerically exact time-evolving matrix product operator (TEMPO) simulations. Finally, we predict an initial-state-independent slowing down of thermalization in the strong-coupling regime of the spin-boson model.

Topological phases support edge states that can be robust to material deformations and other perturbations. While well-studied in quantum systems, topological phases have also been observed in stochastic and biochemical systems, yet it remains unclear which of their properties remain similar or different from those in quantum systems. In this paper, we derive analytical expressions for the spectral properties of simple quantum and stochastic models on the same lattice to rigorously characterize these complex systems. Intriguingly, we find that non-reciprocity moves states away from the steady-state in stochastic systems while clustering states at zero-energy in quantum systems. In contrast, making the system more topological does the opposite: it clusters more states around the steady-state in stochastic systems but moves states away from the zero-energy state in quantum systems. These results provide control parameters for selection and modulation of different purposes while quantifying the size of gap which protects the longest-lived states. Lastly, we discover a mode unique to stochastic systems that we dub the topologically emerging state, which persists across different models and dimensions, including in the presence of non-equilibrium currents.

Despite the long history of the theory of Bose-Einstein condensation, there exist till nowadays some slippery points that are often misunderstood and result in confusion. The report touches some of these points, explaining the following: Global gauge symmetry breaking is the necessary and sufficient condition for the existence of Bose-Einstein condensate. There is no any ``grand canonical catastrophe". The stability of the ideal Bose gas depends on the spatial dimensionality and the shape of a trap. Symmetry-broken averages cannot be neglected. The so-called ``Popov approximation", ascribed to Popov, suggesting to neglect anomalous averages, is neither an approximation nor has anything to do with Popov. There are no thermodynamically anomalous fluctuations in stable equilibrium systems. Representative statistical ensembles are equivalent.

We predict thermoelectric and current rectification effects in hybrid junctions formed by Ising superconductors and materials hosting valley-polarized states. Both effects originate from the interplay of intrinsic Ising spin-orbit coupling, spin-splitting from an exchange or Zeeman field, and valley polarization. The resulting transport signatures provide experimentally accessible probes of valley-polarized states in van der Waals heterostructures, such as junctions of few-layer transition metal dichalcogenides and twisted bilayer or rhombohedral graphene.

We develop a rigorous system-agnostic method to predict quantum thermalization in an overwhelming fraction of accessible pure states in a many-body system, entirely in terms of certain out-of-time-ordered correlators of few-body observables. In contrast to previous rigorous results on thermalization with semiclassical counterparts, our method is not limited to statistical averages of observables, such as time averages in ergodicity or state averages in mixing. Moreover, consistent with such approaches, we retain the advantage of not requiring a detailed knowledge of energy eigenstate structure or thermodynamically large times, which can become intractable for systems with more than a handful of particles. Our approach is centered on a geometric result that connects thermalization to the alignment of high dimensional subspaces in a Hilbert space, which is determined by the saturation of "controllably nonlocal" out-of-time-ordered correlators. This formalism reduces the problem of establishing pure state quantum thermalization at finite times in almost all complex many-body states to a theoretically or experimentally accessible study of few-body correlators, even in thermodynamically large systems.

Advances in intensity-based microscopy techniques have improved our ability to quantify particle motion at microscopic scales, enabling insight into diffusion and collective dynamics. Building on this foundation, we introduce a novel real-space approach that analyses intensity fluctuations within virtual observation boxes of variable size on microscopy images. By correlating these signals, we uncover distinct temporal regimes in the mean square changes of intensity, $\langle ΔI^2(t) \rangle$, which are strongly dependent on the box size compared to the particle width. For small boxes or short timescales, $\langle ΔI^2(t) \rangle$ scales with the mean-square displacement, while for longer timescales and larger boxes, it scales with its square root. We develop a general theoretical framework that captures these regimes and explicitly apply it to a dilute colloidal suspension imaged with confocal microscopy as an experimental model system. This allows for a robust extraction of diffusion coefficients and physical insights into particle dynamics. Our method complements intensity-based and real-space analysis, offering a tool for studying individual and potentially collective behaviour directly from image intensities, even in systems where individual particles cannot be resolved.

We demonstrate that wave amplification enables even weak nonlinearities to reshape linear wave-packet transport in nonreciprocal systems. We study the dynamics of bulk Gaussian wave packets in the Hatano--Nelson model with onsite cubic nonlinearity. We show that the interplay between nonlinearity and amplification generates growing frequency shifts that drive the wave packet through three successive dynamical regimes: an early nonlinear-skin regime with coherent propagation, an intermediate wave-mixing regime driven by mode resonances, and a self-trapping regime in which part of the packet localizes while the remainder ballistically spreads along the system favored direction. The crossover time scales are set by the width and average spacing of the eigen-frequency spectrum. Crucially, within the nonlinear-skin regime, we derive analytical predictions for the wave-packet dynamics and show that nonlinearity couples amplification, dispersion, and nonreciprocity, thereby modifying the magnitude of the wave-packet acceleration and introducing an explicit time dependence into its evolution. Focusing nonlinearities suppress the acceleration and cause it to decrease in time, whereas defocusing nonlinearities enhance it and cause it to increase. We further show that nonlinear interactions typically break down the wave packet before the non-Hermitian jump can occur. Our results provide a route toward accurate control of waves in nonreciprocal metamaterials.

We study the thermodynamics of the (2+1) dimensional Gross-Neveu model inspired from graphene. We focus on the entropy density of the Gaussian fluctuation beyond the mean field. The full in-medium, momentum-dependent evaluation reveals that the fluctuations give a substantial contribution, even comparable to that of the mean field. We argue that the back-reaction from the fluctuations to the mean field should be included, which reduces the contribution mainly coming from the Landau-damping region. To treat this self-consistently, we use the generalized version of the Beth-Uhlenbeck approach for the entropy density. Compared with the standard Beth-Uhlenbeck formulation, the generalized version suppresses the low-energy contributions while preserving the bound-state effects. The fractional entropy carried by bound excitons and free fermions reveals a sharper crossover of the degrees of freedom in the generalized version, which is consistent with Mott-transition physics in two-dimensional materials.

Some notable systems, such as room-temperature superconductors and materials for controlled nuclear fusion, require an accurate description of finite-temperature quantum matter. Stochastic path integral methods are finite-temperature and numerically exact, but scale poorly with system size due the notorious Fermion sign problem. To somewhat mitigate this, we use a hydrodynamical backflow coordinate transformation. Our first attempt was a continuous normalizing flow machine learning approach to determine the optimal parameters. We found this to reduce the error of the total energy, approximately, three times at medium sign severity. Numerical issues challenged training effectively. Thus, a semi-analytic approach was developed to estimate the optimal parameters. We do this by using a derived expression dependent on a Bosonic observable. Hence, the calculation of these values does not have a sign problem. The resulting backflow transformations reduce the problem by multiple orders of magnitude, specifically, in the case of a harmonically trapped, two-dimensional electron gas at finite-temperature. The total energy of the system agrees with previous, backflow untransformed, studies and we calculate energies for up to 32 electrons. The limiting factor is found to be, primarily, the $O(N^3)$ calculation of the Jacobian, stemming from the coordinate transformation of the backflow. A more thorough implementation may further improve this scaling. Otherwise, a pathway for simulating electron systems at currently unreachable regimes is obtained. Finally, as a specific practical use case in energy storage systems, the quantum capacitance for graphene quantum dot materials is calculated.

The observation of a misalignment between the applied heat flux and the measured temperature gradient in insulating solids induced by magnetic field has become a subject of experimental investigation, theoretical speculation, and unsettled controversy. To identify the origin of this phonon thermal Hall effect, we performed a comparative study of longitudinal and transverse heat transport in crystalline (quartz) and vitreous (silica) SiO$_2$ using identical experimental set-ups and thermometers. A finite signal was detected in the crystalline samples and none in the amorphous sample, within our resolution. The cleaner crystal exhibited a larger thermal Hall conductivity than the dirtier one, ruling out disorder as the driver of the effect. On the other hand, the amplitude of the transverse thermal resistivity is almost identical in the two crystalline samples (W$_{\perp}$/B$\approx 10^{-6}$ m.K.W$^{-1}$.T$^{-1}$). We show that in a phonon gas, as in a molecular gas displaying the Senftleben-Beenakker effect, heat is conducted through two channels, and argue that a thermal Hall response is unavoidable whenever these channels differ both in entropy production and in their coupling to the magnetic field. Under such conditions, the conserved energy current and the non-conserved entropy current cease to be parallel. Finally, the magnitude of the transverse thermal resistivity can be accounted for by a surprisingly simple picture. The heat flux induces a tiny drift velocity of the lattice nuclei, the magnetic field exerts a transverse Berry force on this drift, and this force is balanced by an entropic restoring force.

Proteins can combine into functional elements in living cells or self-assemble into unwanted structures in a number of diseases. The resulting aggregates often display filamentous morphologies across a large range of protein shapes and molecular interactions. This has led to the suggestion that filament formation could be a generic outcome of the aggregation of geometrically complex, ill-fitting objects, although such a mechanism has not been demonstrated in three dimensions. To address this problem, we theoretically study the self-assembly of three-dimensional identical, ill-fitting deformable subunits mimicking globular proteins in solution. In our model, self-assembling subunits incur deformations that accumulate as the aggregate size increases and can eventually hamper further assembly. We analytically predict the ground state morphologies of the resulting aggregates as a function of the subunit adhesivity and elasticity by mapping their mechanics onto those of two incompatible, interconnected networks. We find that zero-dimensional clusters, three-dimensional bulks as well as symmetry-broken one-dimensional filaments and two-dimensional layers can all form depending on assembly parameters. Incompressible, moderately adhesive subunits favor filaments. These findings hint at a generic pathway to control self-assembly in three dimensions and suggests that such mechanisms could be investigated in more realistic protein models.

We study Frobenius algebras of operator fields and introduce a novel notion of duality for them. We show that, under the assumption that the operator fields forming the Frobenius algebra are mutual symmetries, the operator fields in the dual Frobenius algebra are also mutual symmetries. This result allows one to construct new infinite-dimensional integrable systems of hydrodynamic type starting from a given one. As the main application, we solve the long-standing Eisenhart--Stäckel problem for any Segre characteristic and in arbitrary dimension: namely, we describe all nondegenerate finite-dimensional integrable systems whose integrals are quadratic in momenta such that the corresponding $(1,1)$-tensors commute as operator fields.

In this paper, we study the asymptotic stability of smooth 1-solitons in the Degasperis-Procesi (DP) equation. Such solutions necessarily exist on a non-zero background, and their spectral and orbital stability has previously been verified by Li, Liu & Wu and by Lafortune & Pelinovsky. Using the complete integrability of the DP equation to establish the strong spectral stability of smooth solitary waves, namely that the origin is the only eigenvalue of the associated linearized operator acting on $L^2(\mathbb{R})$ and that, moreover, in appropriate exponentially weighted spaces the non-zero spectrum for the linearized operator admits a spectral gap away from the imaginary axis. This spectral gap result {is then} upgraded to an exponential decay estimate on the semigroup associated with the linearized operator, establishing a linear asymptotic stability result in exponentially weighted spaces. Finally, we outline analytical challenges with extending our result to the nonlinear level.

Vegetation in semi-arid environments self-organizes into striking spatial patterns -- bands, spots, labyrinths, and gaps -- with characteristic wavelengths on the order of tens to hundreds of meters. Existing reaction-diffusion models postulate nonlinearities and transport laws from qualitative physical reasoning, making it hard to distinguish essential structural features from artifacts of the chosen forms. Here we show how energy-balance and water-conservation principles can constrain the admissible model class before a specific closure is chosen. These constraints motivate a family of semilinear closures; an Euler--Lagrange representative yields a fourth-order vegetation equation coupled to quasi-steady water transport on a one-dimensional hillslope. Linear stability analysis identifies three instability mechanisms: classical water-mediated feedback, energy-balance spatial coupling, and water deflection by vegetation gradients. Their balance depends on terrain geometry. On slopes, the water-mediated coupling dominates and the model reproduces two empirical observations: pattern wavelength increases with aridity, and vegetation bands migrate uphill. On flat terrain, the energy-balance spatial coupling can drive instability independently. Numerical simulations confirm the linear predictions, and exploratory continuation reveals a narrow hysteresis region consistent with subcritical bifurcation.

This study investigates the existence and stability of limit cycles resulting from self-excited oscillations in linear multi-degree-of-freedom systems subjected to discontinuous, state-dependent forcing. Using the method of averaging and slow-flow phase-plane analysis, analytical expressions are derived for the amplitudes and stability boundaries of limit cycles in a two-degree-of-freedom system. The analysis demonstrates that stable limit cycles may exist in all natural modes, with the steady-state response governed by initial conditions in regimes of multistability. A central contribution of this work is the identification and analytical characterization of the stability-axis-flipping (SAF) bifurcation, which serves as the governing mechanism for the exchange of stability between modes. The framework is then systematically extended to systems with higher degrees of freedom, confirming that the SAF bifurcation remains a universal feature, even under varying feedback configurations. The steady-state dynamics, summarized through stability maps and validated by numerical simulations, delineate the existence and stability regions of modal limit cycles as functions of key system parameters. These results provide efficient criteria for guiding optimization studies to mitigate or generate limit cycles at targeted frequencies in flexible mechanical structures.

In this paper, we study the nonlinear dispersive waves including the rarefaction and dispersive shock waves in the discrete modified KdV equation through the numerical simulations of the dispersive Riemann problems. In particular, we propose distinct quasi-continuum models to approximate both the spatial profiles and distinct edge features of these two specific dispersive wave structures. Whitham analysis is performed to construct a closed system of partial differential equations which describe the slowly-varying dynamics of all the relevant parameters associated with the periodic traveling waves of the proposed quasi-continuum models. We then perform reduction on such modulation system to obtain a system of two simple-wave ordinary differential equations which lead to the DSW-fitting method that shall provide useful theoretical insights on different edge characteristics of the dispersive shock waves. Furthermore, we compute analytically the self-similar solutions corresponding to the dispersionless systems of the quasi-continuum models, which can be utilized to approximate the numerically observed rarefaction waves of the discrete mKdV equation. A systematic numerical comparison of these theoretical findings with their associated numerical counterparts finally demonstrate the good performance of the proposed quasi-continuum models in approximating both nonlinear dispersive wave patterns.

The multiplex network paradigm has been instrumental in revealing many unexpected phenomena and dynamical regimes in complex interacting systems. Nevertheless, most of the current research focuses on undirected multiplex structures, whereas real-world systems predominantly involve directed interactions. Here, we present an analytical framework for attaining optimal synchronization in directed multiplex networks composed of phase oscillators, considering both frustrated and non-frustrated regimes. A multiplex synchrony alignment function (MSAF) is introduced for this purpose, whose formulation integrates structural properties and dynamical characteristics of the individual directed layers. Using this function, we derive two classes of frequency distributions: one that yields perfect synchronization at a prescribed coupling strength in the presence of phase-lag, and another that optimizes synchronization over a broad range of coupling strengths. Numerical simulations on various directed duplex topologies demonstrate that both frequency sets substantially outperform conventional distributions. We also explore network optimization through a directed link rewiring strategy aimed at minimizing the MSAF, along with a swapping algorithm for optimally assigning fixed frequencies on both layers of a given directed duplex network. Examination of synchrony-optimized directed networks uncovers three notable correlations: a positive relationship between frequency and out-degree, a negative correlation between neighboring frequencies, and an anti-correlation between mirror node frequencies across directed layers.

In the present work we revisit the problem of the dark solitary wave pinned in the discrete nonlinear Schr{ö}dinger equation. In a number of recent studies, the methodology of exponential asymptotics was attempted to be utilized in this problem, however the results were not found to be fully in agreement with associated multiprecision numerical computations. Here we resolve this conundrum by finding precise exponential asymptotics for the pinned dark solitary waves. Moreover, we reconcile the relevant result with a general theory of pinned dark solitary waves in the {\it continuum} nonlinear Schr{ö}dinger equations in the presence of external potentials.

We present $\frac{m^{2}}{4}+\frac{m}{2}+\frac{1-\left(-1\right)^{m}}{8}$ homogeneous $(3m-2)$-parameter families of Liouville integrable $(2m)$- and $(2m-1)$-dimensional Lotka-Volterra systems. We also study inhomogeneous versions of these systems.

In heterogeneous networks of coupled oscillators, phase frustration typically prevents the emergence of complete synchronization in the Sakaguchi-Kuramoto (SK) model. In this study, we propose an analytical framework to overcome this barrier and induce complete synchronization in oscillators governed by phase-frustrated bi-harmonic coupling. We derive a general set of natural frequencies correlated with the network's degree heterogeneity, along with the parameters involved in the bi-harmonic coupling function that lead to complete synchronization $(r=1)$ in the presence of the harmonic coupling terms $(K_1, K_2 \neq 0)$. On top of that, we found hysteresis in the synchronization transition in the case of scale-free networks, indicating a first-order (discontinuous) phase transition, whereas Erdos--Renyi networks exhibit a second-order (continuous) synchronization transition. Furthermore, we use mean-field approximation to determine the critical coupling strength for the synchronization transition in the absence of first-harmonic coupling $(K_1=0)$. Here, the obtained optimal natural frequencies scale linearly with the node degree, and the critical coupling strength for the onset of synchronization is derived analytically from the self-consistent equations. In this specific regime, we observe distinct dynamical disparities: the second harmonic drives an explosive first order (or second order) transition for the second order parameter $(r_2)$, while the first order parameter $(r_1)$ remains suppressed in the forward direction but emerges during the backward transition. These findings remain robust with higher-order harmonic coupling schemes, as well as across a diverse range of synthetic and empirical networks, including scale--free, Erdos--Renyi, Zachary Karate Club and C.elegans neural network, demonstrating their general applicability.

We introduce a continuous one-parameter family of elliptic sine-Gordon equations (SGE) characterized by the modulus $0 \le m \le 1$ of Jacobi elliptic functions and analyze some of its properties and obtain its kink solution for various values of modulus $m$. These elliptic SGE have the novel property that while in the limit $m = 0$ they go over to the integrable sine-Gordon equation, in the $m = 1$ limit they go over to the integrable sine hyperbolic-Gordon equations (SHGE).

We study the phase-space organization of the planar elastic pendulum as a function of its two dimensionless control parameters: the reduced energy $R$ and the squared frequency ratio $μ$. By randomly sampling the isoenergetic volume to classify trajectories as oscillatory, rotational, or chaotic across the $(μ, R)$ parameter plane, we obtain a global portrait of the coexistence and competition between dynamical regimes. The chaotic fraction is not uniformly distributed across the parameter plane but concentrates in a well-defined central cloud whose ridge follows a linear relation in the $(μ, R)$ plane and whose maximum does not exceed $70\%$ of the available phase space. The order-chaos-order transition is not a global property of the parameter plane but occurs specifically in the central region surrounding this cloud: along paths that traverse it, oscillatory orbits progressively give way to chaotic trajectories, which in turn yield to rotational orbits as the energy grows, revealing a clear sequential mechanism underlying the transition. The onset of rotational motion is gradual rather than sharp, reflecting a strong dependence on initial conditions. By decomposing the total energy into spring-like, pendulum-like, and coupling contributions, we establish a direct correspondence between the coupling power and the abundance of chaotic trajectories, showing that enhanced inter-mode energy exchange is a reliable indicator of dynamical complexity. These results provide a comprehensive and quantitative map of the dynamical regimes of the elastic pendulum, clarifying the structure of the chaotic cloud and connecting it to the underlying mode-coupling mechanisms.

We study the emergence of periodic oscillations through a Hopf bifurcation in a scalar diffusion equation on the half line coupled to a dynamic boundary condition. Our results quantify the effect of delay through the buffering in the diffusive field on boundary kinetics, drawing a parallel to the emergence of oscillations in delay equations. Technically, the Hopf bifurcation occurs in the presence of essential spectrum induced by the diffusive field, preventing a simple approach via center-manifold reduction. The results are motivated by observations in biological systems where dynamic boundary conditions arise when modeling surface dynamics coupled to bulk diffusion.

This work establishes a large deviation principle for the spectral measure of the Lax matrix associated to the periodic Toda chain of $N$ particles, subject to a generalised Gibbs measure. This large deviation principle is governed by a rate function which can be regarded as a generalisation of the free energy of the system. Such a large deviation principle is proven both for the model when the momentum is constrained to be zero and when it is allowed to fluctuate. Moreover, the large deviation principle is proven directly on the level of the representation of the generalised Gibbs partition function given in terms of the variables realising the classical separation of variables, \textit{i.e.} rectify the equations of motion. As such, this work paves the way towards the computation of the thermodynamic limit of dynamical correlation functions in the Toda chain subject to generalised Gibbs ensemble statistics.

Hamiltonian systems lie at the heart of modeling the physical world. Their defining scalar, the Hamiltonian, encodes both energy conservation and symplectic geometry in its phase-space trajectories. Recent deep learning approaches model Hamiltonian systems by embedding their properties either in the architecture or in the loss function. However, they typically ignore that: i) a Hamiltonian carries units of energy and/or ii) that every integrable Hamiltonian admits a canonical transformation to action-angle coordinates in which the dynamics reduce to a simple rotation on an invariant torus. We propose BuSyNet, a deep learning architecture that combines these two constraints via a dimensionally-consistent, symplectic transformation. A symplectic layer maps input trajectories to lower-dimensional latent action-angle variables, which are then combined with system parameters to discover a symbolic Hamiltonian expression in units of energy. Evaluated on the harmonic oscillator and the Kepler two-body problem (in 2D and 3D), BuSyNet recovers concise, closed-form Hamiltonians that outperform state-of-the-art neural architectures in long-term prediction accuracy and stability, while maintaining interpretability.

We present a comprehensive study of optical solitons supported by spiral potentials in media with the cubic-quintic (CQ) nonlinearity. A variety of families of stationary states, including fundamental and high-order (excited) in-phase, out-of-phase, and hybrid-phase ones, are found. The linear stability analysis, corroborated by direct simulations, demonstrates that all upper-branch nonlinear states in potentials with different azimuthal indices are \emph{completely stable}, which rarely occurs in soliton physics. Our findings suggest spiral potentials as an effective means for multistable optical trapping, with potential applications in all-optical data processing.

Synchronization is a ubiquitous phenomenon in nonequilibrium systems. One intriguing example found in every-day life is lifts installed next to each other, that move closely and arrive almost simultaneously during a busy time. However, the basic mechanism behind this lift synchronization is yet to be elucidated. Here, we investigate the effective interaction acting between the lifts quantitatively. Through the analysis on the time-series data obtained by numerically solving a rule-based discrete model of lifts, in which passengers at each floor show up stochastically and call a lift that is expected to arrive first, we find that the effective interaction acting between the lifts consists of not only attraction but also repulsion. By changing the parameters of the rule-based model, we are successful to tune the ratio of these competing interactions and to control the dynamics of lifts, realising the transition between in-phase and anti-phase synchronizations. Our strategy is applicable to the data of real lifts, and thus it is expected to help controlling lift systems. We believe that this study provides a novel approach to design optimal transportation, which is of great importance in improving sustainability of social systems.

Gray swans, plausible but unobserved extreme events, broaden our understanding of the range of hazards beyond those observed during the short observational record. They are useful for dynamical studies, synthetic training data, emergency planning, infrastructure design, and insurance hazard assessment. We propose a method to produce gray swans from the observational record using gradient descent on a loss function with a differentiable weather prediction model. Minimizing the loss corresponds to perturbed initial conditions that produce a measurable outcome at a future time, subject to constraints, such as the size of the initial perturbations. We illustrate the method by altering hurricane Fiona (2022), which tracked northward over the Atlantic Ocean, to produce a gray-swan outcome similar to hurricane Sandy (2012), which made landfall on the East Coast of the United States after a unique westward turn. The Fiona gray-swan solution, involving small perturbations to reanalysis initial conditions, produces an extratropical cyclone with a Sandy-like track, a warm core, and a minimum sea-level pressure more than 20 hPa lower than Sandy. Perturbations to the extratropical state are more important than to the hurricane, leading to interactive strengthening, and merger, of an upper-level trough and the hurricane. Similar gray swans are found for four other Atlantic hurricanes. A major weakness of this work is that the hurricane core is not resolved by the model used for optimization, and the impact of this is unknown. Furthermore, although these solutions present plausible outcomes, they do not inform on their probability of occurrence.

Optical atomic clocks provide exceptionally accurate and precise signals for timekeeping and precision measurements, but they require high-power, free-space laser configurations that limit scalability. We introduce and explore a scalable infrastructure for strontium (Sr) optical-lattice clocks that incorporates co-design of atomic-beam slowing and a magneto-optical trap (MOT) from an effusion source, generation of complex, three-dimensional free-space laser configurations with a photonic integrated circuit (PIC) and metasurface (MS) optics, and laser stabilization to a frequency-comb supercontinuum generated with integrated nonlinear photonics. With these elements, we realize MOTs of all stable strontium isotopes ($^{84}$Sr, $^{86}$Sr, $^{87}$Sr, $^{88}$Sr) with populations commensurate with natural abundances, demonstrating precise beam control and robustness. Access to laser-cooled alkaline-earth atoms with scalable integrated photonics enables system engineering for optical clocks, quantum sensing, and quantum information, and our experiments demonstrate extensible technologies that advance toward a Sr optical clock largely free of bulk optics.

Accurate assessment of central hemodynamics is essential for diagnosis and risk stratification, yet it still relies largely on invasive measurements or on indirect reconstructions built from population-averaged transfer functions. While conventional methods are valuable in clinical practice, they face limitations, particularly in personalized medicine. Physics-informed methods address these by integrating physical principles, reducing the need for extensive data. In this work, a fully noninvasive, patient-specific framework is developed that combines a validated 1-D model of the systemic arterial tree with physics-informed neural networks (PINNs). This model performs the inverse solution of the flow and pressure fields within the arterial network, given minimal noninvasive measurements of pressure from a cuff reading and trains in 4000 iterations, at least 10x faster than the current state-of-the-art models due to several model enhancements. We validate the model predictions against our 1-D solver, yielding a near perfect correlation, and perform additional tests on a clinical dataset for the identification of important central hemodynamic parameters of cardiac output $CO$ and central systolic blood pressure $cSBP$, with correlations of $r=0.847$ and $r=0.951$, respectively. Moreover, the model is able to tune the patient-specific coefficients of the terminal resistance $R_T$ and compliance $C_T$ while training, treating them as learnable parameters. The inverse PINN model is able to solve the entire tree of 8 arteries with a single network, costing 5-10 minutes of computational time. This significant performance boost compared to traditional iterative inverse methods holds promise towards applications of personalized cardiac output monitoring and hemodynamic assessment via noninvasive approaches like wearable devices.

The 2-μm waveband, emerging as a highly promising candidate for optical communication, offers an extended wavelength window for high-speed optical transmission. Despite its potential, the development of integrated electro-optic (E/O) modulators operating at this wavelength range has been limited. Such E/O modulators are crucial for high-speed optical communication systems at the 2-μm waveband. In this work, we propose and experimentally demonstrate high-performance E/O absorption modulators based on a graphene-silicon slot waveguide. Our approach enables wideband, high-speed, efficient, robust and compact modulators at both 2-μm and 1.5-μm wavebands. This work represents a significant advancement towards the realization of high-speed integrated E/O modulators for optical communication systems operating at the 2-μm wavelength range.

The propagation of the transverse electric (TE) and transverse magnetic (TM) waves in an effectively two-dimensional (2D) isotropic medium is described by Bergmann's equation of acoustics. We develop a dynamical formulation of the stationary scattering of these waves and explore its application in the study of the low-frequency behavior of the scattering data. Specifically, we introduce a suitable notion of fundamental transfer matrix for TE and TM waves in 2D. This is an integral operator $\widehat{\mathbf{M}}$ that carries the information about the scattering properties of the medium and admits a Dyson series expansion involving a non-Hermitian Hamiltonian operator. For situations where the inhomogeneities of the medium are confined to a layer of thickness $\ell$, we use the Dyson series for $\widehat{\mathbf{M}}$ to construct the series expansion of the scattering amplitude in powers of $k\ell$, where $k$ is the incident wavenumber. We derive analytic expressions for the leading- and next-to-leading-order terms of this series, verify the effectiveness of their application to a class of exactly solvable models, and use them to study low-frequency invisibility. In particular, we develop a low-frequency cloaking scheme which is applicable for both TE and TM waves. Our results have immediate applications in the study of low-frequency scattering of acoustic waves in a 2D fluid as these waves are also described by Bergmann's equation.

The absorption of light by materials is one of the most fundamental processes in optics and condensed-matter physics. Here we investigate whether laser light is absorbed by a crystalline material as an electromagnetic wave or as localized photon energies. We excite the first-order phase transition of vanadium dioxide with laser pulses of sufficient frequency to overcome the band gap but with insufficient pulse energy to overcome the latent heat. According to Maxwell's equations and Bloch theory, no transition should occur, because nowhere in the material is enough energy. Nevertheless, we observe with ultrafast electron diffraction for short times a disordered crystal geometry with nanometer-sized spots of switched material. Their amount matches approximately to the number of photons in the absorbed laser wave. Two optical experiments confirm this phenomenon, and simulations of single absorbed photons reproduce all measurements results. Although laser light and Bloch electrons are extended quantum objects, the energy of the individual photons is localized into nanometer dimensions, enabling local consequences at substantially higher energy than average.

We present a molecular extension of our recently proposed Green's function embedding method, interacting-bath dynamical embedding theory (ibDET), for computing charged excitation energies at the $GW$ and EOM-CCSD levels. Starting from atom-centered impurities, we construct bath representations that capture the frequency-dependent entanglement between the impurity and its environment and can be systematically improved via the construction of cluster-specific natural orbitals. Utilizing a $GW$ or coupled-cluster Green's function solver, the self-energy of the full system is assembled from all embedding problems to obtain the interacting Green's function. We show that ibDET provides accurate spectral properties with much reduced cost for a broad range of systems, including conjugated molecules and nanoclusters. Compared with full-system results, the errors in the predicted ionization potentials and electron affinities are around 0.1 eV or smaller, while each embedding problem includes only a small fraction of the total orbital space. This work provides an efficient and scalable framework for computing spectral properties of molecular systems.

Power grids must modernize to meet climate goals while maintaining reliable and stable operating conditions. Yet progress is hindered by a limited understanding of the stochastic processes underlying grid frequency and phase-angle fluctuations, which are induced by the growing penetration of renewable generation, consumer demand fluctuations, and market trading. This issue is particularly acute in Africa, where grids often face weak investment. Here, we present results from a newly collected, large-scale, high-resolution dataset of grid frequency and phase angles for the United Kingdom and South Africa, comprising close to one billion data points. Using superstatistical modeling, we treat market-driven power fluctuations as a slowly varying parameter driving grid dynamics and incorporate nonlinear frequency control. As a result, we derive an analytical model that reproduces multimodal frequency distributions previously obtained from numerical simulations, as well as heavy-tailed fluctuations and double-exponential frequency autocorrelation decays, all in excellent agreement with experimental measurements. Beyond frequency, we also address the so far largely overlooked problem of characterizing spatial phase-angle fluctuations. By comparing our predictions with measurement data, we demonstrate that a low-dimensional effective grid model accurately fits South African data despite the grid's complexity. We also highlight significant differences between the grids of South Africa and the United Kingdom. Our results clarify how energy markets and control policies shape grid dynamics across countries with contrasting infrastructure maturity.

We embed general boundary value problems for the time-harmonic Maxwell equations into the elliptic boundary value theory. This is achieved by introducing two new scalar functions to the electromagnetic field and imposing additional boundary conditions, after which the problem becomes elliptic. The results are applied to general problems for Maxwell's equations in bounded and unbounded domains, as well as to the transmission problem with inhomogeneities on the right-hand side of the equations and at all boundaries. Relations between the inhomogeneities of the elliptic problem are established that provide a one-to-one correspondence between the solutions of Maxwell's problem and the solutions of the elliptic boundary value problem.

This paper presents an experimental and theoretical study of the nonlinear behavior of imperfect interfaces in multilayer structures using an equivalent vibro-acoustic approach. The multilayer system is modeled through a Zig-Zag formulation, in which interfacial coupling conditions, stress continuity and displacement discontinuity, relate the kinematics of adjacent layers while preserving an independent description of each layer. This framework significantly reduces the number of kinematic unknowns without compromising the model accuracy. An equivalent Kirchhoff-Love plate formulation is then introduced to derive a frequency-dependent bending stiffness representative of the global structural response. Experimental measurements of the transverse displacement field are performed using laser vibrometry and processed via the Corrected Force Analysis Technique (CFAT).The results demonstrate that the dynamic response of a three-layer beam with imperfect interfaces depends on the excitation level. In particular, variations in the equivalent bending stiffness are observed, revealing the nonlinear nature of the interfacial behavior. The proposed methodology is applied to a glass-epoxy-glass multilayer beam under various excitation levels.

We report a continuous-wave Ho:YAG thin-disk laser operating at the hundred-watt power level. In multimode operation, the laser delivers a maximum output power of 230 W, with a slope efficiency of 35.9% and an optical-to-optical conversion efficiency of 35.3%. In single-mode operation, an output power of 152.3 W is achieved, with beam quality factors M2 of 1.08 and 1.06 in the horizontal and vertical directions, respectively.

Quantum dots (QDs) offer significant potential for applications in quantum information and optoelectronic devices; however, conventional time-resolved spectroscopy cannot generally simultaneously extract both long-lived relaxation dynamics and short-lived quantum beats from ensemble measurements. This limitation arises from the inherent trade-off between temporal resolution and total acquisition time. Here, we demonstrate that asynchronous optical sampling based on a fiber-delivered frequency comb enables simultaneous observation of QD dynamics across multiple timescales. By integrating a galvanometric scanner, we achieve spatial mapping over a $1 \times 1$-\si{\milli\meter}$^2$ area at 441 discrete points in 30.1~min, a measurement that would otherwise require more than 12~days. At each location, both quantum beats and relaxation lifetimes are resolved, giving physical insights into QD ensembles that were previously inaccessible and paving the way for rapid feedback in device fabrication.

The design of the coordination environment of magnetic ions is key to achieving properties such as large magnetic anisotropy and slow magnetic relaxation, but a systematic exploration of the relevant chemical space for these compounds is missing. Here, we automatically extract all entries of mononuclear Dy coordination complexes from crystallographic databases and use multireference ab initio methods to compute their magnetic anisotropy. In addition, we generate and simulate magnetic anisotropy for 25k new molecules with the general formula [Dy(H$_2$O)$_5$L$_2$]$^{n-}$ and pentagonal bipyramidal coordination geometry, a motif selected as very promising. While no molecule with record magnetic anisotropy is serendipitously identified in crystallography databases, molecules with crystal field splittings over 1600 cm$^{-1}$ are identified by systematically exploring new organic ligands. This corresponds to a ~100% increase of magnetic anisotropy over the reference compound, ~30% over any known pentagonal bipyramidal Dy complex, and approaching record values of pseudo bi-coordinated Dy ions. This study demonstrates that the fine-tuning of Dy's second coordination sphere by organic ligands design can significantly improve magnetic anisotropy and that automated computational screening is key to accelerating this chemically non-intuitive process.

This Letter studies the optimization of a wireless communications system empowered by a periodically time-modulated reconfigurable intelligent surface, coined time-Floquet RIS (TF-RIS), in the presence of mutual coupling (MC) among the RIS elements. In contrast to a conventional RIS whose elements may be reconfigured between signaling intervals, a TF-RIS periodically modulates its elements within a signaling interval, thereby inducing frequency conversion. Periodic time modulation is particularly attractive for harmonic backscatter communications to avoid self-jamming. Based on time-Floquet multiport network theory, we formulate an MC-aware optimization problem for binary-amplitude-shift-keying (BASK) harmonic backscatter communications with practical 1-bit-programmable TF-RIS elements. We propose a general discrete-optimization algorithm and evaluate its performance based on realistic model parameters. We systematically examine the performance dependence on the time resolution of the periodic modulation and the number of retained harmonics. Benchmarking against an MC-unaware approach reveals the importance of MC awareness for the more challenging optimization problem of simultaneous desired-harmonic-channel-gain maximization and undesired-harmonic-channel-gain minimization.

Nitrogen-vacancy (NV) centers in nanodiamonds are excellent nanoscale sensors for measuring parameters such as temperature, magnetic field, and viscosity in complex fluidic environments, including living cells. However, the rapid motion of nanodiamonds in such dynamic systems imposes a significant challenge for continuous, real-time tracking and sensing measurements. Here, we present a fast single particle tracking (SPT) method featuring a tetrahedral detection geometry for time-efficient parallel fluorescence collection using four avalanche photodiodes (4-APDs), which eliminates the temporal latency of traditional sequential scanning. We demonstrate an improvement of about an order of magnitude in the temporal resolution and the upper limit of measurable diffusion coefficient compared to previously reported nanodiamond tracking methods based on single APD. The SPT is integrated with multi-parameter quantum sensing based on optically detected magnetic resonance (ODMR) of NV centers. The sensitivities of ODMR-based temperature and 3D rotation sensing are evaluated at different diffusion coefficients, which shows no significant degradation within our measurement range. We apply the system for thermorheology measurements in glycerol/water mixtures under thermal ramps. Additionally, we perform simultaneous translation and rotation tracking in live cells, revealing correlated translational and rotational dynamics. This approach advances multi-parameter nanoscale sensing for soft matter and biological applications, paving the way for real-time nanoscale sensing in highly dynamic fluidic environments.

Angle-selective optical devices are of importance to several applications such as photovoltaics, high-sensitivity photodetectors and displays. There are several approaches to realizing angular selectivity, but it remains challenging to obtain isotropic responses over large spectral bandwidths in optically thin structures. We introduce a dispersion engineering approach coupled with topology optimization to design 2D metastructures, leveraging guided-mode resonances (GMRs), that exhibit isotropic angular selectivity over relative bandwidths of approximately 20%. We experimentally demonstrate metastructures with complementary angular selectivities, either scattering light strongly near normal incidence and transmitting efficiently at higher incident angles, or vice versa. A key finding is that these designs enable operation over spectral bandwidths greater than the GMR linewidths would suggest, a result of carefully tailored interactions between the Fabry-Perot background and resonantly scattered light. This work marks a significant step forward for the realization of broadband, angle-selective scattering in readily fabricated structures of subwavelength thickness, and enables new possibilities in sensing, analog information processing, high-efficiency photovoltaics, and displays.

Fusion reactors can permanently remediate mercury by using it as a neutron multiplier: each (n,2n) reaction reduces the neutron number towards ${}^{197}$Hg which quickly decays into stable gold, irreversibly removing it from the environment while generating substantial economic value. Fusion energy is therefore not merely environmentally benign, but anti-polluting through the continuous consumption of an environmental pollutant. The history of nuclear fission demonstrates that environmental concerns can be decisive obstacles to low-carbon power deployment, suggesting that integrated pollution remediation fundamentally improves the policy calculus for fusion energy. We show that at high neutron flux (achievable in muon-catalyzed and inertial confinement fusion), nuclear reactions make all mercury isotopes eligible for gold transmutation, incentivizing mercury recovery and valuing the world mercury extractable stock at ${\sim}\$200$ trillion, exceeding all in-ground gold reserves. Co-producing gold alongside electricity can triple a fusion plant's revenue, aligning economic incentives with complete, permanent mercury remediation.

Maximally-localized Wannier functions are quantum wavefunctions resembling atomic orbitals that are used to describe electrons in condensed matter. Since their introduction in 1997, these functions have become ubiquitous in ab initio materials simulations, including applications in linear-scaling methods, strongly-correlated electron systems, quantum transport, electron-phonon interactions, and topological materials. Despite their widespread adoption in a vast software ecosystem, Wannier functions have not yet attained their fullest potential in the presence of entangled bands, as their optimization remains challenging and labor-intensive. Here, we introduce a universal meta-optimization method that leverages workflow abstraction and machine learning techniques like differential evolution and Bayesian optimization to generate globally optimized Wannier functions without human intervention. We demonstrate this approach through three applications: (i) autonomous interpolation of entangled band structures with millielectronvolt accuracy starting from coarse Brillouin zone grids, (ii) thousand-fold acceleration of fully ab initio Boltzmann transport calculations via the use of minimal coarse Brillouin zone grids, and (iii) ultra-fast high-throughput calculations of high-precision Wannier functions for large materials libraries. This work brings calculations that previously required supercomputers within the reach of personal computers.

Efficient simulation of strongly correlated electrons has become a routine tool in molecular electronic structure theory due to recent advances in approximate configuration interaction (CI) techniques. Nonetheless, the quantitative and predictive description of molecular electronic states remains a significant challenge due to the difficulty of computing all-electron correlation beyond CI. Here, we describe a new open-source implementation of the anti-Hermitian contracted Schrödinger equation (ACSE) for use in accurate simulation of all-electron correlation in molecules. In contrast to standard approaches via multireference perturbation theory, the scaling of the ACSE does not depend on the complexity of the strongly correlated reference wavefunction. Furthermore, the ACSE employs the exact electronic Hamiltonian, rather than an approximate perturbative Hamiltonian. Our benchmark results demonstrate good accuracy for main group and transition metal systems, in weakly and strongly correlated regimes, with various basis sets, and for ground and excited states. The results suggest that the ACSE has potential as a scalable and robust technique for simulating all-electron correlation in molecular ground and excited states.

Coherent control of neutrons via Bragg diffraction forms the foundation of perfect crystal neutron interferometry, facilitating both fundamental tests of quantum mechanics and applications in quantum information science. In cavity geometries, perfect crystals enable neutron confinement and have been employed in precision measurements of spin-orbit interactions and for neutron electric dipole moment (nEDM) searches. However, in these conventional configurations, neutrons undergo a single pass through the crystal geometry, placing a physical constraint on both crystal and in-flight interaction times and measurement sensitivity. In this work, we introduce a neutron loop cavity that coherently recirculates neutrons through repeated Bragg reflections between perfect silicon crystal blades. This structure is predicted to achieve a neutron survival probability of $\sim64~\%$ for 10,000 Bragg reflections, corresponding to confinement times on the order of seconds. We propose a Schwinger interaction measurement that achieves a $π$ spin rotation in 800 Bragg reflections, representing more than a tenfold improvement in sensitivity over recent measurements. Further applications include high-sensitivity nEDM searches targeting the $10^{-27}~$e$\cdot$cm scale, as well as competitive experimental tests of neutron parity violation, the neutron lifetime, and the quantum Zeno effect with neutrons.

Nonlinear contact dynamics are widely regarded as intrinsically nonlinear systems whose behaviour depends strongly on geometry and impact conditions. Here we show that any one-dimensional conservative contact system satisfying monotone energy-consistent conditions admits two complementary structures: (i) a canonical action-angle representation in physical time, and (ii) an exact harmonic oscillator representation under an energy-based coordinate transformation combined with time reparametrisation. This reveals a hidden linear structure underlying nonlinear contact interactions. Building on this result, we derive a unique universal damping law that preserves linear dissipative dynamics in the transformed harmonic space, and establish a rigorous, closed-form lower bound for the critical timestep in numerical simulations. The framework generalises classical power-law contact models and provides a unified basis for restitution control across arbitrary geometries, recovering known exact solutions as explicit monomial special cases.

Asymmetric sound absorption is essential for advanced acoustic manipulation. However, current frequency modulation and broadbanding highly depend on geometric reconfiguration, leading to inevitable structural complexity that impedes their practical applications. Here, we propose a tunable, highly efficient, asymmetric ventilated acoustic system comprising two heterogeneous resonators. Specifically, it couples a highly dissipative space-coiling resonator (SCR) as a dark mode for energy consumption, alongside a weakly damped Helmholtz resonator as a bright mode acting as a reflective soft boundary. Theoretical and numerical analyses reveal strong asymmetry within the deep-subwavelength region (with a resonator size of approximately λ/9.4), achieving 99% absorption for left-incident waves and 98% reflection for right-incident ones. Furthermore, the SCR introduces an interesting degree of freedom for acoustic tuning. Simply rotating the resonator induces a 92% absorption drop (~11 dB attenuation), functioning as an "Acoustic Switch". Moreover, this rotation significantly shifts the operating band. By parallel-coupling multi-angle isomorphic resonators, we achieve efficient broadband absorption (>0.8) from 325 to 375 Hz, offering an attractive paradigm for tunable acoustic metasurfaces and ventilated absorbers.

Confocal microscopy is the cornerstone of cellular biology and biomedical research due to its non-destructive imaging, compatibility with live cells, sensitivity, optical sectioning, and subcellular resolution. To meet the demand for rapid three-dimensional imaging, we propose a novel approach using a mode-selective photonic lantern (MSPL). This fiber-based device transforms single-mode light into multiple linearly polarized modes, allowing simultaneous detection of multiple planes. Using a four-port MSPL to manipulate three group modes (LP$_{01}$, LP$_{11}$, and LP$_{21}$), we demonstrate high-throughput imaging simultaneously with multiple planes. This technique exploits differences in focus sections across modes, enabling individual multi-plane detection via a spatial division multiplexer, with some trade-off in resolution and field of view.

Retractions serve as an indicator of failures in research integrity, yet most analyses focus on absolute counts rather than risk per paper. We use one of the largest open bibliographic databases to develop incidence metrics normalized by population: retractions per publication and per active author annually. Applying an epidemiological framework that models counts with exposure, we find evidence of exponential growth in retraction incidence, with approximately a 5-year doubling time at both the paper and author levels. These patterns vary significantly across fields, publishers, and countries. While scientific output is becoming more democratized globally, retractions are concentrated in fewer countries, creating a "concentration" paradox that calls for targeted monitoring. Despite exponential growth, the absolute incidence remains low (0.12% in 2021), allowing for corrective intervention. Incidence-based monitoring provides a framework for evaluating policies that safeguard research integrity at scale.

Solar Radiation Modification (SRM) may be the only way to limit global warming in the coming decades, leading to increased interest in the subject and to the expansion of related research & development (R&D) activity. Defining the safety and controllability requirements that any SRM system should meet is crucial for directing R&D activities and enabling governments to make informed decisions on the development and possible implementation of such systems. We present an initial proposal for this set of requirements, which also guides Stardust's R&D, as a basis for further discussion and consideration. While we focus on SRM systems based on Stratospheric Aerosol Injection (SAI), the proposed principles may be applicable more broadly.

Magboltz is widely used to compute electron transport properties in gas mixtures for detector applications. Its text-based workflow, however, can be a barrier for routine use, especially for users who are not already familiar with the program. We present Magboltz-GUI, a Python-based graphical user interface for defining gas mixtures, configuring simulation parameters, running Magboltz, and visualizing or exporting the resulting. The tool is designed as a lightweight frontend for common tasks in research and teaching environments involving gaseous detectors, including micropattern technologies such as Micromegas. This paper describes the software implementation, main interface components, and its availability as an open-source distributed package via Python tools.

We demonstrate temporal pattern formation in a coherently driven fiber ring cavity whose effective finesse is continuously reconfigured using distributed Raman amplification. We achieve an effective finesse of up to $\mathcal{F}_{\mathrm{eff}}\approx800$, corresponding to a linewidth of approximately 725 Hz ($Q\approx2.7\times10^{11}$) at 1555 nm. By exploiting the resulting increase in effective photon lifetime, we excite stable temporal cavity solitons and generate a low-repetition-rate frequency comb with a spacing of 580~kHz. Finally, we analyze the impact of the Raman loss-compensation mechanism, particularly its associated noise and show that a trade-off exists between soliton excitation threshold and stability.

This paper presents lineshapes for molecular Optical-Optical Double Resonance (DR) Spectroscopy with arbitrary strength for both pump and probe field using the steady-state solutions for the 3-level density matrix. When the Doppler broadening can be neglected, the results are analytical, and the probe spectrum is a pair of Lorentzian lines that display Autler-Townes splitting, and each has an angular frequency half-width half maximum equal to the relaxation rates, which are all assumed equal. When Doppler broadening is introduced, one must resort to numerical integration except for the limit of weak pump and probe fields. When the Doppler width is assumed much larger than the pump and probe Rabi Frequencies, the calculated DR lineshapes are found to be Lorentzian with a strong pump field limit that is proportional to the pump Rabi frequency, what is commonly known as power broadening. However, the width does not equal the Rabi frequency and is different for co- and counter-propagating pump and probe fields. Furthermore, that broadening is largely inhomogeneous, despite the Lorentzian shape. The saturation power is found to be about 4 times higher than for the bare probe transition with the same relaxation rate, dramatically lower than that expected if the width is interpreted as homogeneous.

A relaxed baseline case, based on the SPARC Primary Reference Discharge (PRD) design point, is used to conduct a thorough investigation for the most unstable low-$n$ MHD instabilities for the first time. The simulations use the high-fidelity 3D extended-MHD code M3D-C1. The linear simulation, by scanning over the resistivity, identifies a dominant internal kink mode at the $q=1$ surface with a toroidal mode number $n=1$. Both the current and the pressure profiles are strongly affecting the kink instability in the baseline case. The linear growth rate is sensitive to the keV-level temperature profile and the on-axis $q_0$ around unity. A simplified 1D eigenvalue solver shows a good qualitative agreement for the observed pressure effects. In 3D nonlinear simulations, the marginally unstable case gives a moderate sawtooth crash soon after $q_0$ drops below unity, likely because of the lack of stabilizing effects in our simulations, such as heating and energetic particles. When both the current and the pressure drives exist (the baseline case), a strong sawtooth is observed, which features a magnetic reconnection event and a hollowed pressure profile. This can be explained by mixing both the Kadomtsev and Wesson models. The actual sawtooth crash may occur in SPARC before $q_0$ drops far below unity due to the sensitive changes of the instability around $q_0\sim 1$. The sawtooth-like oscillations shown in low-$β$ simulations also provides an opportunity to investigate periodic sawtoothing timescales in SPARC. This work forms a basis for understanding particle and heat transport under the influence of MHD instabilities, which can be essential for properly assessing the performance of the SPARC tokamak and future fusion pilot plants.

High-throughput solid-state nanopore experiments generate continuous MHz-rate data streams in which only a small fraction of data contains informative molecular information. This creates storage and processing bottlenecks that limit experimental scalability. We introduce Data Sieving, a GPU-accelerated acquisition framework that integrates real-time event detection directly into the measurement pipeline and selectively stores and allows real-time analysis of snapshots around molecular translocations. The system employs a lightweight rolling-average and min-max trigger to identify event candidates in parallel across channels. This architecture reduces stored data volume by up to 98% while preserving complete molecular signatures across a wide temporal range, from microsecond-scale protein dynamics to second-scale nucleic acid nanoparticle events. Continuous baseline monitoring enables autonomous closed-loop actuation; in high-concentration DNA experiments, automatic declogging restored pore conductance, reducing the time spent in a non-productive clogged state to near-zero and without interrupting parallel measurements. Validated across DNA, protein, and nucleic acid nanoparticle measurements, Data Sieving links data storage directly to molecular information content rather than experiment duration, enabling scalable, real-time operation of parallel nanopore sensors. The approach provides a hardware-agnostic foundation for long-duration, high-bandwidth single-molecule experiments and other event-driven sensing platforms. By using algorithms intrinsically compatible with low-latency digital architectures, this framework provides a clear path toward high-bandwidth, highly multiplexed recording across hundreds of individual nanopore channels in both solid-state and biological pores.

Accuracy, variationality, and convergence underpin the reliability of modern electronic structure methods, yet definitive benchmarks in the relativistic regime remain elusive due to the absence of numerically exact full configuration interaction (CI) references. Recent algorithmic advances in the CI framework, enabled by the small-tensor-product (STP) decomposition approach, have dramatically extended the tractable size of the configuration space, making numerically exact CI calculations feasible in large active spaces previously beyond reach. In this work, we employ the recently developed STP-CI framework to perform large-scale numerically exact CI calculations and directly benchmark relativistic coupled cluster and density matrix renormalization group methods. Definitive benchmarking of approximate relativistic electronic structure methods is ensured through the application of the gap theorem, which provides rigorous error bounds on the CI reference and establishes a controlled standard for assessing accuracy, variationality, and convergence.

Some algorithms for the numerically exact treatment of fermion determinants are summarised. This is not supposed to be a review, rather a concise handbook. The audience is expected to have a basic understanding of how to put fermions on a computer. We primarily discuss different ways to work with the fermion matrix in the "sausage" (Green's function) formulation for quantum Monte Carlo (QMC). We emphasise the need for varied approaches in different space-time volume regimes. In particular, for small spatial volumes we describe a numerically stable method based on dense matrix operations. It is designed specifically to deal with very low temperature regimes. On the other hand, for (relatively) large volumes we describe a highly efficient and scalable sparse matrix approach.

A lithium droplet transport and evaporation model has been developed within the Direct Simulation Monte Carlo code OpenEdge. This model integrates gravity, collisional ion drag, orbital-motion-limited charging, energy-balance evaporation, and an anisotropic rocket recoil force using a Strang-split integrator. Validation against analytical drag-gravity solutions and independent RK45 evaporation integration demonstrates relative errors below 0.00001 for droplet radii of 1.5, 2.5, and 3.5 mm. Simulations of ensembles containing 100000 droplets, launched from inner and outer divertor surfaces in SOLPS-ITER plasma background for the CAT tokamak reactor concept, indicate that transport outcomes are determined by initial size, velocity, and launch location. Outer-divertor droplets predominantly redeposit locally, whereas inner-divertor droplets reach the low-field-side wall. Smaller droplets lose most of their mass to evaporation before reaching the core, while larger droplets retain their mass and redeposit on nearby tiles. Both one-way and iterative two-way coupling frameworks map the evaporated lithium onto the SOLPS-ITER mesh as volumetric sources, facilitating self-consistent evaluation of lithium droplet impacts on edge-plasma performance.

Guided mode resonance (GMR), the resonant coupling of free-space light into leaky waveguide modes, is traditionally achieved with periodic patterned structures. However, this approach makes its key properties such as quality factor (Q-factor) fabrication-dependent and non-tunable. Here, we introduce a time grating platform, i.e., a homogeneous waveguide whose refractive index is modulated periodically in time, that allows tunable GMRs through temporal modulation engineering rather than spatial structural redesign. We show that the Q-factors of these GMRs diverge as the modulation depth vanishes. Furthermore, unconstrained by energy conservation, the resonances exhibit near-unity reflection for fundamental harmonics and values exceeding 40 for first-order harmonics. Our findings not only apply to yield a giant Goos-Hänchen shift over 103 times wavelength without sacrificing the reflection magnitude, but also open new avenues for related phenomena such as bound states in the continuum, unidirectional GMRs and beyond.

Thin-film lithium niobate (TFLN) has emerged as a versatile integrated photonics platform, combining strong electro-optic and nonlinear effects. Among TFLN devices, ring resonators play a central role in filtering, modulation, and nonlinear optical processes. However, intrinsic loss, which ultimately limits ring performance, is most often summarized by single-valued metrics, and its statistical variability across resonances has received limited attention. Here, we show that intrinsic loss rates in monolithic TFLN ring resonators follow a statistical distribution, comprising a baseline loss and a tail arising from discrete loss events. This behavior is revealed by characterizing 2233 resonances, using an adiabatic waveguide-ring coupling architecture that selectively excites the fundamental mode and yields clean spectra in the ultra-high-Qi regime. We find the most probable intrinsic loss rate ki = 2 pi x 10.4 MHz, indicating operation in a low-loss regime comparable to state-of-the-art thick silicon nitride platforms.

Protein rotational kinetics are essential for understanding macromolecular behavior in crowded environments, yet measuring these dynamics at solid-liquid interfaces remains a significant challenge due to low signal strengths. Here, we experimentally demonstrate a label-based optical technique for measuring rotational diffusion kinetics using an all-dielectric multilayer stack that sustains both transverse electric and transverse magnetic polarized surface electromagnetic waves. We introduce the concept of Fluorescence Recovery after Orientational Photobleaching, a rotational analogue to the standard translatory fluorescence recovery after photobleaching technique, which utilizes anisotropic photobleaching via resonant transverse electric excitation followed by real-time monitoring of the orientational relaxation towards isotropy. Our ratiometric analysis of the transverse electric and magnetic polarized fluorescence components allows for a distance-independent estimation of the rotational friction coefficient. Applying this method to covalently bound neutravidin, we observe a rotational friction coefficient (about 5.8E-18 J s) significantly higher than in bulk solutions, highlighting the impact of surface anchoring and molecular crowding. The proposed approach provides a robust, high-sensitivity platform for resolving biomolecular dynamics in complex interfacial environments.

While optical injection locking (OIL) of vertical-cavity surface-emitting lasers (VCSELs) has been widely studied in the past, the polarization dynamics of OIL have received far less attention. Recent studies suggest that polarization locking via OIL could enable novel computational applications such as polarization-encoded Ising computers. However, the inherent polarization preference and limited polarization switchability of VCSELs hinder their use for such purposes. To address these challenges, we fabricate VCSELs with tailored oxide aperture designs and combine these with bias current tuning to study the overall impact on polarization locking. Experimental results demonstrate that this approach reduces the required injection power (to as low as 3.6 μW) and expands the locking range. To investigate the impact of the approach, the spin-flip model (SFM) is used to analyze the effects of amplitude anisotropy and bias current on polarization locking, demonstrating strong coherence with experimental results.

Spectral vision task plays a pivotal role in extracting discriminative spectral-spatial features from high-dimensional data, enabling fine-grained identification beyond human vision. Traditional methods usually involve first collecting rich spectral-spatial information and then using complex algorithms to digitally process it into scene classification and recognition. However, the complexity of processing massive three-dimensional (3D) hyperspectral datasets poses challenges for algorithms. Here, we demonstrate a compressive Hyperspectral Phasor Imaging with Single-pixel detection (HyPIS) that leverages highly compressed spatial-spectral data to achieve spectral task. Two optical encoders are used for wavelength-dependent sine- and cosine-encoding that transforms spectral signals into a two-dimensional (2D) phasor plot. By applying spatial-temporal illumination patterns, a single-pixel detector is enough to reconstruct the phasor image of the object. This allows to directly generate pixel-wise spectral task, bypassing 3D hyperspectral data. Our experiments show that HyPIS can perform real-time classification and recognition tasks of different scenes, reducing the required amount of data by two orders of magnitude, and it can still accurately classify under low light and uneven lighting conditions. This work develops a completely new spectral technology that enables spectral tasks to be performed without obtaining high-resolution hyperspectral datasets, holding promise for spectral applications in mobile devices, robotics, and satellite technologies.

AI applications in fusion is a maturing field, playing a key role as surrogate models and digital twins to overcome computational expense limitations and insufficiently characterised phenomena, and expanding the horizon for real-time applications. The IAEA is supporting this activity through the AI for Fusion Coordinated Research Project (CRP), a five-year initiative launched in 2022, which involves 24 institutions across 11 countries. A key goal is to support the development of modern data infrastructure required to enable the development of agnostic AI models that can be safely extrapolate into the parameter space of future fusion power plants. The IAEA is playing an active role in contributing to the data infrastructure with the Fusion Data Lake project. A modern data platform to enable the development of AI workflows in line with FAIR data principles. The platform comprises three major components: 1. An international data catalogue; 2. A centralised medium-term storage; and 3. A data federation of the various fusion data platforms around the world. The current proof of concept (PoC) demonstrates the data cataloguing and federation capacity by integrating with the UKAEA's MAST Data Catalog. Currently, the second phase of the PoC will demonstrate scalability by integrating two additional experimental fusion device catalogues. This report presents a high-level project overview, including: - Technical architecture and design, collaborations and contributions, and the PoC solution; - Data and metadata model development and ontological concepts; and - The approach to data governance and terms of service. This illustrates the approach, results, and direction of the work, highlighting the high potential value to the fusion community of increasing the visibility and accessibility of the numerous international experimental data sets.

Purpose: This study compares the hierarchical structure of scientific teams across countries and investigates factors associated with the observed cross-national differences. Design/methodology/approach: Drawing on 150,817 publications with author contribution statements, we focus on the 15 countries with the largest volume of scientific publications, examine cross-country variations in the proportion of tall teams, and analyze how this proportion correlates with other factors. Findings: Scientific output from East Asia is dominated by tall teams, which persist after controlling for team size, indicating that this pattern cannot be fully accounted for by the prevalence of larger teams in these countries. Cultural factors, measured by Power Distance, as well as the observed funding patterns of major basic science agencies, are associated with the dominance of tall teams in East Asia. Research limitations: This study is limited by its reliance on publications with author contribution statements, which may introduce selection bias; its focus on cultural and funding factors, while leaving other institutional contexts unexamined; and its use of a leadership concentration measure that does not capture other dimensions of hierarchy. Practical implications: Understanding cross-national differences in research team structures and their associated cultural and institutional factors can inform science policy and team management. Originality/value: This study provides a systematic cross-national comparison of team hierarchy and offers a mechanistic understanding of the dominance of tall teams in East Asia, highlighting associations with cultural and funding factors.

Gas diffusion layers (GDLs) play a crucial role for the performance of proton exchange membrane fuel cells (PEMFCs). The utilization of composite foam-rib flow fields (CFRFFs) can alter the reactant gas transfer pattern, hence improving the efficiency of under-rib reactant gas transfer and water drainage. The impact of the cathode and anode GDL thicknesses (h_{c,GDL} and h_{a,GDL}) on the performance of CFRFF design is investigated by three-dimensional multiphase non-isothermal numerical simulation in this study. The results indicate that for the conventional rib flow field (CRFF) design, there is an optimal h_{c,GDL} for optimal cell performance, while for the CFRFF design, as h_{c,GDL} becomes thinner, the cell performance increases, and the trend is dominated by the variation of the oxygen concentration. Under a thin GDL, the rib width of the CRFF design should be as small as possible to minimize concentration polarization loss, while the rib width of the CFRFF design can be slightly larger. Furthermore, by decreasing the thickness of h_{a,GDL} in both the CRFF and CFRFF designs, there is an increase in the dissolved water content in the ionomer of the cathode CL and a subsequent decrease in the Ohmic polarization loss.

Multiple x-ray beams generated by interference processes in perfect crystals were imaged with a resolution of about 100nm using tele-ptychography in the diffraction direction. These multiple wave-fields, also known as x-ray diffraction echoes, are related to the process known as the Pendelloesung effect and are described by dynamical diffraction theory. The echoes are produced by the constructive interference of diffracted x-rays at the exit surface of the crystal sample. In the imaged diffraction peak, we observed 10 echoes maxima with a total signal length of 78 um. Which translates into a total temporal delay in the signal of less than 108 fs.This makes the echoes of high importance for x-ray optics at x-ray Free Electron Laser sources, as the effect could be used for future ultrafast x-ray beam splitters. In addition to this application, echoes can be exploited to follow ultrafast processes in single crystal micro-structures such as melting or strain propagation.

The terahertz (THz) spectral regime offers unique opportunities for next-generation biochemical sensing due to its non-destructive, label-free probing capability and strong sensitivity to molecular vibrations. However, conventional THz biosensors remain hampered by intrinsically low-quality factors and limited sensitivity, severely restricting their utility for trace-level biochemical and chemical detection. Here, we report an ultrasensitive THz metasurface biosensor that harnesses quasi-bound states in the continuum (QBICs) with sharp resonances and enhanced light-matter interactions to overcome these limitations. As a proof of concept, the device achieves label-free detection of a sulfur-containing amino acid cysteine, with an ultrahigh sensitivity of 492 GHz/RIU and an ultralow detection limit down to 0.00025 mg/mL. The synergy between QBIC-induced field confinement and meticulous structural optimization of the metasurface underpins this performance, marking a significant advance over conventional THz metasurface biosensing schemes. These results establish QBIC-based metasurfaces as a promising platform for ultrasensitive and high-precision biochemical and chemical sensing, with broad implications for medical diagnostics, food safety, and environmental monitoring.

In this paper, we propose a method to analyze correlations in pandemic-related data across different geographical regions, relying on the analysis of correlations for non-stationary time series, which are typical of pandemic data. Unlike traditional epidemiological approaches focused on medical and modeling perspectives during a pandemic, our method emphasizes post-pandemic analysis to assess how societal responses; such as lockdowns, travel restrictions, mobility patterns, and vaccination campaigns, manifest in the collective behavior of regions. These insights can inform future public health strategies and enhance understanding of the complex dynamics underlying pandemic spread and control.

The p-type surface conductivity of H-terminated diamond (HD, H-diamond) has created new path ways for developing diamond based electronic devices as well as chemical and bio-sensors. However, the hydrophobic nature of the HD surface can negatively impact device performance due to its low wettability. Herein, we report the study on polymer electrolyte-gated field effect transistors (EGFETs) fabricated using pristine and partially O-terminated HD films. The HD surface is transformed from hydrophobic to moderate hydrophilic by partial O-termination. Also, the sheet resistance of the HD surface increases from 7.6 to 18.7 k-Ohms per sq. while the sheet hole density decreases from 10.5 to 4.8 x 10^12 cm^-2 upon partial O-termination. Consequently, the ON - OFF ratio of the EGFET devices decreases from ~ 40 to 14 and the maximum transconductance declines from of -150 to -7.9 micro-seimens per V, but the areal capacitance increases from ~ 7.8 to 27.1 microFarad per cm^2 with partial ozonation on HD surface. In addition the in situ Raman measurements in HD EGFET provide direct experimental evidence of a gating-induced blue shift and linewidth broadening of the diamond Raman band which are associated with strong electron phonon coupling. This work highlights the significant impact of the partial O-termination on the performance of the HD EGFET devices and effect of electrochemical gating on the phonon behaviour of the H-diamond.

To investigate the physics of picosecond gain-switching dynamics in single-mode lasers under femtosecond optical pumping at room temperature, we designed and fabricated first-order surface-grating GaAs distributed-feedback (DFB) lasers with five systematically varied grating periods (120-124 nm), corresponding to lasing wavelengths of 825.7-849.5 nm (1.502-1.459 eV). The 124-nm-period device, closest to the quantum-well gain peak among the investigated devices, exhibited the highest output power and spectral bandwidth. Among all devices, the 122-nm-period DFB laser (838.2 nm, 1.480 eV) generated the shortest pulses, despite lasing at a higher photon energy and lower output power than the device closest to the gain peak. All devices exhibited characteristic down-chirp behavior that increased with excitation power. The shortest pulses had a chirped pulse width of 6.6 ps and a chirp rate of 0.13 meV/ps, whereas spectrally resolved measurements revealed a minimum pulse width of 3.8 ps (2.3 ps after deconvolution of the detection time resolution) near the central photon energy of the pulse spectrum. Numerical simulations revealed temporally and spatially resolved dynamics of photons, carriers, gain, and refractive index, reproducing the experimental results qualitatively and quantitatively. Furthermore, a mechanism for generating the shortest pulses at photon energies above the gain peak was identified and attributed to higher differential gain, saturation gain, and a higher transparency carrier density in the high-energy region of the gain spectrum. These experimental and theoretical results elucidate the intrinsic dynamics of picosecond pulse generation in gain-switched DFB lasers and provide design guidance for short-pulse generation and computational tools applicable to both optical and electrical pumping.

Thermal imaging is crucial for night vision but fundamentally hampered by the ghosting effect, a loss of detailed texture in cluttered photon streams. While conventional ghosting mitigation has relied on data post-processing, the recent breakthrough in heat-assisted detection and ranging (HADAR) opens a promising frontier for hyperspectral computational thermal imaging that produces night vision with day-like visibility. However, universal anti-ghosting imaging remains elusive, as state-of-the-art HADAR applies only to limited scenes with uniform materials, whereas material non-uniformity is ubiquitous in the real world. Here, we propose a universal computational thermal imaging framework, TAG (thermal anti-ghosting), to address material non-uniformity and overcome ghosting for high-fidelity night vision. TAG takes hyperspectral photon streams for nonparametric texture recovery, enabling our experimental demonstration of unprecedented expression recovery in thus-far-elusive ghostly human faces -- the archetypal, long-recognized ghosting phenomenon. Strikingly, TAG not only universally outperforms HADAR across various scenes, but also reveals the influence of material non-uniformity, shedding light on HADAR's effectiveness boundary. We extensively test facial texture and expression recovery across day and night, and demonstrate, for the first time, thermal 3D topological alignment and mood detection. This work establishes a universal foundation for high-fidelity computational night vision, with potential applications in autonomous navigation, reconnaissance, healthcare, and wildlife monitoring.

We examine precipitate-induced dynamic strain aging in precipitation-hardened Al-Cu alloys by combining atomistic simulations, kinetic Monte Carlo, and analytical rate theory. Atomistic simulations were used to characterize (1) the energetics of nearest neighbour Cu<->Al exchanges at dislocation - precipitate junctions and (2) the subsequent change in obstacle strength. For robustness, the simulations were performed with two distinct interatomic potentials. The resulting catalog of local Cu-Al exchange events was used as input for a kinetic Monte Carlo model of the time-dependent evolution of obstacle strength during dislocation pinning at the precipitate. The predicted strengthening kinetics were then embedded in an analytical dynamic strain aging model to predict the strain-rate sensitivity parameter. On the whole, the modeling predicts a low strain-rate sensitivity across a broad range of intermediate quasi-static strain rates, consistent with experimental observations for precipitate-strengthened alloys. The results therefore identify a mechanistic origin of the low strain-rate sensitivity in precipitation hardened aluminum alloys, emerging directly from the kinetics of dislocation-precipitate interactions when nearest neighbour Cu<->Al exchanges are considered.

We present TUNA, an open-source quantum chemistry program specifically designed for atoms and diatomic molecules. Within this narrow molecular domain, a broad and consistent set of electronic structure methods and calculation types is available. Energies, optimisations, vibrational frequencies, response properties, coordinate scans and ab initio molecular dynamics trajectories can be accessed through an intuitive command-line interface. A single principle underlies TUNA: once a method can be used to evaluate the energy, all properties follow from numerical differentiation. This makes the program both a transparent teaching platform and a compact environment for benchmarking methods on diatomics $\unicode{x2014}$ among the most simple yet instructive systems in quantum chemistry. Reference implementations including density functional theory, many-body perturbation theory and coupled cluster theory, supported by detailed theoretical documentation, make TUNA an accessible foundation for developing improved methods and algorithms in electronic structure.

The implementation of the Shakhov Bhatnagar-Gross-Krook (SBGK) method in the open-source particle code PICLas is extended for modeling of polyatomic molecules, as well as mixtures including atoms and molecules, while accounting for non-equilibrium in the internal degrees of freedom. The conservation properties of the model are shown and the model parameter for the Prandtl number is derived. In order to determine the viscosity and thermal conductivity of gas mixtures, the first approximation of the transport properties using collision integrals is employed. The model is verified with simulation test cases of a supersonic Couette flow and a hypersonic flow around a 70° blunted cone with different flow parameters and gas compositions. The results are compared to the Direct Simulation Monte Carlo (DSMC) method as well as the Ellipsoidal Statistical BGK (ESBGK) method to assess the accuracy of the model, where overall good agreement is achieved. In particular, the proposed SBGK model captures the shock in front of the 70° blunted cone more precisely than the ESBGK model.

We investigate non-Hermitian photonic crystals in which the lossy and lossless constituents share the same real permittivity and differ only in their imaginary part. We characterize the complex band structure and reflection response of both one-dimensional (1D) and two-dimensional (2D) systems, and show that introducing even a small amount of material loss opens a quasi bandgap at the Brillouin-zone boundary. This quasi bandgap, absent in the lossless limit of the same structure, gives rise to sharp reflectivity peaks whose origin we explain through second-order perturbation theory. As an application of this behavior, we demonstrate a selective reflector combining a conventional photonic-crystal waveguide with a non-Hermitian photonic crystal, achieving wavelength-selective reflection with broadband absorption.

Optical biosensors are indispensable in medical and environmental diagnostics, yet existing approaches are fundamentally limited in their sensitivity due to ensemble-averaged measurements. Digital biosensing has emerged as a promising solution for resolving individual binding events, thereby providing signals at very low analyte concentrations down to the single-molecule level. Here, we present a novel concept for digital optical biosensing empowered by dielectric Mie voids, combining nanoparticle-based contrast enhancement and deep learning for ultrasensitive biomarker detection. The resonantly trapped light in the air cavities of the periodic Mie void arrays ensures strong overlap between the near-fields and the single gold nanoparticles that are captured on the surface in the presence of the protein biomarker. Remarkably, this strong interaction creates high-contrast digital signals for the precise counting of single nanoparticles located both within and outside the voids, yielding efficient use of the entire sensor area for high sensitivity. We employ deep-ultraviolet (DUV) lithography for the scalable and low-cost production of Mie voids in silicon wafers and automated image analysis with a convolutional neural network for robust nanoparticle counting. As a proof of our concept, we demonstrate the detection of an important disease biomarker, interleukin-6 (IL-6), from small sample volumes at concentrations as low as 1.84 pg/ml, within the physiological range of healthy individuals. Owing to its scalability, precision, and adaptability, our digital nanophotonic biosensing approach based on silicon Mie voids establishes a versatile route for applications ranging from bioanalytics to health and environmental monitoring.

The design of triboelectric nanogenerators (TENGs) for efficient energy harvesting requires predictive models that capture the interplay between surface roughness, real contact area, and electrostatic behaviour across diverse tribolayer materials and roughness levels. To address this demand, this paper presents a multiphysics finite element framework that couples mechanical contact analysis with electrostatic simulations, considering exact surface roughness representations rather than idealised statistical approximations. Compared with optical interference microscopy measurements, the framework predicts the real contact area ratio more accurately than analytical models. The proposed approach captures the electrostatic behaviour by scaling the TENG surface charge density with the real contact area ratio between the rough tribolayers, computed for a given mechanical load. This method improves agreement with experiments for open-circuit voltage and capacitance relative to approximate analytical models. To represent the TENG circuit, a time-dependent ordinary differential equation is integrated, enabling evaluation of electrical responses under varying load conditions and elucidating the roles of surface roughness, mechanical load, contact-separation frequency, and resistive load. The framework provides a robust, scalable tool for performance optimisation across dielectric materials, mechanical behaviours, and operating conditions and is readily extendable to other surface-dependent energy-harvesting devices.

To achieve net-zero emissions, cities must transition away from reliance on private vehicles. However, car-centric urban growth has transformed the automobile from a convenience tool into a necessity for accessing essential services, creating significant "car dependency". This study introduces a novel Car Dependency Index (CDI) that quantifies the accessibility gap between private and public transport across 18 cities in Europe and North America. Utilising high-resolution geospatial data and numerical simulations, we reveal pronounced spatial inequalities, showing that car dependency remains a primary driver of car ownership even when accounting for income. A ``what-if" simulation of the planned metro expansion in Rome predicts a reduction of approximately 60,000 commuting vehicles, yet highlights that isolated interventions have localised impacts. We conclude that systemic, network-level transit expansions are essential to dismantle car-based systems and foster equitable, sustainable urban mobility. Our framework provides policymakers with an objective, scalable tool to identify viable areas for car-free zones and target infrastructure investments effectively.

Machine-learning interatomic potentials are widely used as computationally efficient surrogates for density functional theory in atomistic simulations, enabling large-scale, long-time modeling of materials systems. We investigate how different fine-tuning strategies influence the prediction of harmonic phonon band structures, thermal properties, and the potential energy surface along imaginary phonon modes. We achieve substantial accuracy improvements with minimal additional data, with as few as 10 additional training structures already yielding significant gains. In addition to existing approaches, we introduce Equitrain, a finetuning framework that implements LoRA-based adaptation. Across 53 materials systems, we show that fine-tuned models consistently outperform both the underlying pretrained model and models trained from scratch. Equitrain achieves the best overall performance, and our results demonstrate that fine-tuning enables accurate phonon predictions.

Our understanding of gender differences in mobility is marked by a clear tension: surveys portray women's movements as more complex than men's, while digital traces suggest less diverse travel. Here, we resolve the contradiction by modeling trajectories as networks of sequential visits, using smartphone traces linked to self-reported gender for 543,155 individuals across 10 countries. We show that the apparent conflict in the literature arises because women's mobility networks are simultaneously more clustered and more home-anchored -- a nuance obscured by aggregate metrics. This pattern arises because women tend to link multiple destinations within single trips, for trips spanning up to 150 km and multiple days. This organization yields systematically higher travel efficiency, measured as distance saved through destination chaining over monthly sequences.

Dynamic modulation of material properties in space and time enables powerful control over wave propagation, yet existing theories largely rely on idealized, nondispersive models. In realistic media, frequency dispersion can strongly reshape wave dynamics, especially near resonances in highly dispersive platforms such as epsilon-near-zero materials. Here, we develop a general frequency transition theory for electromagnetic scattering at moving interfaces between dispersive media. From phase continuity, we derive nonlinear frequency transition relations and show that dispersion fundamentally reshapes the space-time scattering landscape, enabling additional propagating solutions with no counterpart in nondispersive systems. Applied to Drude, Lorentz and double-Drude media, the theory reveals how resonant dispersion, material loss and negative-index branches reorganize the scattering channels. For the two-wave scattering class, we further introduce a mixed-domain formulation that combines time-domain interface kinematics with frequency-domain constitutive relations, yielding closed-form scattering coefficients. These results establish a unified framework for dispersive space-time scattering and open opportunities for dispersion-based transition engineering in realistic materials.

The non-Hermitian skin effect (NHSE), characterized by extensive boundary accumulation of eigenstates under open boundary conditions, has emerged as a central phenomenon in non-Hermitian physics. Conventionally, the NHSE arises from either non-reciprocal couplings or onsite gain and loss combined with synthetic gauge fields. Existing studies, however, have been largely confined to frameworks with Abelian-coupling, leaving the role of non-Abelian couplings essentially unexplored. Here, we demonstrate that non-Abelian-couplings can generate the NHSE, giving rise to a time-reversal-symmetry-protected Z2 skin effect with pseudospin-dependent boundary localization and dynamical pseudospin separation. Experimentally, we implement a representative four-level model using a programmable topolectrical circuit and directly observe both the predicted NHSE and the boundary-induced pseudospin-inversion reflection. Our work establishes a fundamental link between non-Abelian coupling and non-Hermitian topology, opening new avenues for realizing non-reciprocity-free topological materials and devices.

Precise simultaneous control of both angular and spatial light-field distributions remains a longstanding challenge in optical design, often requiring complex multi-element configurations. In this work, we propose a compact single-lens solution that achieves unified angular-spatial modulation through the co-optimization of double freeform surfaces. The problem is formulated as an extended caustic design that enforces prescribed irradiance patterns on two distinct receptive planes, where the dual-plane constraint implicitly defines the directional characteristics of the light field while preserving spatial accuracy. This framework eliminates the need for auxiliary optical components while delivering performance comparable to that of conventional multi-lens systems. Comprehensive numerical simulations verify the method's effectiveness, demonstrating accurate and stable control of both angular and spatial light-field properties. The proposed approach establishes a practical foundation for compact, high-performance optical systems and provides a promising route toward integrated angular-spatial light-field engineering.

At low temperatures $T$ where $1/T=β\gg1$ the naïve implementation of determinant quantum Monte Carlo (DQMC) methods suffers from loss of precision and numerical instabilities when evaluating the fermion determinant. This instability propagates into the calculation of observables that rely on the evaluation of the inverse of the fermion matrix, or the Greens function. For DQMC methods that rely on the Hamiltonian Monte Carlo (HMC) algorithm, an additional complication comes from evaluating the force terms required for integrating Hamilton's equations of motion, since here loss of precision and numerical instabilities are also prevalent. We show how to address all these issues using various choices of matrix decompositions, allowing us to simulate at $β\gtrsim 90$, which corresponds to room temperature for graphene structures. Furthermore, our implementation has numerical costs that scale similarly to the naïve implementation, namely as $\mathcal{O}(N_x^3N_t)$, where $N_x$ ($N_t$) is the number of spatial (temporal) sites.

Exponential growth describes an extremely rapid process ubiquitous across mathematics and diverse physical, biological, and technological systems. Here, we introduce a class of fractal-inspired lattices that combine long-range periodic order with self-similar hierarchy, establishing a structural motif that enables exponential scaling of topological boundary states. We demonstrate this phenomenon in (i) a quasi-one-dimensional lattice chain constructed from Koch-curve unit cells and (ii) a two-dimensional periodic tiling lattice composed of Sierpinski-gasket unit cells. We show that, for suitable coupling parameters, both the number of topological boundary states $N_{\ell}$ and the number of topological minigaps $M_{\ell}$ grow exponentially with the fractal generation index $\ell$. We find that $N_{\ell}$ is an integer multiple of $M_{\ell}$, with the integer determined by the underlying symmetry. This hierarchical scaling law is captured by multi-topological-phase theory and confirmed experimentally in laser-written photonic lattices. Our results identify fractal hierarchy as a materials architecture principle for controlling boundary-state multiplicity, revealing an interplay between topology, self-similar geometry, and periodic order. More broadly, this work suggests a route to synthetic materials and integrated photonic platforms in which large numbers of robust boundary modes can be engineered within compact architectures.

This paper presents a p-adaptive high-order hybridizable discontinuous Galerkin spectral element method (HDG-SEM) for solving the Poisson equation in electrostatic plasma simulations using particle-in-cell (PIC) schemes. This approach enables element-local refinement of the polynomial degree, concentrating computational effort specifically in regions with strong gradients. Thus, the method significantly reduces the global number of degrees of freedom compared to uniform high-order methods. The proposed method is implemented in the open-source framework PICLas and validated through a series of benchmark test cases, including a dielectric sphere and a one-dimensional plasma sheath. Finally, a two-dimensional axisymmetric simulation of an ion optic demonstrates the method's capability to efficiently model complex plasma phenomena but also highlights current limitations.

Reliable confinement and stable performance of tokamak fusion plasmas require accurate real-time magnetic shape control. A promising route to reduced latency and increased flexibility in plasma control systems (PCS) is to emulate physics-based controllers using neural networks. In prior work, we have demonstrated that virtual circuits (VCs), which define the poloidal field coil current vectors able to modify each plasma shape parameter independently, can be accurately emulated with neural network models trained on a large library of simulated Grad-Shafranov equilibria. This enables magnetic controllers to accurately adapt to evolving plasma equilibria, in contrast to pre-set VC schedules whose performance degrades upon departure from their reference equilibria. Here, we investigate the performance and robustness of these emulators in closed-loop simulations using the FreeGSNKE Pulse Design Tool (FPDT): a framework that couples the FreeGSNKE evolutive equilibrium solver with a virtual PCS. The FPDT models the coupling between controllers, plasma current and shape response, and actuator constraints. Using the emulated VCs within the FPDT, we demonstrate effective in-silico control of MAST Upgrade (MAST-U) plasma scenarios and show that the emulators are robust in the presence of input measurement uncertainty and under different update frequencies. These results establish the viability of neural network emulated VCs for closed-loop plasma shape control, representing a key step toward real-time deployment in the MAST-U PCS.

Reconfigurable metalenses capable of large focal length tuning, fast response times, and high focusing efficiency while maintaining diffraction-limited operation are highly desirable for next-generation adaptive imaging systems. Phase change chalcogenides provide a promising platform for such devices by exploiting the reversible amorphous-to-crystalline transition to achieve non-volatile tuning with relatively fast switching. However, extending these approaches towards the visible spectrum is challenging because of the reduced meta-atom dimensions, stringent phase coverage requirements, intrinsic material absorption and the need to simultaneously preserve focusing efficiency and image quality across multiple material states. Here, we present a dynamically tunable metalens based on $Sb_2$$S_3$ operating at the edge of the visible spectrum. The design framework combines finite element computations with a genetic algorithm-based inverse design approach to achieve robust phase control in both amorphous and crystalline states. The resulting metalens shows diffraction-limited performance with a minimum focal length tunability of $33\%$, focusing efficiencies of approximately $40\%$ and Strehl ratios of $0.80$ and $0.77$ in the amorphous and crystalline states, respectively.

The 15-minute city is a powerful planning concept to counter car-dependence by promoting active mobility to amenities and fostering inclusive urban environments. However, this policy has challenges in amenity-poor urban peripheries. Public transport remains underexplored in this discourse despite its role in distant access. Here, we propose a framework that incorporates public transport into the 15-minute city model using openly available data. By comparing Helsinki, Madrid, and Budapest, we demonstrate that multimodal mobility substantially increases access to amenities and enhances socio-spatial integration within a 15-minute reach. Although urban periphery benefit significantly from radial or high-speed public transport lines in their social mixing potential, such lines alone do not improve their access to amenities. These findings underscore the need to optimize polycentric public transport networks that can improve inclusive urban accessibility and complement active mobility in polycentric cities.

A rapid expansion of system flexibility is essential to integrate increasing shares of renewable energy into future energy systems. However, flexibility needs and technology-specific contributions to flexibility remain poorly quantified in energy system modelling. Existing methods are not widely applied, leaving key questions unanswered: which flexibility technologies are critical for climate neutrality, and what are the cost implications of alternative deployment strategies? To address this gap, we apply a correlation-based flexibility metric to a high-resolution, sector-coupled model of the German energy system, covering its transformation towards climate neutrality. For our default scenario, we find that daily flexibility needs increase by a factor of 3.7 between 2025 and 2045, driven primarily by the expansion of solar PV. By 2045, stationary batteries provide 38% of daily flexibility, while flexible electric vehicle charging contributes 30%. Systems with constrained flexibility increase system costs by 6.9%, electricity prices by 14 EUR/MWh and trigger 47% higher hydrogen and e-fuel imports compared to an unconstrained system in 2045. In contrast, scenarios with high shares of flexible electric vehicle charging, vehicle-to-grid, and industrial demand-side management achieve system cost reductions of 3.3%, while also reducing import dependence. Higher flexibility also reduces electricity price ranges, decreases average electricity prices by 3 EUR/MWh, and reduces backup capacity by 22% (22 GW). Overall, our results highlight the decisive role of specific flexibility technologies in achieving cost-efficient and energy-secure climate-neutral energy systems, providing quantitative guidance for policy and investment decisions.

A novel particle merging algorithm for rarefied gas dynamics simulations is proposed that can conserve arbitrary velocity and spatial moments of the particle distribution via solving a non-negative least squares problem. An extension that preserves both exact and approximate collision rates is also derived. The algorithm is applied to the simulation of several model rarefied gas dynamics problems, where it exhibits noticeably lower merging-induced error in key macroscopic quantities.

Anisotropic photonic time crystals, enabled by periodic temporal modulation of a uniform anisotropic medium, exhibit asymmetric momentum-bandgap structures and offer unique control over light-matter interactions. Here, we introduce and construct temporal Weyl points in APTCs within a synthetic three-dimensional space defined by two phase parameters and the quasi-frequency. The temporal response reveals robust Fermi arcs linking TWPs of opposite topological charge. Unlike spatial counterparts, these Fermi arcs emerge only after the first temporal supercell comprising multiple periods of APTCs, reflecting causality. We further show that TWPs generate a directional near-zero radiation trajectory in momentum space with tunable radiation from stationary charges embedded in APTCs, while the associated Fermi arcs robustly suppress radiation at selected directions and frequencies. Our findings establish temporal Weyl physics in photonic time crystals and uncover new opportunities for topological control of light-matter interactions through the time dimension.

Vector beams with spatially structured polarization and intertwined spin-orbital angular momentum (SAM-OAM) provide powerful degrees of freedom for tailoring light-matter interactions. While such structured beams are well established in the visible and infrared regimes, extending them to the extreme-ultraviolet (EUV) and soft X-ray (SXR) domains at relativistic intensities remains a major challenge. Here, we investigate the generation of higher-order harmonic vector beams driven by relativistic laser-plasma interactions. Combining theoretical analysis with three-dimensional particle-in-cell simulations, we elucidate the underlying physical mechanisms governing the transfer and conversion of polarization and orbital angular momentum during harmonic generation. We demonstrate that both the polarization topology and OAM of the emitted harmonics can be deterministically controlled by the topological charges of the driving field. Owing to the intrinsic properties of vector beams, either few-cycle driving pulses or vector polarization gating applied to multi-cycle pulses enable the production of intense isolated attosecond pulses featuring spiral wavefronts and spatially tailored polarization states. These results establish a pathway toward high-intensity structured light sources in the EUV and SXR regimes and open new opportunities for ultrafast and strong-field light-matter interaction studies with engineered angular momentum.

We investigate inviscid numerical instabilities that arise in simulations of axisymmetric flow over a hypersonic sphere in an inert, calorically perfect gas at low specific heat ratio ($γ\approx 1.1$--$1.2$). We show that when the density ratio across the bow shock is high and the computational mesh is relatively coarse, numerically induced traveling-wave instabilities of the carbuncle type can develop in the shock layer near stagnation for inert gases. These instabilities, not previously documented in the literature, are noteworthy because bow shock oscillations are also observed experimentally in polyatomic gases exhibiting post-shock thermochemical relaxation. When such gases are modeled as inert with an effectively low $γ$, our results emphasize the need for caution to avoid conflating genuine physical instabilities with numerical artifacts in simulations.

Laser speckle, the granular intensity pattern arising from random optical interference, provides a high-dimensional encoding of spectral information that can be exploited for precision metrology. Speckle-based spectrometers have advanced rapidly owing to their compact footprint, mechanical robustness and alignment agnostic nature, yet their spectral resolution has remained limited to the picometre scale. In this work, we break this limit by employing an integrating sphere as a multiply scattering cavity with access to a high range of path lengths to enhance spectral sensitivity. At 780$\,$nm, the resulting device achieves a resolution of 6$\,$fm, corresponding to a resolving power of $1.3\times10^8$, representing an approximately 80-fold improvement over previous implementations. This ultra-high resolution enables clear discrimination of laser sidebands generated by an electro-optical modulator, with extracted sideband powers agreeing with expected values to within 1%. It further permits the first direct speckle-based measurement of the hyperfine structure of the $\text{D}_{2}$ transition in $^{85}\text{Rb}$, with transmission spectra differing by no more than 3.6% from independent wavemeter-referenced measurements. These results establish speckle as a new platform for ultra-high precision spectroscopy, radio-frequency spectrometry, and microwave photonics.

TESS (Transiting Exoplanet Survey Satellite) has produced long-term photometry for millions of stars across the sky. In this work, we present an asteroseismic catalogue of 19,151 red giants in the TESS Continuous Viewing Zones using sectors 1--87 (Years 1--7). We visually assessed the power spectra for oscillations, and then applied the computationally efficient nuSYD method to confirm reliability. We identified an increase of 80% in the number of previously known oscillating red giants at a TESS magnitude $>$ 8. We determined the frequency of maximum power ($\rm ν_{max}$) and the large frequency separation ($\rm Δν$) using the pySYD pipeline, achieving typical precisions of 1.5% and 1%, respectively. We classified the stars into Red Giant Branch (RGB) and Core Helium Burning (CHeB) classes using a Convolutional Neural Network. Using spectroscopic data for 10,298 stars with reliable asteroseismic measurements, we have been able to measure stellar mass and radii with precisions of 7.5% and 2.8%, which is comparable to that from 4-yr $Kepler$ data. A comparison of the seismic radii with Gaia radii shows excellent agreement. With three years of TESS data, the asteroseismic parameters are precise enough to identify the RGB bump and delineate the Zero Age Helium Burning edge. Combined with astrometric data, these parameters reveal established trends across the Galactic plane, providing a valuable set of uniformly determined asteroseismic parameters for Galactic Archaeology.

The bacterial flagellar motor (BFM) is a rotary molecular machine driven by the ion electrochemical potential across the cell membrane. Recent cryo-EM structures reveal a cogwheel-like architecture in which multiple stators engage a large rotor. A longstanding puzzle is the directional asymmetry of its torque-speed relation: concave in counterclockwise (CCW) rotation but nearly linear in clockwise (CW) rotation. Here, we develop a stochastic mechanochemical model that explicitly incorporates rotor-stator coupling and detailed ion translocation kinetics. By integrating physiological torque-speed data with recent measurements of rotor-stator relative motion, we show that under physiological conditions the motor operates in a tight engagement regime, rendering the torque-speed relation largely insensitive to the specific form of mechanical interactions. This finding rules out differences in rotor-stator mechanics as the origin of CW-CCW asymmetry. Guided by cryo-EM structures, we propose a contact-dependent gating mechanism in which the MotA-FliG interaction modulates the ion release rate of the MotB subunit proximal to the FliG ring. Molecular dynamics simulations indicate tighter MotA-FliG contact in the CW motor, implying a reduced ion release rate compared to CCW. Our model demonstrates that differential gating strength accounts for the observed asymmetry: stronger gating in CCW shortens torque-free waiting phases, enhances torque generation, and produces a concave torque-speed curve, whereas weaker gating in CW yields lower torque and a linear relation. This structure-based framework quantitatively links molecular asymmetry to motor function and identifies specific interfaces for targeted perturbation and mutational studies.

A new neural network model for a quasilinear saturation rule has been developed to map linear gyrokinetic data to nonlinear saturated potential magnitudes to predict the total energy and particle fluxes. The training dataset is taken from the high resolution simulation database generated from nonlinear gyrokinetic turbulence simulations with the CGYRO code for developing the SAT3 model. This new model, named SAT3-NN, overall is able to capture the 1D saturated potential magnitudes of the dataset more accurately than SAT3, as depicted by lower percentage errors in the peak locations and peak values of the 1D saturated potentials. The resulting fluxes also had smaller deviations from the nonlinear CGYRO data as compared to previous saturation models such as SAT0 - SAT2. Consistent with SAT3, SAT3-NN is able to recreate the anti-gyroBohm scaling of fluxes seen for the TEM-dominated cases considered.

Machine learning has become a useful tool for studying phase transitions in statistical systems.For the two-dimensional classical XY model, however, the topological character of the Berezinskii-Kosterlitz-Thouless (BKT) transition and pronounced finite-size effects make it nontrivial to extract robust size-dependent pseudo-critical temperatures from configuration data. Existing studies often stop at phase classification, leaving open how standard neural-network outputs can be turned into quantitatively testable observables. Here we develop a machine learning-assisted framework for the 2D XY model that uses standard network outputs to extract the size-dependent sequence of pseudo-critical temperatures T(L). Specifically, we generate Monte Carlo configurations using embedded cluster updates, train a standard ResNet18 only on samples from the Quasi-ordered Phase and the Disordered Phase, and determine T(L) from bootstrap-averaged probability curves using the 50% crossing criterion. We then analyze the finite size drift of this temperature sequence using BKT-motivated scaling and compare it with susceptibility-peak temperatures. The resulting temperature sequence shows a systematic finitesize drift consistent with BKT-type behavior and remains in the same fluctuation window as the susceptibility peak, supporting its interpretation as a finite-size pseudo-critical temperature. More broadly, this framework provides a practical route for converting standard neural-network outputs into physically interpretable finite-size observables in systems with strong crossover or topological transition signatures

Acousto-optic (AO) interactions provide a powerful interface between the microwave and optical domains, enabling functionalities such as optical switching, non-reciprocal propagation and efficient microwave-to-optical transduction. Integrated demonstrations to date have largely relied on thin-film lithium niobate (TFLN), which offers strong piezoelectric response and low optical loss performance. Here, we establish lithium tantalate on insulator (LTOI) as a scalable platform for integrated acousto-optics. LTOI combines intrinsically low birefringence, high optical damage threshold, strong electro-optic and Kerr nonlinearities, and superior acoustic quality factors with a mature high-volume manufacturing base. We demonstrate for the first time acousto-optic modulation on the LTOI platform. By exploiting the anisotropy of surface acoustic waves, we reveal a direct correlation between acousto-optic modulation efficiency and the electromechanical coupling coefficient of lithium tantalate. In particular, acoustic excitation along the crystal Z-axis enhances the higher-order R1 mode, yielding the highest modulation efficiency. Our Mach-Zehnder interferometers achieve a modulation efficiency of 0.68 $\mathrm{\mathbf{V \cdot cm}}$, while racetrack resonators reach 0.022 $\mathrm{\mathbf{V \cdot cm}}$ -representing, to the best of our knowledge, the lowest $\mathrm{V_πL}$ demonstrated in non-suspended ferroelectric platforms. This record performance directly enables microwave-to-optical conversion without suspended structures, establishing LTOI as a robust and scalable platform for integrated acousto-optics with broad applications in communications, signal processing, and quantum information technologies.

Convergent evolution provides powerful evidence for natural selection, yet its molecular basis is typically sought in protein-coding amino acid substitutions. Whether adaptive pressures can drive the convergent evolution of synonymous codon usage bias (CUB) to override phylogenetic history remains a fundamental question. Here, we investigate this within the rapidly radiating fern family Thelypteridaceae by establishing a comparative framework that integrates chloroplast phylogenomics with dimensionality reduction of codon usage, morphological data, and divergence time estimation. Our results reveal that chloroplast CUB patterns are strikingly incongruent with the phylogeny of this family. Instead, they partition species into distinct clusters that strongly correlate with a convergently evolved morphological trait, lamina base architecture, a key adaptation whose radiation we date to the early Neogene. This convergent molecular signal is driven by a specific subset of photosynthesis-related genes (ndhJ, psaA, and psbD), which exhibit a high density of type-specific, third-position codon substitutions. These findings demonstrate that CUB can serve as a powerful, quantifiable indicator of adaptive history, revealing a cryptic layer of molecular convergence linked to the regulation of protein synthesis. Our work providing a new framework for uncovering adaptive histories obscured by complex evolutionary processes.

Speech production requires the rapid coordination of a complex hierarchy of linguistic units, transforming a semantic representation into a precise sequence of articulatory movements. To unravel the neural mechanisms underlying this feat, we leverage recordings from eight 3.2 x 3.2 mm 64-microelectrode arrays implanted in the motor cortex and inferior frontal gyrus of two patients tasked to produce twenty thousand sentences. We show that a hierarchy of linguistic features are robustly encoded in most of these small cortical patches. Contrary to our expectations, instead of a clear macroscopic organization between patches, we observe a multiplexing of phonetic, syllabic and lexical representations within each cortical patch. Critically, this coding scheme dynamically changes over time to allow successive phonemes, syllables and words to be simultaneously represented without interference. Overall, these results, reminiscent of position encoding in transformers, show how small cortical patches organize the unfolding of the speech hierarchy during language production.

We study high-dimensional mediation analysis in which exposures, mediators, and outcomes are all multivariate, and both exposures and mediators may be high-dimensional. We formalize this as a many (exposures)-to-many (mediators)-to-many (outcomes) (MMM) mediation analysis problem. Methodologically, MMM mediation analysis simultaneously performs variable selection for high-dimensional exposures and mediators, estimates the indirect effect matrix (i.e., the coefficient matrices linking exposure-to-mediator and mediator-to-outcome pathways), and enables prediction of multivariate outcomes. Theoretically, we show that the estimated indirect effect matrices are consistent and element-wise asymptotically normal, and we derive error bounds for the estimators. To evaluate the efficacy of the MMM mediation framework, we first investigate its finite-sample performance, including convergence properties, the behavior of the asymptotic approximations, and robustness to noise, via simulation studies. We then apply MMM mediation analysis to data from the Alzheimer's Disease Neuroimaging Initiative to study how cortical thickness of 202 brain regions may mediate the effects of 688 genome-wide significant single nucleotide polymorphisms (SNPs) (selected from approximately 1.5 million SNPs) on eleven cognitive-behavioral and diagnostic outcomes. The MMM mediation framework identifies biologically interpretable, many-to-many-to-many genetic-neural-cognitive pathways and improves downstream out-of-sample classification and prediction performance. Taken together, our results demonstrate the potential of MMM mediation analysis and highlight the value of statistical methodology for investigating complex, high-dimensional multi-layer pathways in science. The MMM package is available at https://github.com/THELabTop/MMM-Mediation.

Dynamic fading channels are modeled at two fundamentally different levels of abstraction. At the physical layer, the standard representation is a correlated Gaussian process, such as the dB-domain signal power in log-normal shadow fading. At the link layer, the dominant abstraction is the Gilbert-Elliott (GE) two-state Markov chain, which compresses the channel into a binary ``decodable or not'' sequence with temporal memory. Both models are ubiquitous, yet practitioners who need GE parameters from an underlying Gaussian fading model must typically simulate the mapping or invoke continuous-time level-crossing approximations that do not yield discrete-slot transition probabilities in closed form. This paper provides an exact, closed-form bridge. By thresholding the Gaussian process at discrete slot boundaries, we derive the GE transition probabilities via Owen's $T$-function for any threshold, reducing to an elementary arcsine identity when the threshold equals the mean. The formulas depend on the covariance kernel only through the one-step correlation coefficient $ρ= K(D)/K(0)$, making them applicable to any stationary Gaussian fading model. The bridge reveals how kernel smoothness governs the resulting link-layer dynamics: the GE persistence time grows linearly in the correlation length $T_c$ for a smooth (squared-exponential) kernel but only as $\sqrt{T_c}$ for a rough (exponential/Ornstein--Uhlenbeck) kernel. We further quantify when the first-order GE chain is a faithful approximation of the full binary process and when it is not, reconciling two diagnostics, the one-step Markov gap and the run-length total-variation distance, that can trend in opposite directions. Monte Carlo simulations validate all theoretical predictions.

To support the high data rates for latency-critical applications, future wireless systems will employ fully digital beamforming multiple-input multiple-output (MIMO) architectures at millimeter wave (mmWave) frequencies. Moreover, mmWave MIMO deployments will coexist with conventional sub-6 GHz MIMO systems, creating opportunities to exploit out-of-band sub-6 GHz information to enhance channel estimation at mmWave frequencies. In this work, we analyze the pilot-aided channel estimation performance of mmWave MIMO systems under various pilot configurations in both static and dynamic environments. We evaluate the system performance in terms of spectral efficiency (SE) for line-of-sight and non-line-of-sight propagation conditions. Simulation results show that incorporating out-of-band sub-6 GHz information yields notable SE gains in both static and dynamic scenarios.

Tomographic synthetic aperture radar (TomoSAR) enables three-dimensional imaging by resolving targets along the elevation dimension, which is essential for environment reconstruction and infrastructure monitoring. A critical challenge in TomoSAR is the severe multipath propagation that causes ghost targets, range offsets, and elevation ambiguities. To address this, this paper proposes an enhanced Newtonized orthogonal matching pursuit (NOMP) algorithm to extract the delay, Doppler, and complex amplitude parameters of each propagation path, effectively separating line-of-sight (LoS) and multipath components prior to TomoSAR processing. Additionally, a height fusion strategy combining TomoSAR estimates with LoS-ground reflection delay-based inversion improves elevation accuracy. Simulation results demonstrate that the proposed method achieves improved positioning and elevation accuracy while effectively suppressing multipath-induced artifacts.

Future wireless communication systems will integrate both sub-6 GHz and millimeter wave (mmWave) frequency bands within multi-antenna architectures to meet the increasing demand for high data rates. In such multi-band systems, reliable information obtained from the sub-6 GHz band can be exploited to support communication at mmWave frequencies. To ensure that both systems experience similar multi-path propagation effects, the sub-6GHz and mmWave antenna arrays have to be colocated and precisely aligned. However, such a configuration may adversely alter the radiation characteristics of the arrays, potentially degrading their performance. In this paper, we investigate the impact of positioning a mmWave antenna structure in front of a sub-6 GHz antenna structure. Through both simulations and measurements, we evaluate how the presence of the mmWave structure affects the radiation pattern of the sub-6 GHz one. The results demonstrate that the influence of the mmWave structure on the sub-6 GHz performance is minor, indicating that co-located configurations are feasible with negligible degradation.

Sensor-based Human Activity Recognition (HAR) underpins many ubiquitous and wearable computing applications, yet current models remain limited by scarce labels, sensor heterogeneity, and weak generalization across users, devices, and contexts. Foundation models, which are generally pretrained at scale using self-supervised and multimodal learning, offer a unifying paradigm to address these challenges by learning reusable, adaptable representations for activity understanding. This survey synthesizes emerging foundation models for sensor-based HAR. We first clarify foundational concepts, definitions, and evaluation criteria, then organize existing work using a lifecycle-oriented taxonomy spanning input design, pretraining, adaptation, and utilization. Rather than enumerating individual models, we analyze recurring design patterns and trade-offs across nine technical axes, including modality scope, tokenization, architectures, learning paradigms, adaptation mechanisms, and deployment settings. From this synthesis, we identify three dominant development trajectories: (1) HAR-specific foundation models trained from scratch on large sensor corpora, (2) adaptation of general time-series or multimodal foundation models to sensor-based HAR, and (3) integration of large language models for reasoning, annotation, and human-AI interaction. We conclude by highlighting open challenges in data curation, multimodal alignment, personalization, privacy, and responsible deployment, and outline directions toward general-purpose, interpretable, and human-centered foundation models for activity understanding. A complete, continuously updated index of papers and models is available in our companion repository: https://github.com/zhaxidele/Foundation-Models-Defining-A-New-Era-In-Human-Activity-Recognition.

With the development of 6G technologies, traditional uniform linear arrays (ULAs) and uniform planar arrays (UPAs) can hardly meet the demands of three-dimensional (3D) full-space coverage and high angular resolution. Spherical antenna arrays (SAAs), with elements uniformly distributed on a spherical surface, provide an effective solution. This article analyzes the issues of traditional arrays, summarizes the advantages and typical structures of SAAs, discusses their potential application scenarios, and verifies their superiority over UPAs via a case study. Finally, key technical challenges and corresponding research directions of SAAs are identified, providing a reference for their research and application in future wireless communications.

The widespread use of unmanned aerial vehicles (UAVs) in low-altitude airspace has raised significant safety and security concerns, motivating the development of reliable non-cooperative UAV surveillance technologies. Integrated sensing and communication (ISAC), enabled by multiple-input multiple-output (MIMO) architectures and orthogonal frequency-division multiplexing (OFDM) waveforms, has emerged as a promising paradigm for leveraging cellular infrastructure to support large-scale sensing without additional hardware deployment. This paper presents the first comprehensive survey dedicated to MIMO OFDM-enabled ISAC for low-altitude non-cooperative UAV surveillance, where the targeted UAVs do not intentionally assist the monitoring system through dedicated signaling or prior coordinate sharing. We first analyze the unique propagation characteristics of low-altitude UAV sensing, including severe clutter, rapid channel variations, and mixed near/far-field effects, and discuss corresponding waveform design principles. We then systematically review existing MIMO OFDM-enabled UAV surveillance techniques along four key dimensions: ISAC system modeling and network optimization, UAV detection and tracking algorithms under single and networked base station (BS) architectures, UAV identification techniques based on micro-Doppler and learning-based approaches, and experimental validations and practical field trials. Subsequently, we summarize open challenges such as sensing under severe clutter and multipath, data scarcity for identification, cooperative multi-BS fusion, and real-world deployment constraints. Finally, we outline promising future research directions toward 5G-Advanced (5G-A) and 6G-enabled low-altitude surveillance systems.

Stackelberg prediction games (SPGs) model strategic data manipulation in adversarial learning via a leader--follower interaction between a learner and a self-interested data provider, leading to challenging bilevel optimization problems. Focusing on the least-squares setting (SPG-LS), recent work shows that the bilevel program admits an equivalent spherically constrained least-squares (SCLS) reformulation, which avoids costly conic programming and enables scalable algorithms. In this paper, we develop a simple and efficient alternating direction method of multiplier (ADMM) based solver for the SCLS problem. By introducing a consensus splitting that separates the quadratic objective from the spherical constraint, we obtain an augmented Lagrangian formulation with closed-form updates: the primal quadratic step reduces to solving a fixed shifted linear system, the constraint step is a projection onto the unit sphere, and the dual step is a lightweight scaled ascent. The resulting method has low per-iteration complexity and allows pre-factorization of the constant system matrix for substantial speedups. Experiments demonstrate that the proposed ADMM approach achieves competitive solution quality with significantly improved computational efficiency compared with existing global solvers for SCLS, particularly in sparse and high-dimensional regimes.

The increasing, high-risk interactions between vehicles and vulnerable micromobility users, such as e-scooter riders, challenge vehicular safety functions and Automated Driving (AD) techniques, often resulting in severe consequences due to the dynamic uncertainty of e-scooter motion. Despite advances in data-driven AD methods, traffic data addressing the e-scooter interaction problem, particularly for safety-critical moments, remains underdeveloped. This paper proposes a pipeline that utilizes collected on-road traffic data and creates configurable synthetic interactions for validating vehicle motion planning algorithms. A Social Force Model (SFM) is applied to offer more dynamic and potentially risky movements for the e-scooter, thereby testing the functionality and reliability of the vehicle collision avoidance systems. A case study based on a real-world interaction scenario was conducted to verify the practicality and effectiveness of the established simulator. Simulation experiments successfully demonstrate the capability of extending the target scenario to more critical interactions that may result in a potential collision.

In this paper, the problem of maximizing the sum-rate is addressed for a multi-user uplink scenario that is assisted by an active reconfigurable intelligent surface (RIS). The maximization is achieved by optimizing the beamforming at the base station, the users' transmit power, active RIS elements phase shifts, and active gains in presence of imperfect channel state information (CSI). The non-convex maximization problem is decomposed into sub-problems and solved via iterative approaches including the Lagrangian method, the projected gradient descent, multi-variate Taylor expansion and fractional programming. Numerical results show that the active RIS is more sensitive to CSI imperfections than passive one at high error variances.

Electrocardiogram (ECG) foundation models represent a paradigm shift from task-specific pipelines to generalizable architectures pre-trained on large-scale unlabeled waveform data. This survey presents a unified and deployment-aware review of foundation models and medical large language models (LLMs) for ECG intelligence in cardiovascular disease (CVD) diagnosis, monitoring, and clinical decision support. The central thesis of this survey paper is that next-generation cardiovascular AI systems will be inherently agentic, requiring the synergistic integration of two complementary model classes: (i) ECG foundation models that act as signal-level interpreters, learning rich electrophysiological representations via self-supervised and multimodal pretraining, and (ii) medical LLMs, trained on biomedical text corpora, that function as knowledge-based reasoning backbones for contextual inference, guideline alignment, and clinical decision support. Thus, the survey systematically reviews existing pool of generalist medical LLMs, as well as ECG foundation models that utilize techniques such as self-supervised learning, multimodal ECG-language alignment, vision transformer architectures, and possess capabilities such as zero-shot classification, automated report generation, and longitudinal risk modeling. Recognizing the constraints of consumer-grade wearable edge devices, we further examine model optimization techniques such as quantization, pruning, knowledge distillation, as well as the role of small language models in enabling low-latency, energy-efficient, and privacy-preserving ECG intelligence on edge platforms such as smartwatches. Finally, we outline future directions in multimodal ECG foundation models, agent-driven monitoring, and explainable, secure edge intelligence, with particular emphasis on real-time, on-device cardiovascular analytics in consumer electronics ecosystems.

Null forming is increasingly essential in modern wireless systems for spectrum-sharing, anti-jamming, and covert communications in contested and congested environments. Achieving deep nulls, however, is far more demanding than conventional beam steering: nulls are intrinsically narrow, and even small phase, timing, or gain mismatches across RF chains can significantly degrade suppression. This work develops and validates a self-calibrating SDR architecture tailored for high-fidelity null forming using a compact reference transmitter directionally coupled to the antenna feeds. We demonstrate the effectiveness of the approach through simulation and experimental measurements on an SDR platform operating from 3.0 to 3.5GHz, a band of growing importance for Department of Defense spectrum-sharing initiatives.

In sixth-generation (6G) networks, the deployment of large numbers of Internet of Things (IoT) users (IU) necessitates efficient resource utilization and reliable connectivity, making resource allocation a critical factor. Specifically, the upper mid-band (FR3) spectrum has emerged as a promising candidate for 6G systems due to its favorable balance between bandwidth availability and coverage. However, translating these spectral advantages into performance gains in dense IoT environments requires intelligent management of interference and propagation impairments. In this paper, we propose a reconfigurable intelligent surface (RIS)-assisted IoT network operating in the FR3 band to enhance coverage and improve signal quality. Furthermore, we formulate a joint power allocation and IU-RIS association problem to maximize the achievable sum rate under practical channel conditions and power constraints. The resulting problem is nonconvex and combinatorial due to interference coupling and binary association variables. To address this challenge, we develop a multiphase resource allocation framework that integrates a successive convex approximation (SCA)-based power allocation scheme combined with a matching-theory-based user association algorithm. Simulation results demonstrate that the proposed scheme significantly outperforms conventional greedy and random search schemes in terms of sum-rate enhancement.