The Inference Report

Procedural generators produce useful verifiable reasoning problems at scale, but have received less attention as data for completion-supervised fine-tuning. We introduce Reasoning Core, a collection of 50 generators spanning mathematics, logic, planning, state tracking, formal languages, structured data, games, causality, and code, with semantic scorers, difficulty controls, and task evaluators. Under a matched completion-supervised protocol, we compare Reasoning Core with Procedural Warmup, Reasoning Gym, and SynLogic across four base-model settings and multiple training durations. In the primary 3B comparison, Reasoning Core achieves the highest mean scores on DROP, LogiQA, and ARC-Challenge, exceeding both the baseline without procedural data and all three alternative procedural collections. Task-level analyses show that semantic validity alone does not ensure training utility, highlighting compact targets and calibrated difficulty as important design factors. We ran audits combining model-assisted review, human adjudication, and regression testing. Applied throughout Reasoning Core development and to the other collections, they reveal subtle mismatches among generation, rendering, targets, and scoring, a reminder that procedural generation alone does not guarantee correctness. The library, generated datasets, and audit material are publicly available.

Long-horizon reasoning requires an agentic runtime that can persist when evidence supports its current approach and pivot when measurements reveal failure, hidden constraints, or a misspecified objective. We present Argus, a persistent, self-evolving runtime in which Manager, Planner, Engineer, and Reviewer execute bounded missions over durable project state. Argus separates stable user intent from operational objectives, constraints, and verification criteria, and admits memories, skills, procedures, verifiers, routing decisions, and rejected routes only after role-owned review and, when available, task-native verification. Model weights remain fixed; self-evolution occurs through persistent runtime state and control policy, with autonomous execution between operator-owned escalation points. Across seven GPT-5.5 benchmark arenas, Argus achieves about 78% on SWE-Bench Pro versus 59% for Direct Copilot while using 1.41 times the aggregate tokens. After verification-gated self-evolution, mature SWE-Bench waves use 21% fewer solve-input tokens and 15% less active workflow time per task than startup waves, while recording 34 verifier recoveries and 22 strict review-loop rescues. Argus also reaches 76.8% on AARRI-Bench and a 28.0-point gap on mathematical data synthesis, with competitive GPU-kernel and language-model-training results. Beyond benchmarks, an optimized RWKV6 kernel was merged upstream; a multi-day mathematics campaign retained falsified routes and proof-backed frontier updates; and six paper pipelines completed 254 missions with 16 stage rollbacks. These results show that a fixed-weight, self-evolving harness can revise, recover, and accumulate verified approaches while producing structured trajectories for future supervised and reinforcement learning.

Context lengths of language models (LMs) have dramatically increased, driven by the demands for in-context learning, self-improvement, and long-horizon agentic workflows. Existing long-context corpora, however, are dominated by books, academic articles, and code repositories, which are finite resources and often scarce in long-distance dependencies. In this work, we introduce OctoLong, a context engineering pipeline that instruments an AST parser, a language server backend, and a package manager to facilitate the recursive retrieval of code references, enabling the curation of dependency-rich code contexts of millions of tokens in length. We then train OctoLong-Instruct, a suite of capable long-context open LMs, derived from base models ranging in size from 600M to 14B parameters, via context-extension mid-training on a ~50B-token mixture containing ~6.2B tokens of OctoLong code contexts, followed by ~10B tokens of instruction tuning. Our training ablations and experimental evaluations against 18 state-of-the-art open-weight long-context LMs show that supplanting just 12% of traditional context-extension corpora with OctoLong data yields substantial gains in long-range retrieval, long-term state tracking, repository-level code understanding, and downstream agentic tasks, while also enhancing API usage in short-context coding scenarios.

Long-horizon reasoning in recent LLMs demands that the model switch between distinct skills inside a reasoning chain, such as first doing a math derivation, then using the result to plan a schedule. We call such problems cross-skill long-horizon tasks: multi-step tasks whose steps require different reasoning skills and depend on earlier outputs. Existing benchmarks often evaluate individual skills, lacking a principled way to measure how well a model switches between skills. We address this gap from both the evaluation and training sides. We introduce Skill Entropy, a measure of the difficulty of switching from one skill to another. We then propose Skill^2-Bench, a benchmark of cross-skill long-horizon tasks built over 558 skills across 9 verifiable and open-ended domains. Each task is assigned a task-level skill-entropy score and grouped into three difficulty levels. Evaluating 8 frontier and 4 open-source models on Skill^2-Bench reveals a skill-switching gap: accuracy decreases on higher-entropy tasks. We then turn skill entropy from a benchmark scale into a training signal. We propose Skill-Entropy RL, an RL framework where the model predicts not only the answer at each step but also the skill used to produce it. The reward combines step-level correctness with a skill-entropy reward that measures the alignment between the model-predicted skill sequence and the gold skill sequence. On Qwen3-4B-Instruct and Qwen3-1.7B, Skill-Entropy RL improves the Skill^2-Bench score from 34.4% to 68.4% and from 14.6% to 40.1%, respectively, outperforming competitive baselines. The same pipeline can be applied to off-the-shelf training data such as OpenR1-Math, indicating that skill entropy is a reusable training signal. Code available at: https://github.com/Gen-Verse/Skill-Entropy-RL

Modern Greek is absent from NVIDIA's Nemotron retrieval models and from major multilingual retrieval benchmarks, despite being important for retrieval-augmented generation (RAG) in legal, energy, financial, and medical applications. We present an end-to-end adaptation of the Nemotron retrieval stack for Modern Greek, including corpus mining, synthetic supervision, retrieval model training, reranker adaptation, reader fine-tuning, and a new benchmark called HERA. Our study shows that a parameter-free BM25 baseline outperforms several off-the-shelf multilingual dense retrieval models on specialist Greek corpora. After fine-tuning on 65,773 Greek retrieval pairs, a Nemotron 1B embedder improves nDCG@10 from 0.362 to 0.835 and substantially outperforms its unadapted counterpart. The learned language competence transfers to general-domain Greek, although the advantage over BM25 remains domain-dependent. We further adapt a cross-encoder reranker and demonstrate consistent improvements across specialist domains. Finally, we LoRA-tune a Nemotron 30B-A3B mixture-of-experts reader for grounded generation, increasing judged answer correctness from 29.4% to 66.9% while significantly improving faithfulness and citation quality. We also introduce HERA, the first large-scale Greek benchmark for retrieval-augmented generation, and release our adapted models and benchmark to support future research on Greek-language RAG systems.

Gradient descent on a factored model $W = UV^\top$ is implicitly biased toward low-rank solutions, while Adam, starting from the same small initialization, is not. We trace the difference to the gauge symmetry of the loss, its invariance under $(U, V) \mapsto (UQ, VQ)$. Gradient flow's low-rank mechanism is available to an optimizer only if that optimizer is gauge-equivariant, a condition necessary for the transfer but not sufficient for low-rank recovery. Gradient descent, momentum, "shared-scalar" Adam, Muon, and Shampoo satisfy it. Adam, RMSProp, and the other coordinate-wise methods do not. A structure theorem characterizes the memoryless equivariant rules as exactly the Gram-determined left preconditioners, and a transfer theorem carries gradient flow's pathwise properties to common-scalar flows. We then sort nine update rules on underdetermined matrix sensing by recovery error against the planted ground truth. A one-parameter family from coordinate-wise to shared-scalar preconditioning restores the bias monotonically, isolating anisotropy as the cause. A "spectral schedule" reconciles two opposing reports about Muon: equal-rate updates recover exactly low-rank targets but lose their edge as the spectral tail grows. In transformers, Adam separates two gauge-equivalent initializations at the first step, where the equivariant optimizers stay at float precision, and ends with the per-head invariants $W_Q^\top W_K$ 56% apart in relative Frobenius distance, a gap no per-head rotation can close. On two hyperspectral datasets at matched training loss, gradient descent cuts held-out error by 43-44% at the lowest sampling density, and at lower effective rank. Basis choice is therefore not a tuning detail but a decision about which interpolant the optimizer selects.

Predicting how a subcortical structure's shape will evolve from a few prior scans could support prognosis and clinical-trial enrichment. Existing longitudinal mesh predictors either extrapolate shape trajectories via high-dimensional embeddings or regress vertex deformations directly. We instead predict the surface's intrinsic geometry in continuous time: a single per-structure graph network predicts the future per-vertex first fundamental form (metric tensor) for an arbitrary causal multiple-visit history and an arbitrary prediction horizon, conditioned on a Fourier encoding of the lead time. The predicted metric is decoded into a surface by a differentiable As-Rigid-As-Possible solver, and the model is trained end-to-end on the rigid-aligned vertex error. Training through the reconstruction keeps the decoded prediction a valid surface and consistently improves it. On 14 subcortical structures from the ADNI dataset, the proposed mesh evolution model (MT-GNN) predicts best among the evaluated methods at every horizon ($-2.29\%$ mean vertex error vs. the temporal mean, $p{=}6.1{\times}10^{-5}$, beating it on 14/14 structures), ahead of geodesic shape regression (DCM, $-0.19\%$) and a mesh transformer (TransforMesh, $-0.45\%$; $p{=}1.2{\times}10^{-4}$), with the lead widening as the horizon grows.

On-Policy Self-Distillation (OPSD) has become a standard post-training approach for improving visual reasoning in multimodal large language models (MLLMs). Existing methods draw privileged information from diverse input sources to guide self-distillation. Yet these designs overlook Modality Imbalance, a challenge inherent to MLLM reasoning. When textual information dominates generation, the model cannot fully integrate its multimodal input. Consequently, carefully designed privileged information remains underused, limiting the effectiveness of OPSD. To examine this limitation, we construct a Positive Teacher with the Zoom-In Image and a Negative Teacher with the Mask Image, which exhibit different degrees of Modality Imbalance. Changes in their reasoning correctness and token logits reveal that Modality Balance can itself serve as privileged information. Motivated by this finding, we introduce OPD-V, a visual OPSD paradigm that instantiates such information through the Positive Teacher and Negative Teacher. Positive Modality-Balance Logits Margins define a Modality-Balance Trust Region that selects the on-policy tokens used for self-distillation. Experiments across 6 benchmarks, 4 MLLM backbones, and 5 post-training methods show that OPD-V consistently improves reasoning performance while reducing training cost.

Achieving local differential privacy in distributed optimization while maintaining low communication cost remains challenging. Existing vector quantization methods, such as vqSGD, use high-dimensional geometric constructions but incur unfavorable dimension-dependent variance. In this work, we propose Subsampled Stochastic TurboQuant (SSTQ), a framework that combines overcomplete equal-norm tight frames, coordinate subsampling, and privacy-aware one-dimensional quantization. SSTQ includes two variants: a Flat Randomized Response version and a Metric-Aware Laplace version, the latter being better suited to higher codebook bit-width regimes. We show that SSTQ achieves optimal mean squared error scaling while using only $\lceil \log_2 N \rceil + b$ bits per client, where $N = Θ(d)$ is the frame size. We also derive a surrogate privacy-aware codebook objective that reduces the codebook-dependent MSE scaling from $O(4^b)$ to $O(2^b)$. Finally, we empirically evaluate SSTQ against established baselines on federated learning tasks using CIFAR-10 and Fashion-MNIST, demonstrating favorable utility and communication efficiency.

Spoken Language Understanding (SLU) is the core component of task-oriented dialogue systems and a pivotal link in achieving seamless human-agent interaction. While traditional SLU can effectively extract user semantics for closed-set tasks after in-domain supervised fine-tuning, it faces significant challenges in leveraging in-context learning for open-domain tasks due to its ambiguous rule definitions. This work proposes Spoken Function Calling (SFC), a novel semantic understanding perspective that optimizes semantic understanding with structured rule definitions, to evolve beyond traditional closed-set SLU. Specifically, we curate and extend a suite of spoken functions based on traditional SLU datasets, construct a multi-agent system to synthesize the SFC-Bench dataset, evaluate the performance of Large Language Models (LLMs) and Large Audio Language Models (LALMs), and enhance the SFC capabilities of LALMs through post-training. Experiments demonstrate that SFC outperforms traditional SLU, substantially enhancing the semantic extraction accuracy for LLMs and LALMs.

Long context reasoning in large language models (LLMs) is usually constrained by the fact that a single inference trajectory has to simultaneously explore the context, store intermediate state, verify evidence, and produce the final answer. This becomes particularly difficult in tasks that require extraction, counting, ordering, or multi-hop reasoning, where an early mistake can propagate until the final response. In this work, we propose Chained Recursive Language Models (Chained RLM), an inference-time architecture, in which the same underlying model is called repeatedly as a sequence of fresh reasoning roots. Each root receives the original problem and context, but does not inherit the full conversational history. Instead, it receives a compact plain-text summary, a plain-text blackboard, and some durable task-specific artifacts written by predecessor roots. The motivation is to manage the context by chopping into partial tasks rather than one large inference response; in each staged computation, intermediate artifacts can be inspected, corrected, and extended by a later fresh inference by the same model. We describe the system model, handoff mechanism, artifact workspace, and evaluation protocol for this system. We study when fresh-context artifact continuation gives a measurable gain in accuracy over direct LLM answering even with recursive tool-calling.

Kinetic model discovery is a central challenge in chemical engineering, as accurate rate expressions are essential for understanding and controlling chemical and biological processes. Symbolic regression (SR) has emerged as a powerful data-driven approach for identifying interpretable kinetic models, but usually operates without domain knowledge, often exploring physicochemically implausible models. Large language models (LLMs) offer a promising avenue for injecting domain expertise into this search. Here, we introduce an LLM-guided SR framework, embedding an LLM module within an iterative SR algorithm for automated kinetic model discovery. The LLM performs two roles at each iteration: (1) a qualitative physicochemical critique of the best SR candidates, and (2) the proposal of new candidate rate expressions guided by the SR-generated models and embedded chemical knowledge. Our framework is evaluated on four in silico case studies of increasing complexity, spanning heterogeneous catalysis and bioprocess systems. Results show the LLM-guided framework reduces iterations to identify the ground-truth model by $41.7-79.3\%$ versus a state-of-the-art SR framework, with the LLM directly proposing the correct model structure in over half of the guided runs. In practical settings, where each iteration typically requires a new wet-lab experiment, this translates into a substantial reduction in experimental effort. Predictive performance on an independent validation set is equivalent between both approaches, with $R^2>0.98$ in all case studies. Ablation studies indicate that both the SR component and the LLM scale contribute to this performance, with a reduced-size LLM largely retaining discovery efficiency. These findings demonstrate that LLMs can effectively inject domain knowledge into scientific model discovery, paving the way toward fully automated, domain-aware kinetic modelling pipelines.

AI-assisted developer tools increasingly mediate programming through chat panels, terminal agents, generated diffs, and streaming status output. These interaction surfaces may create visual accessibility barriers for blind, low-vision, and color-vision-deficient developers, yet little is known about how such barriers are reported in public tool ecosystems. We analyze issues and forum discussions from five AI developer tool ecosystems: GitHub Copilot in VS Code, Cursor, Claude Code, OpenAI Codex, and OpenCode. From 2,652 keyword-retrieved candidates, a three-model ensemble identified 600 unanimously positive visual accessibility reports. A stratified manual sanity check supported this conservative selection. Topic modeling and qualitative analysis identified three recurring categories: screen-reader and assistive-technology barriers; visual presentation, contrast, and differentiation problems; and readability, scaling, and control limitations in AI-specific interfaces. The relative prominence of these concerns varied across ecosystems and reflected differences in editor, terminal, chat, diff, and agent interaction surfaces. An exploratory metadata analysis further identified differences in reporter activity and, across the GitHub-based ecosystems, maintainer participation and closure processes. These findings show that the accessibility record of AI developer tools is shaped by both their interaction design and the reporting and maintenance practices of their surrounding ecosystems.

Can computer vision help make classrooms safer? In this pilot study, we investigate privacy-aware and computationally efficient classroom incident recognition from CCTV-style observations. This setting remains underexplored, with limited benchmarks and few methods designed for the privacy, efficiency, and generalization demands of real-world deployment. We introduce a novel hybrid benchmark combining generative CCTV-style videos with real-world classroom pose data, and propose a lightweight, but robust motion-reasoning framework motivated by the observation that many incidents differ more in motion direction, speed, acceleration, and intensity than in pose alone. To that end, our method first constructs hierarchical kinematic representations of human actions. Our method then distills hierarchical, multi-order kinematic reasoning from a large teacher into a much smaller single-order student, enabling efficient per-person inference while preserving expressive motion understanding. Experiments show that our model outperforms substantially larger baselines at less than one-tenth of their computational cost, while also demonstrating stronger out-of-domain motion reasoning and zero-shot synthetic-to-real generalization. We will publicly release the benchmark, codebase, and supporting tools to facilitate further research in privacy-aware classroom safety.

The Subspace Constrained Mean Shift (SCMS) algorithm is a popular nonparametric method for extracting density ridges, which serve as a low-dimensional representation of high-dimensional data. It is a widely held belief in the literature that SCMS trajectories converge to the classical density ridge, which we call the "static ridge", defined via the density gradient and the eigenvalues and eigenvectors of the density's Hessian. In this paper, we demonstrate that this assumption does not hold in general, as the static definition fails to account for the rotation of the trailing eigenspace along the continuous flow of the algorithm's underlying vector field. To resolve this, we propose a paradigm shift by introducing the "stable ridge", a novel geometric structure defined through the lens of dynamical systems and the Jacobian of the projected density gradient. We prove that this stable ridge is the true theoretical target of the SCMS algorithm. Building upon this foundation, we develop a generalized SCMS framework utilizing a constant step size, establishing its uniform R-linear convergence and topological surjectivity onto the stable ridge. We further derive the rates of convergence for estimating the stable ridge in terms of the Hausdorff distance. Finally, we expose that the original SCMS algorithm suffers from polynomial-time computational complexity, which is caused by implicitly coupling the step size to the smoothing bandwidth via the Mean Shift operator, and demonstrate how our generalized framework provides a statistically consistent and more efficient solution.

In partially observable reinforcement learning, agents face a dual bottleneck: they must explore to encounter rewarding states and retain that experience in memory to optimize their policies. Exploration bonuses and memory architectures are traditionally evaluated in isolation, leaving their interaction unmeasured, and standard notions of sparse reward conflate temporal signal density with what the reward actually supervises. We present a controlled study crossing episodic exploration bonuses with diverse neural memory architectures across three environments that vary how the content of memory is acquired. An identical bonus signal yields three distinct interaction patterns: it amplifies architectural capacity differences where memory content must be actively discovered and retained unsupervised; equalizes architectures to a shared ceiling where the content, once sought out, is a single reward-supervised cue; and is null where the observation stream is purely scheduled. Controlled reward manipulations verify that these patterns track reward structure rather than density: a dense reward neutralizes a bonus only if it directly supervises the required latent memory, and a small avoidable penalty on exploratory actions (leaving the optimum unchanged) induces policy convergence to suboptimal stationary states, which either bonus resolves. We then formalize reward sparsity with observation-anchored reward machines, separating structural sparsity (an automaton reproduces the return without the task-required history) from potential sparsity (the one-step reward misprices local exploratory actions); the resulting vocabulary organizes the three regimes by the retention burden each task exposes. Together, these results show exploration and memory are complements, not substitutes: a bonus induces exposure, and only memory converts exposure into return.

Near-term quantum hardware limits circuit depth and often imposes geometrically local connectivity for quantum generative models, restricting the output distributions accessible to shallow unitary Born models. Introducing stochasticity into a unitary quantum Born model can improve the empirical generative performance of the resulting channel model and, for a restricted small-scale architecture, has been proven to represent a strictly larger family of distributions than its unitary counterpart. However, whether such randomness provides a provable separation at fixed shallow depth for arbitrarily large systems has remained open. Here, we show that shared classical randomness, a comparatively weak resource from entanglement theory, is sufficient to establish such a strict scalable representational separation over the corresponding shallow unitary Born model. More specifically, we augment bounded-connectivity shallow unitary circuits, followed by computational-basis measurements, with spatially separated local Pauli operations, whose joint application is controlled by a single classically sampled random bit. The resulting shallow-depth channel model generates long-range correlations in the classical output distribution that no purely unitary shallow-depth model with bounded connectivity can reproduce. For one-dimensional nearest-neighbour architectures, reproducing such distributions with a purely unitary model can require depth $Ω(N)$ in the worst case. We further show that measurement-based quantum computation (MBQC) provides a natural implementation of the required shared classical randomness through suitable adaptation of the random measurement outcomes. Numerical experiments on MBQC-based generative models support the analytical results.

AI-supported care planning can help clinicians, patients, caregivers, and care teams coordinate complex decisions across clinical, functional, psychosocial, and environmental needs. However, many AI systems present recommendations as fixed outputs, limiting stakeholders' ability to inspect, challenge, and revise plans when they conflict with clinical judgment, patient values, or real-world feasibility. We present CoPlan - a Co-Intelligent and Contestable Interface for Human-AI Care Planning. CoPlan uses a multi-agent workflow in which specialized AI agents generate candidate interventions and supporting or challenging arguments, while human care planners can accept, reject, modify, or add arguments before final plan generation. Through this design, CoPlan combines co-intelligence, in which humans and AI agents contribute complementary expertise, with contestability, where recommendations remain open to inspection, revision, and justification. We demonstrate CoPlan in an aging-in-place care planning scenario. The system supports adaptive care team recruitment, role-based argument review, final care plan generation, and practical follow-up through scheduling agents. This work contributes a contestable care planning interface and a design framing for trustworthy human-AI care planning that preserves human agency and clinical accountability.

Deep neural networks have shown impressive success in NLP tasks owing to their complex structure and huge number of edges. Achieving state-of-the-art performance in natural language processing with a large pre-trained model such as BERT is expensive and time-consuming, carries a large carbon footprint, and is difficult to realize on machines with minimal computational capability. This creates a barrier to training complex models for resource-constrained languages such as Bengali. However, in a complex neural model, not all edges are equally impactful, and the contributions of some of them can be neglected. Pruning promises to reduce the memory footprint of regular networks, shorten the training time of ever-growing networks, and increase inference efficiency without sacrificing comparable performance. In this work, we introduce BnBERT-iPET, a sparse few-shot language modeling approach for Bengali, and experimentally show that a lightweight few-shot-learned language model retaining only 10% of the edges of an initial model such as BERT can perform neck and neck with much larger models on challenging tasks for a resource-constrained language such as Bengali. By learning from few shots through iterative pattern exploiting training and achieving 90% sparsity with the Lottery Ticket Hypothesis pruning technique, our pruned BnBERT-iPET model proves to be a tough competitor to state-of-the-art language models such as Bangla Electra, Indic-BERT, and XLM-RoBERTa on downstream tasks over standard benchmark datasets of the Bengali language.

Data assimilation (DA) uses Bayesian inference to update the state of a numerical forecast model with observed data. In this study, we propose a fundamentally different, unified approach to atmospheric data assimilation. We use latent video flow-matching to sample temporally consistent trajectories from a prior trained using ERA5 reanalysis (69 variables over an 8-day window). We also use posterior sampling to assimilate real observation sources, such as those from the NOAA Integrated Global Radiosonde Archive and the Integrated Surface Database. Because the prior generates a continuous trajectory, it naturally propagates information between observed and unobserved frames. Therefore, we can perform various DA tasks, such as filtering and smoothing, simply by changing the observed frames. Moreover, we generate full-state ensemble forecasts directly from sparse observations, achieving performance competitive with state-of-the-art observation-to-forecast models.

Long-horizon search agents must make multiple sequential actions (steps) to search, retrieve, verify, and integrate evidence to reach a final answer. However, existing methods for training these agents typically treat all steps within a trajectory uniformly during both supervised fine-tuning (SFT) and reinforcement learning (RL), failing to distinguish useful actions from erroneous or redundant ones. In this paper, we propose Answer-Backtracked Credit Assignment (ABC), a fine-grained credit assignment framework for training long-horizon search agents by converting sparse trajectory-level outcomes into dense step-level supervision that rewards useful actions (even in failed trajectories) while suppressing erroneous or redundant actions. Specifically, given a potentially obscure query and its corresponding ground-truth answer, ABC first performs Answer-Backtracked Clue Recovery, which traces back from the answer to recover intermediate clues required to solve the question. It then applies Clue-Anchored Step Scoring to evaluate each search step against these clues, converting sparse binary outcome supervision into dense step-level rewards. Based on these rewards, we develop ABC-SFT, which reweights the loss of each turn, and ABC-GRPO, which uses the step-level scores as rewards in GRPO. Building on this framework, we train ABSeeker based on Qwen3.5-4B with only 8.5k examples. ABSeeker achieves 37.3% on BrowseComp and 39.1% on BrowseComp-ZH. With context management, the scores further improve to 55.3% and 52.9%, respectively, significantly outperforming same-scale (4B) agents and even matching the performance of larger ones (approximately 30B). These results demonstrate the effectiveness of answer-backtracked step-level credit assignment for training long-horizon search agents.

Reasoning about necessity and possibility depends on assumptions about accessibility between worlds and about which objects exist at each one. The same inference may therefore hold under one modal system and fail under another. Evaluating language models on such problems requires testing whether their judgments follow the stated semantics rather than a familiar logic. We construct paired modal problems with identical premises and conjecture but different frame or domain conditions; automated reasoning verifies opposite labels. A balanced core prevents the semantic condition alone from revealing the answer. On this core, four of five recent models perform below the condition-only baseline under direct prompting. Yet enabling reasoning mode raises DeepSeek V4 Flash from 4.4% to 88.1% on unchanged prompts. Following stipulated modal semantics thus depends strongly on inference mode as well as model identity. When frame conditions are omitted, models often agree but fit different familiar logics best. We release the formulas, oracle artifacts, countermodels, and responses.

Agents for long term reasoning require a memory that can be efficiently and effectively updated over time, as new facts and external feedback continue to arrive. Recently, graph memory has been adopted to offer structural organization for multi-hop retrieval and reasoning. However, existing methods store all memories in a flat graph, and accumulated historical memories can introduce irrelevant contexts and increase the cost of evidence selection during retrieval. Moreover, they typically update memory units independently, requiring repeated unit-wise rewrite to cover related changes. To address these issues, we propose HiGram, an evolving hierarchical graph memory framework with path-level localization and rewriting. Specifically, we first propose a hierarchical graph memory, which organizes the memory into coarse-to-fine architecture composed of upper-level nodes and MemoryUnits, thereby reducing the amount of irrelevant information during retrieval. We further propose MicroGraph-based path-level localization, which leverages query and update conditioned MicroGraphs to identify support subgraph and evidence path before rewrite. Finally, we propose a coordinated rewriting method that jointly revises intra-unit memory and inter-unit dependencies, enable valid dependency structures updating in the localized evidence path. Experiments on benchmarks for long-term conversational question answering and conflict-aware memory evaluation demonstrate that our method demonstrate substantial improvements over baselines in answer quality and token efficiency. Besides, our method improves answer accuracy and query-valid evidence selection under dynamic, static, and conditional conflicts.

Muon has recently emerged as a promising alternative to AdamW for language model pretraining by orthogonalizing momentum matrices using Newton-Schulz iterations. Although Muon mitigates gradient anisotropy, it does not explicitly account for the curvature geometry of the loss landscape and may therefore remain sensitive to curvature anisotropy. We bridge this gap by proposing MALT (Muon Augmented by Lightweight Two-sided Preconditioning), which uses lightweight diagonal preconditioners to reduce the sensitivity of Muon to curvature anisotropy. Specifically, MALT uses two-sided diagonal preconditioners with low memory and computational overhead to approximately capture the curvature geometry of the loss landscape. It orthogonalizes the preconditioned momentum using Newton-Schulz iterations and maps the result back to define the update direction, while norm grafting controls the update magnitude. To improve the robustness of MALT to stochastic gradient noise, we further propose MALTER (MALT with Adaptive stEpsize Rescaling). Convergence guarantees are provided for MALT in the stochastic non-convex setting. Experiments on GPT-2 Small, Medium, and Large pretraining show that the proposed methods outperform Muon while maintaining nearly the same memory footprint and wall-clock time.

Language models differ in how safely they behave and these differences are measured by safety benchmarks. But aggregated benchmark scores are hard to trust and interpret, because benchmarks duplicate one another, correlate heavily, and models may sandbag when they detect evaluation. To address these issues, we draw on Item Response Theory (IRT), a statistical toolkit for measuring these latents from performance on items with inferred psychometric properties. We fit IRT models to eight safety benchmarks across 192 language models, the largest psychometric analysis of LLM safety evaluations to date, and contribute three results. First, we find that three interpretable factors of refusal strictness, truthfulness, and contextual harm explain most of the variance between models across benchmarks. Second, psychometrically selected items recover full benchmark scores with lower error than random subsets of the same size, and roughly ten adaptively chosen items suffice for several individual benchmarks, cutting evaluation cost by 97-99%. Third, IRT supports audits of individual models, showing that it can be used to detect naive sandbagging and changes of model behind APIs. Overall, we show IRT is a ready-made toolkit for reading, reducing, and auditing safety benchmarks, which we recommend frontier labs and evaluators adopt.

Systems that automate scientific discovery must repeatedly decide which experiment to run, which hypothesis to test, which tool to build, and when to stop. Many systems make these decisions by maximizing a myopic score such as expected information gain per unit cost or a learned plausibility score. We identify a structural limitation of this approach. Some actions are constructive: they acquire an epistemic capability (an instrument, assay, pipeline, simulator, or abstraction) whose value lies not in the information returned immediately but in the future actions it makes available. When the least-cost route to a confident answer requires a chain of such constructions, a planner that scores actions only by information obtainable within a bounded horizon cannot value the first construction: it yields no information within the horizon and is dominated by any measurement with positive information, however small. We formulate goal-directed discovery as a stochastic shortest-path problem in belief space in which constructive experiments change the downstream action graph, and prove that for every lookahead depth d there is an instance on which every myopic information-maximizing planner has an unbounded approximation ratio, and a related instance on which it never reaches the goal. The mechanism is a capability-indistinguishability lemma: within the horizon, acquiring a capability can be observationally indistinguishable from paying for a null action. This establishes capability gating as a reachability axis of difficulty distinct from curvature (submodularity) and information order (adaptivity gaps). We introduce CG-Plan, an incremental replanner with a capability-aware cost-to-go heuristic h = h_cap + h_exp. In a controlled testbed, the performance gap appears only under gating, persists for every fixed horizon, and arises when near-miss hypotheses come from a data-consistent proposer.

Diffusion policies are a powerful policy class for continuous control, but their iterative denoising process creates a substantial computational bottleneck. Reducing this cost requires adapting the number of denoising steps to the difficulty of each action while preserving task performance. We introduce Prefix-Optimal Generative Policies (POGP), a framework that learns a prefix value function at every intermediate denoising step through a Bellman-style recursion over the denoising chain. The prefix value function serves two purposes: it provides an auxiliary training objective that encourages intermediate outputs to become high-quality actions, and it enables a test-time stopping rule that terminates denoising when additional steps are unlikely to produce meaningful improvement. Across four MuJoCo environments and comparisons with 12 baselines, POGP reduces the required number of denoising iterations by approximately 2.7-fold while retaining near-full task performance. Compared with state-of-the-art dynamic diffusion baselines, prefix training also improves final task performance by approximately 3.5%. These results indicate that supervising intermediate denoising steps is useful not only for adaptive early stopping, but also as an auxiliary objective that improves the learned policy.

Critic-free group-based reinforcement learning has become a scalable approach for post-training large language models. However, most existing methods allocate the same number of rollouts to every task and trajectory state, even though some rollouts provide much more useful learning signals than others. Recent work has started to treat rollout generation as an adaptive decision, but two important limitations remain. First, intervention strategies are often based on fixed heuristics and therefore cannot adjust as the policy changes during training. Second, these methods usually decide only how many rollouts to generate, without explicitly controlling where and how to intervene. To address these limitations, we propose Recoverability-Aware Intervention Learning (RAIL), a training-time framework that learns how to generate rollouts based on the improvement produced by each intervention. RAIL models intervention selection as an online contextual-bandit problem and trains a recoverability controller using intervention traces collected through a shadow-to-live procedure. This allows the controller to keep learning while the underlying policy evolves. We evaluate RAIL in terms of effectiveness, adaptivity, expressiveness, and efficiency. Across multiple settings, RAIL consistently improves performance under limited rollout budgets. These results show that recoverability-aware intervention provides a principled way to generate more informative and less redundant rollouts, leading to stronger learning signals during post-training.

Recent advances in machine learning have enabled training of wireless foundation models, which aim to support tasks such as channel estimation, beam prediction, and localization based on wireless signals. Existing wireless foundation models typically pretrain on channel tensors using masked reconstruction over subcarriers, antennas, or time but ignore the physical characteristics of wireless propagation. In this work, we propose to instead use multipath propagation as the fundamental pretraining object. We present MultiPathFormer, an autoregressive foundation model that represents each transmitter-receiver link as an ordered sequence of continuous-valued path tokens and pretrains with next-path prediction. We introduce an Environmental RAG (retrieval-augmented generation) mechanism and a first-path codebook on top of the transformer backbone, leveraging environment knowledge to improve path statistics estimation like delay and power by up to 59%. MultiPathFormer pretrained on 27 environments transfers to unseen users and, after scenario-specific fine-tuning, outperforms training the corresponding models from scratch in new environments. Across downstream tasks, it outperforms SOTA channel-based foundation models, achieving 5.57 m mean localization error, 0.914 top-3 beam accuracy, 0.994 line-of-sight classification accuracy, and 0.561 channel estimation NMSE. These results show that path-level pretraining can learn reusable representations of wireless propagation.

The spread of misinformation is widely perceived as a threat to democratic deliberation, yet how political elites' rhetorical commitments to truth shift alongside the rise of populist actors remains poorly understood. Analysing 4.5 million tweets and 59,170 parliamentary speeches by German political elites between 2015 and 2025, we measure evidence-based and intuition-based rhetoric using a validated distributed dictionary representation. Across both arenas, intuition-based language has become more prominent, and right-leaning actors consistently exhibit the lowest Evidence Minus Intuition (EMI) scores. The parliamentary entry of the extreme-right Alternative for Germany (AfD) in 2017 coincides with sharp downward shifts in EMI across the broader chamber, while a more gradual decline is observed on Twitter. These findings document an association between far-right visibility and a changing approach to truth in elite discourse in a multiparty European democracy.

Video Anomaly Detection (VAD) is inherently challenging due to the scarcity of anomalies and the large visual variability in surveillance footage, including changes in lighting, viewpoint, and human appearance. To mitigate visual noise and address privacy concerns, recent work has shifted to pose-based VAD, which focuses on motion dynamics rather than raw video data. However, existing pose-based approaches model human behavior in continuous latent spaces, limiting their ability to learn compact motion patterns necessary for robust behavior analysis. We address this by proposing Vector-Quantized Video Anomaly Detection (VQ-VAD), a novel human-centric anomaly detection framework that learns discrete motion representations. VQ-VAD adapts Vector-Quantized GAN (VQ-GAN), originally developed for image generation, to operate on keypoint sequences and construct a motion codebook of normal behavior. Trained exclusively on normal motion sequences, VQ-VAD detects anomalies by identifying high reconstruction errors when an observed motion sequence cannot be mapped to the learned codebook. We conduct extensive experiments across three complementary evaluation settings, including in-domain, cross-domain, and cross-dataset generalization, on four anomaly detection benchmarks. VQ-VAD achieves strong in-domain accuracy (81.83% on HR-SHT [15]), effective cross-domain transfer from CMU Panoptic [14] (76.69% on HR-SHT [15] without retraining), and competitive cross-dataset robustness. The code base for this work is available at https://github.com/TeCSAR-UNCC/VQ-VAD.

Small open-weight language models increasingly run in private, offline, and cost-sensitive settings, where the key deployment question is not only what a model answers but when it should defer to a human. We study whether verbalized confidence can support risk-controlled deferral, evaluating eleven instruction-tuned models from three families, 0.5B to 14B parameters, on ARC-Challenge and TruthfulQA with 25,168 local predictions. Three theoretical results delimit what calibration can provide: strictly monotone calibration preserves the risk-coverage frontier and error-detection AUROC; temperature scaling cannot calibrate models whose confidence stays above one half while accuracy falls below it; and a Clopper-Pearson procedure converts a 200-question calibration set into a finite-sample risk certificate under an i.i.d. deployment assumption. Empirically, eight of 22 model-task pairs hit the temperature-scaling infeasibility floor within one percentage point of the predicted bound. Platt scaling reduces ECE to as low as 0.02, yet certified autonomy at a 20% risk budget is granted to only three model-task pairs and to none at 10%. We also identify and repair an answer-ordering artifact in the multiple-choice form of TruthfulQA. Calibration gives confidence semantics; certified deferral determines when small models are safe to use.

The semiconductor industry is undergoing a dual revolution: the shift toward heterogeneous 2.5D chiplet systems and the integration of Large Language Models (LLMs) into Electronic Design Automation (EDA) flows. While these paradigms offer unprecedented benefits in yield, modularity, design productivity, etc., they radically expand the hardware attack surface. This paper provides a unified analysis of these frontiers, ranging from attacks on chiplet systems (including hardware stacks for LLM acceleration) across architectural, logical, and physical levels, to various exploits against LLM-driven EDA pipelines. To secure chiplet systems, we review a powerful defense approach that leverages 2.5D split manufacturing and active interposers for physically isolated Root of Trust (RoT) architectures. To secure LLM-driven EDA pipelines, we first identify native threats and then review state-of-the-art defense techniques. Finally, we discuss how LLM systems can advance hardware security efforts for modern systems, including chiplets.

Structured input files such as JSON, DOT, OBJ, INI, S-expression, and TinyC are widely used in software systems, but small corruptions can cause parsers to reject otherwise useful data. Repairing such inputs is important because malformed configuration, program, and data files can interrupt testing, analysis, deployment, and downstream automation even when most of the original content remains intact. Existing repair techniques can produce structurally valid inputs, but they often rely on deletion or repeated search, which may lose original content and result in semantic incorrectness. This paper presents RepairFormer, a transformer-based framework for structured input repair. The approach formulates repair as a supervised sequence generation task and uses format tags, oracle validation, and boundary-localized repair to generate valid outputs while preserving content. The boundary workflow focuses generation on the detected fault region, reducing the input size, and supporting repair of longer files. In evaluation, RepairFormer achieves a 88% in repair and 94% in recovery, showing strongest content preservation when repairs are successful. Additional experiments on our benchmark shows RepairFormer repairs 97.57% and recovers 94.29% with 5x faster runtime compared to state of the art.

We investigate the transferability of Earth weather foundation models to planetary atmospheres by adapting the GraphCast graph neural weather forecasting model to Mars. While GraphCast achieves state-of-the-art performance for terrestrial forecasting, its applicability to non-Earth environments remains unexplored. Using the Mars Climate Database (MCD), which provides global atmospheric fields across vertical altitude levels (similar to Earth pressure levels), we evaluate zero-shot and fine-tuned GraphCast predictions of Martian temperature and wind fields. Zero-shot forecasts produce a surprisingly accurate depiction of current conditions but fail to reproduce diurnal variability and rapidly decay toward climatological mean states. To address this limitation, we fine-tune GraphCast using MCD variables and top-of-atmosphere solar radiation forcing while holding humidity constant. Fine-tuning enables rapid learning of Martian thermal variability. Within as few as 10 training epochs, the model begins to capture the diurnal cycle and forecasts up to 10 days reproduce seasonal and vertical temperature structure. Prediction quality improves with training sample size and exhibits sensitivity to seasonal initialization. These results demonstrate that Earth-trained AI weather models can be adapted to simulate Martian atmospheric dynamics, providing a pathway toward rapid planetary weather prediction to support mission operations, dust storm risk mitigation, and future human exploration.

Against the backdrop of violence in police interactions with the U.S. public, we explore how deferentially police officers speak to virtual characters depicted as Black adult males in vir- tual reality (VR) simulations. We evaluate the effect of seeing and communicating with these characters through a causal in- ference lens, where the assignment of the Black man character to a police officer and simulation is the treatment variable. Our (marginal) average treatment effect AT E measures the social impact of the character on the deference of officer statements with each turn of the conversation. Soberingly, we find that most officers speak less deferentially to Black man characters, except for White, biracial, and multiracial female officers, es- pecially in settings where the VR character was known to be a suspect. Across a full conversation of a typical VR scene, these marginal AT Es can result in notable changes in def- erence of tone (two to several points difference on a scale of 0-10), above and beyond that due to the initial effect of per- ceiving a Black male character. Even more disconcerting is that this can contribute to conversation breakdowns that po- tentially result in violence or danger to both the public and the police. We also explored the capabilities of large language models (LLMs) for ATE estimation. From our methods com- parison analysis, including model validation against synthetic data, we provide unique scientific insights on LLM-assisted methodologies for ATE estimation. As such, for ATE esti- mation with multilevel data with text, we recommend mixed effects models with the inverse propensity treatment weighted (iptw) approach, which utilized an LLM for text feature cre- ation. While we also tested LLMs for finetuning prediction models ultimately for ATE estimation, we conclude they are an area for further development and refinement.

Released aligned large language models remain vulnerable to malicious downstream finetuning. Existing defenses are largely designed for the fine-tuning-as-a-service (FTaaS) paradigm or rely on downstream users to follow additional safety procedures, and therefore do not directly address the setting we study: a provider controlled partially protected open-weight (PPOW) release setting in which most weights remain trainable while a small safety-critical component is preserved at release. We propose a Unidirectional Safety Gate (USG), instantiated as a Null Space Cubic Layer together with an Inverse Adapter inserted after the final Transformer layer. During downstream fine-tuning, the cubic layer suppresses or blocks gradients from harmful samples whose hidden states fall in a calibrated protected region, while the Inverse Adapter restores the base model's forward behavior. In practice, we calibrate a threshold using defender-held harmful data, allowing protection to generalize to nearby in-distribution harmful samples. Across six evaluated model-dataset settings, USG keeps post-finetuning attack success rate close to the pre-release level under a fixed release threshold, while maintaining high safe-pass rates on easier settings and exhibiting a clearer safety-utility trade-off on unsafe samples from BeaverTails. These results suggest that release-time representation-space blocking can raise the cost of malicious downstream adaptation without requiring downstream cooperation. The code is available at https://github.com/OpenCausaLab/Gradient-Immunity.

Understanding the effectiveness of quantum compilation techniques requires visibility into the entire transpilation process, not just the final circuit metrics. This demonstration presents an MLflow-inspired autologging framework for Qiskit that automatically captures compiler provenance, including transpilation stages, pass-level execution data, backend characteristics, compiler configuration, and execution results. The framework extends the QProv provenance model with compiler-specific information and stores the collected data in an MLflow Tracking Server for analysis and visualization. By eliminating manual instrumentation, the proposed approach improves compiler observability and supports reproducible evaluation of quantum compilation workflows.

Sparse matrix kernels are fundamental to scientific computing, graph analytics, and machine learning. Their GPU performance depends strongly on the input sparsity pattern and execution strategy. For the same SpMM on the same matrix, cuSPARSE exhibits a 350x performance gap between CSR and Blocked-ELL. Our study of multiple data formats, specialized systems, and sparse compilers shows that no single implementation consistently dominates across sparsity patterns and operators. This motivates a system that can adapt its representation, execution strategy, and hardware mapping to each workload and target GPU. We present SparseDitto, an LLM-based system that constructs a GPU kernel for each matrix, operator, and target GPU. SparseDitto supports SpMV, SpMM, and SpGEMM within a unified design framework. A lightweight additive model ranks established strategies using structural features of the input matrix. An architecture-aware planner then proposes several candidate designs. Coding and verification agents implement and refine them using measurements from the target GPU. Across three sparse operators and a diverse set of matrices, SparseDitto achieves a geometric-mean speedup of 2.68x over cuSPARSE on an NVIDIA RTX PRO 6000 GPU, with a maximum of 146.61x. On an NVIDIA H200 GPU, it achieves 2.79x, with a maximum of 78.5x. Its generated SpMM kernels also accelerate full-batch GCN training by up to 3.39x.

Football score forecasting combines a strong statistical core with a difficult contextual edge. Dynamic Poisson-family models estimate team strength, expected goals, and coherent score probabilities, but do not directly understand roles, tactical matchups, motivation, or how a first goal changes behaviour. Large language models (LLMs) can reason about such concepts, yet are not calibrated probability engines. We combine both components through an auditable information harness. This paper documents four iterations: V1, a dynamic score-driven Dixon-Coles baseline; V2, which maps LLM contextual ratings back into expected-goal parameters; V3, which replaces scalar correction with goal-by-goal simulations over a frozen score-candidate set; and V4, which adds shared first-breakthrough and post-goal cascade judgments, time-aware stopping, and deterministic tail candidates. The harness defines input semantics, supplies pre-match evidence, and constrains the LLM to an inspectable reasoning route. On a chronological replay of the first 150 matches of the 2025-26 English Premier League, V1 achieved 10.0% Top-1 and 26.7% Top-3 exact-score accuracy. V3 reached 12.0% and 30.0%, while V4 reached 14.7% and 30.7%. V4 increased candidate coverage from 77.3% to 84.7%, although no added tail candidate became a Top-3 exact hit. V1's native 1X2 distribution achieved 53.3% argmax accuracy, 0.9878 log loss, 0.5870 Brier score, and 0.2095 ranked probability score. These results are exploratory: the development slice is not an untouched benchmark, and temporal input isolation cannot exclude outcome memory in a closed LLM. The contribution is an auditable hybrid architecture, a clear design evolution, and negative findings showing where football-aware simulation does and does not improve score selection.

A central question in language acquisition is whether linguistic biases can emerge from general learning mechanisms operating over underdetermined input. Artificial Language Learning (ALL) studies have shown that human learners reliably generalize beyond the evidence provided, including by preferring scope-homomorphic noun phrase modifier orders. In this work, we investigate whether language models exhibit the same bias under similar conditions. We create a controlled learning environment in which models are trained on a corpus where all noun phrases containing multiple modifiers have been removed, eliminating direct evidence about modifier ordering, and are then evaluated on multiple modifier sentences. Across three model sizes, we find that they consistently prefer scope-homomorphic orders despite never observing them during training. These preferences vary in strength by modifier type. To investigate the source of these preferences, we examine noun-modifier association strength using pointwise mutual information (PMI). While PMI reflects known modifier-ordering patterns, it does not explain the models' ordering preferences. These findings demonstrate that LMs can recover human-like linguistic generalizations from impoverished input and provide a controlled framework for investigating the mechanisms underlying such biases.

High-quality annotation of artworks is essential for computational art research, yet extracting implicit semantics remains challenging due to the reliance on culturally grounded meanings and deep contextual knowledge behind the images. Current AI-assisted annotation tools often lack assistance or rely on one-way workflows where experts have to perform extra manual calibrations to improve AI models, resulting in limited efficiency. To address this, we propose Bidirectional Human-AI Augmentation(BiHAA), a closed-loop framework in which skills and domain knowledge base evolve through real-time interaction and bidirectional HAI augmentation. Informed by a formative study with 20 artwork annotators from different backgrounds, we implement this framework in ArtAnno, an artwork annotation system driven by a multi-agent architecture. The system includes a Proactive Agentic Support Module, where AI augments humans through semantic mining and label suggestion, and an Interaction-Driven Evolution Module, where human expertise continuously enhances the AI through distilling annotation trajectories into reusable experience. Evaluation through a user study and two case studies demonstrates that our framework and system improve annotation efficiency, enable knowledge accumulation, and reduce the effort of information seeking and verification for annotators with limited domain expertise. We conclude by discussing broader implications and future directions.

Joint Energy-Based Models (JEM) unify classification and generation within a single network and support out-of-distribution (OOD) detection. Canonical JEM training relies on stochastic gradient Langevin dynamics (SGLD); a theoretically motivated alternative, the Predictor-Corrector (PC) sampler, has not previously undergone a systematic replication test on the canonical model. We reproduce canonical JEM on WideResNet-28-10 without normalisation layers on two independent runs and test whether PC retains its theoretical advantage without an annealed noise schedule, across three protocols: PC replacing SGLD throughout the roughly 130 training epochs; cold-start generation (FID); and refinement-style multi-OOD detection (AUROC). The reconstruction reaches 92.88% test accuracy and buffer-FID 44.46 (canonical: 92.9% and 38.40). We document two failure modes: catastrophic late-training divergence via the canonical outlier-buffer mechanism (both SGLD runs and, with the same signature, both PC runs), and run-dependent SVHN OOD-discrimination dynamics. No method-level advantage of PC over SGLD is observed on any protocol: at inference the absolute AUROC difference stays below 0.007 across all ten checkpoint-OOD pairs and the FID difference below 0.5; on the training protocol a hierarchical seed-by-image bootstrap gives a 95% confidence interval on the macro-averaged AUROC difference that contains zero, while a seed-level equivalence test with two runs per method cannot establish formal equivalence. The data are consistent both with equivalence and with a small directional effect. This practical indistinguishability is theoretically expected: under fixed noise the PC predictor step degenerates by construction, so its guarantees do not transfer to canonical JEM.

Short-term load forecasting (STLF) play a vital role in the electric power industry. It serves infrastructure that European and German law designate as critical. Determinism, reproducibility, and auditability are engineering requirements rather than optional extras. STLF is no longer purely an accuracy problem. It is also a software-engineering and compliance problem. This paper describes results from a 41-day live challenge that evaluated a complete STLF pipeline for the aggregated German transmission-grid load. The pipeline is based on the open-source Python library spotforecast2-safe, which implements the EU-AI Act Requirements in Safety-Critical Environments by design. The pipeline predicts the 24 hourly load values of a target day from European Network of Transmission System Operators for Electricity (ENTSO-E) data. It includes anomaly detection and gap-aware data preparation, calendar and weather covariates, a recursive multi-step forecasting algorithm, and hyperparameter tuning. Forecast accuracy is measured against the official ENTSO-E day-ahead forecast. The EU-AI act compliant spotforecast2-safe pipeline beats the ENTSO-E baseline. In-context models show competitive performance. Transparent, low-cost, and auditable local models (referred to as macl2l in this paper) are competitive with more than 100-million-parameter large, energy-intensive pre-trained foundation models such as chronos-2. The challenge infrastructure, the complete submission history of all teams, and the frozen final leaderboard are publicly available.

Predicting the existence and type of links (edges) between nodes in a multi-relational graph is key for applications from social interaction prediction to knowledge relationship identification. Enhancing local features with relevant global information is crucial for accurate link prediction, yet it remains challenging. We address this by modeling the relationship between node pairs as node influence. That is, whether the node influence can be propagated and what type of influence is propagated indicates where and what type the edge is, which will be the most relevant local and global information to predict the edges. To this end, we extend the Susceptible-Infectious-Recovered (SIR) epidemic model to capture the influence propagation of nodes on a large scale through sub-graph structures. Subsequently, these sub-graphs are compressed using virtual edges, thereby substantially reducing the computation associated with utilizing the global graph structure. Finally, we propose the Influential Graph Neural Predictor, referred to as IGNP, a link prediction framework guided by influence propagation. Extensive experiments demonstrate the superiority of the proposed method, which outperforms strong baselines by a large margin on the widely used and real-world datasets.

Representation theorems in decision theory establish that behavior satisfies certain axioms if and only if it can be rationalized by a well-defined objective. I argue that this ``if and only if'' structure provides a potentially useful foundation for label-free evaluation and regularization of LLMs and other AI systems. Axiom compliance can be checked from the model's own responses to synthetic choice problems, with no external labels or human feedback, and the penalties are readily computable. Because the axioms are necessary and sufficient, the resulting checks exhaust the implications of the relevant rationality standard for the elicited data: a model that passes cannot be rejected on rationality grounds by any further test of the same data. I discuss three instantiations: probabilistic coherence via a theorem of de Finetti, preference rationality via Afriat's theorem, and subjective expected utility via a theorem of Echenique and Saito (2015), each yielding a continuous penalty that is zero whenever behavior can be rationalized. Since coherence does not restrict which objective rationalizes behavior, these penalties complement rather than replace other evaluation and training signals.

Metamodels are instrumental in reducing the computational burden associated with nested reliability analyses and optimization loops in Performance-Based Risk Optimization (PBRO) of structures under stochastic loads. In this context, stochastic emulators are particularly useful because they approximate response distributions while accounting for the intrinsic stochasticity of the simulator. Among these methods, Stochastic Polynomial Chaos Expansion (SPCE) is especially attractive because it does not require replications of nonlinear analyses at fixed input conditions. However, SPCE may present limitations in accurately representing extreme responses in the tails of structural response distributions. To address this limitation, this study proposes a framework that combines Generalized Stratified Sampling (GSS) with SPCE. The GSS scheme partitions the input space into strata according to the intensity of the hazard, improving the representation of extreme responses, while independent SPCE emulators are trained within each stratum. The conditional exceedance probabilities estimated in each stratum are then recombined using the total probability theorem to evaluate the probabilistic constraints. The proposed GSS-SPCE framework is applied to the optimal design of buckling-restrained brace cross-sectional areas in a two-story steel building. The objective is to minimize the initial construction cost while satisfying prescribed probabilistic performance constraints. Results show that the proposed framework accurately estimates structural response distributions, including their tail regions, while substantially reducing the number of nonlinear model evaluations required for PBRO.

Mental health professionals have raised concerns about risks of psychological harm from interaction with large language models (LLMs), including "delusional spirals" in which concerning human and LLM behaviors reinforce each other over time. With growing public use of LLM-powered chatbots, there is an urgent need to build evaluations grounded in real-world episodes of psychological harm experienced by users. We developed DelusionEval, an evaluation protocol that tests a model's tendencies to exhibit behaviors linked to promoting user delusions. We prompt each model with 589 unique conversation histories from 18 participants, comprising 12,591 messages from users who experienced delusions and psychological harm. We find that the tendency of an evaluated LLM to exhibit delusion-linked behavior does not reliably correlate with model size, release date, or the presence of test-time reasoning. However, extending the context of prior messages substantially increases rates of delusion-linked behaviors, providing evidence for the importance of context in LLM safety evaluation. For example, the rate of failing to discourage self-harm when the user expresses suicidal ideation increases from 30.0% to 41.1% when an additional 350 messages are prepended to the conversation history. All model families (e.g., GPT, Claude) exhibit substantial rates of delusion-linked behaviors. Within families, later, larger, or higher-reasoning models are not uniformly better across all behavior categories. Our results raise concerns regarding the potential psychological impact of LLMs and the need for more rigorous studies of real-world human-AI interaction.

Vision offers a critical axis for advancing foundation models, driving a shift towards natively unified multimodal pretraining. Despite this momentum, the design space and the fundamental mechanisms of how modalities interact during unified training remain underexplored. We provide empirical clarity through a systematic exploration of multimodal pretraining. Our controlled experiments on both synthetic and large-scale real-world datasets yield four key insights into the physics of multimodal pretraining: (i) Knowledge Flow: We disentangle how language, visual understanding, and visual generation transfer knowledge across modalities, revealing distinct patterns of influence and asymmetry; (ii) Synergy vs. Competition: We show that data "complexity" largely determines whether modalities are synergistic, identify architectural choices that promote synergy: such as shared attention and normalization with modality-specific feed-forward layers, and find that these behaviors generalize across different visual tokenizer designs; (iii) Early Unification: Unifying modalities from the very early stages and training them jointly is shown to be more effective than late alignment or sequential training. This process uncovers a vision laziness phenomenon, where delayed integration leads models to rely on language priors; (iv) Recipes: We derive efficient pretraining recipes that achieve strong generative performance using only 5% of the compute budget. These core findings are subsequently validated at scale by training multiple 13.5B MoE models on 2T tokens. We hope this study provides a principled foundation for understanding and scaling multimodal pretraining.

Analog circuit design automation using reinforcement learning (RL) has emerged as a promising approach for reducing manual effort. However, many existing RL-based methods focus on single-objective optimization. Even methods designed for multi-objective (MO) problems often reduce multiple design specifications to a single scalar reward. This simplification limits the ability to capture the true Pareto trade-off among competing objectives and often leads to suboptimal designs. Moreover, requiring the model to be retrained from scratch whenever the desired MO specifications change remains a key limitation. To address these challenges, we present ORACLE, an open-source RL-based framework for MO analog circuit design optimization that replaces scalar reward optimization with vector-valued learning and preference-aware conditioning. ORACLE represents a true MO analog circuit design optimizer that uses a preference vector to specify the relative weights of multiple objectives, enabling a single trained model to generate designs across diverse trade-off settings without retraining. We further propose two preference-guidance strategies, namely normalized-weight guidance and cosine-aligned guidance, to improve convergence. In addition, we incorporate a large language model (LLM)-guided action selection mechanism to filter actions that are likely to lead to suboptimal designs or increased runtime. Our results show that, on multiple circuit topologies with 2,000 test cases, ORACLE reduces runtime by 20.4x - 104.4x compared to state-of-the-art approaches. It also meets 99.9% of the 2,000 target specifications, and achieves 5.1x - 318.6x better figure of merit in the resulting output specs.

We show that tiny transformers can profitably employ a simple form of Chain of Thought, which we call protoreasoning, allowing us to study step-by-step reasoning on ~1M-parameter models and opening up opportunities for much more detailed experimentation and analysis than is feasible for larger models. Current Large Language Models exhibit impressive step-by-step reasoning, but we have yet to understand its generality, i.e., when and how LLMs learn genuinely general algorithms rather than "bags of heuristics." Such questions are hard to settle on compute-intensive frontier models trained on opaque data. To work at model scales far below the threshold for natural-language competence, we define reasoning-friendly tasks on Dyck languages (sentences of correctly nested brackets). We find that protoreasoning traces substantially close the out-of-distribution generalization gap, and ablations confirm that the trace's content, not merely its extra tokens, drives the gain.

SciCode is the standard measure of the scientific-coding ability of language models: research-level problems that demand both frontier scientific theory and its implementation as working numerical code. It is a component of the Artificial Analysis Intelligence Index and a standing evaluation in government and national-laboratory suites. Yet its scores have recently plateaued: the strongest 2026 models cluster tightly around 60\% subproblem accuracy, and a successor model ties its predecessor. We trace this stagnation to defects in the benchmark itself. A per-problem, domain-expert audit of all 65 test problems uncovers 263 defects; 192 of them, spread across 91\% of the main problems, cause correct, instruction-following solutions to be wrongly rejected---through non-reproducible gold answers, over-tight tolerances, or self-contradictory specifications. Critically, 78\% of these score-suppressing defects require specialized physics or mathematics knowledge to detect, not mere clerical proofreading. We corrected every confirmable defect to produce SciCode-Verified. The corrections add only the specifications a well-posed problem requires, repair grading, and tighten the tests that were too lenient; every change is recorded with its justification and independently re-checked by a second domain expert. We re-evaluate twelve frontier model snapshots on the corrected benchmark and find a substantial recovery: subproblem accuracy rises from 45--60\% to 84--98\%, and main-problem accuracy from 9--27\% to 69--92\%. State-of-the-art models are far more proficient in scientific coding than SciCode has suggested---the bottleneck was not model capability, but the quality of the evaluation instrument. We release SciCode-Verified with its complete audit trail as the corrected public standard.

The capabilities of an LLM agent depend not only on its model but on the harness: the executable program that constructs context, invokes tools, verifies results, and recovers from failure. Recent work shows that evolving the harness yields persistent improvements without updating model weights. Existing approaches, however, assume that all execution experience can be routed to a single optimizer, which evolves one harness along a sequential trajectory. Real agent ecosystems violate that assumption: users, organizations, and environments generate isolated streams of experience that cannot be pooled, so the experience most worth learning from is exactly the experience that cannot be directly centralized. We introduce EvolveNet, a paradigm of collaborative harness evolution that moves experience extraction to the data. A shared harness is broadcast to data-local agent deployments, each of which evolves it on its own workload. Only the resulting program adaptations are composed into an updated shared harness and redistributed, so that every participating agent inherits operational experience discovered by the others. By shifting the aggregation boundary from raw workloads to learned adaptations, EvolveNet keeps workloads local and allows multiple evolutionary searches to proceed concurrently with reduced serial depth. Because independently modified programs cannot be averaged like model parameters and may conflict when composed, EvolveNet introduces scope-typed, evidence-guided program aggregation. Across five settings spanning text-to-SQL, data-science coding, competitive programming, software engineering, and agentic workflows, EvolveNet improves the shared harness in all five, with the largest gains under heterogeneous workloads, and ablations attribute the improvement to composition of adaptations from different agents rather than to selecting among them.

Interactive video world models are essential for long-horizon planning and exploration, yet they suffer from compounding errors. Post-training methods such as reinforcement learning (RL) can improve these models, but they hit a verification bottleneck: for arbitrary action sequences, no ground-truth future state exists to measure long-term drift. Our key insight is that reversible action cycles make this verification possible: a sequence composed with its inverse must analytically return to the initial state, yielding annotation-free supervision on long-horizon correctness. Building on this, we introduce WorldCycle, a self-verifiable RL framework that constructs closed action cycles and their repeated executions from ordinary action sequences, and optimizes two complementary rewards: a spatial closure reward enforcing symmetry between mirrored forward and reverse segments, and a temporal consistency reward aligning states across repeated cycle executions. These rewards force the model to learn actions as consistent state operators rather than memorized temporal patterns, and extend naturally to out-of-distribution composite action cycles that the base model handles poorly. We further release CycleBench, a diagnostic benchmark for state-returning ability under complex action structures. WorldCycle reduces state returning drift by up to 44% and lifts composite-action accuracy nearly 4x over the base model, providing a vital foundation for physically grounded world models.

Reinforcement learning (RL) post-training improves the reasoning capabilities of large language models, but autoregressive rollout generation remains a major efficiency bottleneck. Speculative decoding can accelerate generation, yet applying it during RL is difficult because the target policy continually evolves: static proposers become stale, while frequent drafter updates add substantial overhead. We introduce SpecRoll, a speculative rollout engine that preserves the target model's sampling distribution while adapting at two timescales. Lightweight future-token heads generate parallel proposals, while our proposed Reflex module uses delayed verifier feedback to perform bounded, trajectory-local hidden-state corrections without backpropagation. A complementary slow path updates the head parameters only when sustained degradation is detected. SpecRoll combines these mechanisms with concurrency-aware sparse-tree verification and exact target verification, leaving the target rollout distribution and GRPO objective unchanged. Across five models ranging from 1.5B to 14B and three mathematical reasoning datasets, SpecRoll achieves 1.26-2.15x generation speedup and 1.21-2.04x end-to-end speedup over vanilla GRPO. It also outperforms FastGRPO in both generation and end-to-end time across all 15 matched settings, with an average pairwise end-to-end gain of 1.18x. Controlled ablations show that the fast and slow adaptation paths provide complementary benefits. Our source code is available at https://anonymous.4open.science/r/SpecRoll-26062006.

Unified Multimodal Relation Extraction (UMRE) aims to identify intra-modal and cross-modal relations between textual entities and visual objects. However, existing UMRE studies still encounter two critical issues: ignoring inherent aleatoric uncertainty causes noise propagation, and deep-seated heterogeneity between distinct modal distributions hinders alignment. To address these issues, we propose the Uncertainty-Guided UMRE Network (UG-UMRE). Specifically, we design an Uncertainty-Driven Unimodal Augmentation (UDUA) module, which models features as Gaussian distributions based on the Variational Information Bottleneck. By incorporating an uncertainty-aware self-supervised contrastive learning mechanism, UDUA effectively filters out noise while maintaining semantic consistency. Furthermore, we introduce the Joint Aleatoric Uncertainty Alignment (JAUA) module as a global semantic pre-calibration mechanism. JAUA leverages probabilistic distribution consistency to construct a shared latent space, eliminating the distributional gap by synchronizing cross-modal statistical properties, thereby laying a robust foundation for fine-grained interaction. Experiments on three benchmark datasets (UMRE, MORE, and MNRE) demonstrate that UG-UMRE achieves state-of-the-art performance. Further analysis validates the pluggable and effective performance of the proposed UDUA and JAUA modules.

The emergence of the ISO standard GQL introduces a powerful query language extending first-order logic with controlled recursion, raising the question of its applicability to evaluation of ontology-mediated queries (OMQs). We focus on OMQs consisting of atomic queries over ontologies expressed in Horn-ALCHI, an expressive Description Logic that is not, in general, first-order rewritable. To address this, we introduce DL automata, a novel formalism that captures the semantics of such OMQs via runs over fact sets. We then identify a large class of DL automata that can be rewritten into unions of conjunctive two-way regular path queries (UC2RPQs), a central fragment of GQL. Our class of automata relies on a stratification of their states, ruling out specific forms of cyclic dependencies known to raise the complexity. This yields a broad class of Horn-ALCHI OMQs that are GQL-rewritable.

We propose a deep learning framework for image restoration from images degraded by both multiplicative Gamma noise and blur. Unlike conventional deep equilibrium (DEQ) models that rely on implicit neural regularization, the proposed method learns an explicit and interpretable regularizer parameterized by geometric priors associated with surface area and mean curvature. To minimize the resulting variational model, we develop a mirror descent algorithm tailored to the commonly used Gamma-noise fidelity terms. Leveraging the Kurdyka-Lojasiewicz property for functions defined in $o$-minimal structures, we establish the global convergence of the generated iterates to a critical point. Experimental results on both grayscale and color image restoration demonstrate that the proposed method consistently outperforms representative model-based approaches while achieving performance comparable to state-of-the-art DEQ models based on implicit regularization, despite requiring substantially fewer trainable parameters.

CheMLFlow is an open-source platform for building and executing end-to-end, high-throughput, and agentic workflows for scientific and technological applications. CheMLFlow targets a common bottleneck in scientific machine learning development, where researchers often need to assemble data acquisition, curation, representation, model training, validation, screening, interpretation, and reporting into a reproducible pipeline, even when their primary research contribution concerns only one stage. CheMLFlow provides modular workflow components, ready-to-run reference pipelines, standardized artifacts, and evaluation outputs that reduce orchestration overhead and support benchmarking across methods and datasets. The platform is designed to be extensible, reproducible, and automation friendly, with pluggable representations and models, deterministic splits, explicit run artifacts, batch execution, and report generation. As scientific software increasingly moves toward agent assisted experimentation, CheMLFlow's configuration driven workflows and structured outputs also provide a practical interface for coding agents to help users construct experiments, inspect results, and summarize findings under human supervision. This article describes the system architecture, core workflows, and benchmarks that reach literature performance for quantum mechanical, physicochemical and bioactivity property prediction, and use cases involving time series datasets demonstrating applications beyond molecular chemistry datasets.

Social media videos often communicate meanings that go beyond their visible actions, captions, or speech. A mundane clip may become humorous, ironic, or satire only through the interaction of multimodal cues and cultural context, making such content a difficult test case for video-language models. In this paper, we introduce \textit{DrivelHub+}, a benchmark for evaluating whether models can infer the implicit, non-linear, and rhetorically layered meanings of social media videos that appear nonsensical on the surface but convey deliberate pragmatic meanings. DrivelHub+ consists of 1,000 videos collected from social media, each annotated with a human-written implicit narrative explanation. Unlike conventional video understanding tasks focused on recognition or description, we present a benchmark that targets contextual multimodal reasoning. We evaluate current video-language models from two perspectives: explanation, where models must explain the pragmatic comprehension of a video in natural language; and representation, where we adapt reasoning-as-retrieval to test whether model representations align videos with their corresponding implicit narratives in both video-to-text and text-to-video retrieval. Our benchmark provides a diagnostic setting for measuring the gap between multimodal perception and pragmatic comprehension, asking whether current models can move beyond describing what is shown to inferring what is meant.

Training LLM agents commonly relies on supervised fine-tuning from expert trajectories or online reinforcement learning over human-specified tasks with handcrafted verifiers. Though effective, both remain bottlenecked by externally specified tasks and supervision signals, limiting the scalability and diversity of agent training. We study an environment learning paradigm in which agents acquire interaction and manipulation capabilities solely through environment interaction, without externally specified tasks. We propose State2State, an environment-derived mid-training method that converts explored environment states into training objectives, challenging agents to reach a specified target state. By deriving tasks from environment exploration and verifying success through rule-based state matching, State2State provides scalable and verifiable training objectives without expert supervision or manual task design. Experiments on ALFWorld and ScienceWorld show that State2State improves agent performance as a standalone environment-learning stage in most settings. As initialization for downstream RL, it further improves final performance and learning efficiency, with promising evidence of cross-environment generalization.

Bayesian causal discovery seeks to determine the posterior distribution of causal theories, which are interpreted as directed acyclic graphs (DAGs) that explain the observed data. The resulting posterior allows systematic reasoning regarding epistemic uncertainty within these theories. Nonetheless, finding such graphs is difficult due to identifiability problems and limited observational data. Furthermore, precisely approximating posterior over graphs is challenging given vast range of potential DAGs. Recent Bayesian approaches have addressed some of these challenges, yet they remain limited as they fail to encode dependencies between edges, and lack principled ways to incorporate domain knowledge as inductive biases during the search process. To overcome these limitations, we propose SVI-DAG, a structured variational inference approach to Bayesian causal discovery using observational data and prior beliefs that uses normalizing flows to model dependencies between edges, supporting expressive and multimodal posterior learning over DAGs. To mitigate mode seeking behaviour in evidence lower bound optimization and promote mode coverage, we use stein variational gradient descent to update the node potentials using a kernel in acyclicity space. We evaluate SVI-DAG against 5 state-of-the-art Bayesian DAG learning methods and demonstrate superior performance in uncertainty quantification while remaining competitive in terms of structural accuracy.

Chain-of-thought (CoT) reasoning offers a window into the decision-making of large language models (LLMs), which can be monitored for target behaviors by reading the reasoning trace, motivating work on CoT monitorability. Latent CoT approaches, however, replace the explicit tokens with a small number of continuous states, lowering inference costs but removing the readable trace this monitoring relies on. Monitoring then requires alternative access to the model, such as probing its activations or verbalizing the latent states back into text, but how much monitorability these alternatives preserve is unclear. We study this question with a hint-based intervention setup, a proxy for behaviors where models exploit biasing input cues, e.g., an inadvertently leaked answer or a belief stated by the user, without acknowledging them. Taking hint-reliance as the monitorability target, we compare monitors across reasoning modes, from explicit CoT to weakly- and strongly-supervised latent CoT, on math reasoning and question answering. We find that, in this setup, monitorability depends more on properties of the task (such as whether the correct answer constrains the supporting reasoning) and the level of access to model internals than on the reasoning mode.

In gradual adaptation, how should the training time on each task change as the number of intermediate tasks increases? We study this question for overparameterized linear regression tasks that change smoothly and share a zero-loss solution. With $N$ tasks and training time $s_N$ on each, the final learning progress converges to a continuum curve when $Ns_N\toτ$. The limiting progress is $Θ(τ)$ for small $τ$ and $Θ(τ^{-1})$ for large $τ$, so both very short and very long training produce little progress. It follows that optimal per-task training times scale as $s_N^\star=Θ(N^{-1})$, equivalently $Ns_N^\star=Θ(1)$. Experiments on gradually rotated MNIST and a natural Yearbook time shift are consistent with less per-task training as the path is divided more finely.

As chart images, tabular data, and visualization code play increasingly important roles across diverse domains, cross-representation understanding across these modalities poses fundamental challenges for AI systems: the relationships across representations are inherently \textit{one-to-many}, supervision is ambiguous and costly, and model optimization lacks a principled signal that is both direction-adaptive and representation-generalizable beyond task-specific objectives. We introduce CoCoEvolve to improve consistency across chart, table, and code representations. Instead of treating cross-representation mapping as a one-to-many problem, we define explicit one-to-one correspondences and optimize models using agreement between representations, without additional annotations. During training, CoCoEvolve@Train performs co-evolution across the chart-table-code cycle, while CoCoEvolve@Test applies the same consistency objective at inference time for test-time co-optimization. We also present CoCoEvolve@Eval, an evaluation suite covering all six cross-representation tasks. Across four benchmarks, CoCoEvolve improves performance in both training-time and test-time settings. Our project page: https://xhguo7.github.io/CoCoEvolve/.

As AI systems become increasingly integrated into diverse interfaces and applications, model-centric audits are insufficient to address risks arising from interactions among system components and deployment environments. System integration has long been central to software audits in safety-critical domains such as aerospace. However, its role in AI auditing remains underexplored. Scanning through 4,259 documents, we present a scoping review of AI audits that treat system integration as a core tenet of evaluation (n = 58). Using reflexive thematic analysis, we analyze their elements, actors, enablers, and constraints. We find that the corpus represents an emerging yet still fragmented form of AI auditing: few existing measures target integration-specific risks; large gaps remain in meeting traditional audit expectations; and access to necessary information and resources significantly influences audit design. Nonetheless, integration can be categorized across three sites (inter-component, system-environment, and multi-system), each serving the functions of risk exploration, risk determination, coordination, and procedural regularity. Deviating from other types of evaluations, these audits assess qualities specific to system integration, including compatibility, completeness, and oversight. This review calls on the AI community to prioritize system integration as a core strategy for addressing AI risk, and to develop audit practices capable of capturing failures across components, environments, and systems beyond the reach of component-level evaluation.

Multilingual large language models exhibit substantial performance differences across languages, while existing adaptation methods often require parameter updates and considerable multilingual training data. We propose an inference-time multilingual steering method that uses pretrained sparse autoencoders to identify and strengthen target-language-related features. Using multilingual parallel sentences, we compare SAE activations across languages and select a small number of layer-specific features associated with each target language. These features are decoded into steering signals and injected into the model's hidden states without additional training. Experiments with Gemma-3-12B-it show average accuracy improvements of 10.9 percentage points on XCOPA, 5.3 points on XNLI, and 1.9 points on MGSM.

Video-to-audio (V2A) generation extends image-to-audio generation (I2A) by introducing consecutive frames that provide essential temporal cues for audio synthesis. However, existing conditional diffusion-based V2A methods typically enhance visual conditioning with additional audio-visual supervision, acoustic structure prediction, or reasoning from large multimodal models, requiring extra networks or strong inductive biases. Inspired by recent advances in visual representation learning, we introduce TD-V2A, which leverages temporal differences (TD) as the key representation that distinguishes V2A from I2A, enriching visual conditioning with minimal architectural modification. We first investigate TD at both the frame and feature levels to identify the most effective representation level at which TD complements visual representations. Based on these findings, we develop a hierarchically continual learning strategy and an annealed temporal differences guidance method to progressively learn and exploit TD information during diffusion training and sampling process, respectively. Extensive experiments on benchmark datasets demonstrate that effectively exploiting TD through our proposed framework significantly improves end-to-end V2A generation quality, even outperforming dedicated V2A representations such as contrastive audio-visual pretraining.

Confidence estimation is essential when LLMs are used for classification, indicating when predictions can be trusted. However, common approaches such as verbalization produce extremely sparse outputs. For instance, Qwen3-32B verbalizes only eight unique confidence values on SST-2, with over half being exactly 95%, a pattern we observe consistently across four datasets and two LLMs. Besides limiting practical utility, we show that this sparsity critically affects evaluation: the choice of interpolation in area under the accuracy-rejection curve (AUARC) dramatically alters rankings, with consistency sampling dropping from best to worst under stepwise versus linear interpolation. We advocate for standardizing stepwise interpolation for a fairer comparison. Under such a fair evaluation, we find that weighting verbalized digits by token probabilities, a method we term verbalization logprobs, addresses sparsity and achieves the best AUARC (+2.3 points over vanilla verbalization) without incurring additional inference cost.

Defensive driving scores are useful only when they preserve distinctions between policies that observe surrounding actors and those that do not. Re-simulation benchmarks may use reference-conditioned forgiveness, under which an agent receives credit when the logged human reference fails a compliance channel. When agent and reference share an unstable rollout transformation, this rule can propagate shared reference failures into broad compliance credit. We audit this risk in NAVSIM v2.2 original scene single-stage scoring. Under the affected documented-stack condition on the audited numerical backend, the route-blind Ignore-All probe and a route-aware actor-blind probe outrank human replay and PDM-Closed over the complete 12,146-token navtest split. A fresh installation following the public specification reproduces rollout divergence on a fixed 32-token diagnostic set. A same-source dependency stack control and an exact-input diagnostic isolate dependency-sensitive numerical behavior in the shared velocity refit. On a 450-token control pool, replacing only the solver eliminates rollout divergence and restores blind-last ordering while keeping forgiveness enabled. Thus, the numerical instability is the direct trigger. Reference-conditioned forgiveness propagates the resulting shared reference failures into compliance credit. We contribute an audit protocol requiring score basis and stack disclosure, blind probes, overwrite reporting, and rollout stability tests before using such scores for defensive driving claims.

Multi-agent LLM systems relay key--value caches instead of text and credit their gains to exchanged ``latent thoughts''. That credit is a claim about \emph{which} example's cache is relayed, not merely that one is. We audit it causally in released systems. The cache is replaced with deranged (mismatched-example), zeroed, and moment-matched random counterparts, under two regimes defined by whether the receiver needs the sender's private information. Where it does, the battery reads ceiling: 100\% against 23--25\% for answer-irrelevant relays on the primary backbone, a contrast replicated across three families, five checkpoints, and a prose document-QA surface. Where it does not, a pre-registered five-seed protocol establishes equivalence within 2.8 points, a margin anchored to the audited system's reported gain, under Holm-corrected TOST on GSM8K and ARC-Challenge across three Qwen3 scales and on MedQA at 8B (one cell shows a small detected advantage inside the margin); a second family shows no detected advantage. A large cache effect need not be a pairing effect. In one natural cell, zeroing the relay costs 14.7 points; a mismatched cache, 0.4. Nor is need sufficient: under the same test, delivered channels span ceiling (LatentMAS's native relay), partial (KVComm's layer subset), and no detected example-specific transfer (C2C's released projector). Benchmark deltas do not by themselves establish latent-thought transmission; establishing it takes a mismatched-cache audit, which we release.

Standard accuracy metrics for VLMs often mask significant reliability failures in sensitive domains. In this work, we utilize a histopathology-validated brain MRI dataset to systematically assess the diagnostic robustness of four VLM families under evidence-preserving perturbations. By reordering anatomical slices and swapping target label positions, we evaluate whether models maintain consistent predictions when clinical evidence remains invariant. Our results reveal significant vulnerabilities in presentation-order stability, with models exhibiting prediction flips in up to 48.9% of cases under simple sequence reversals. We further identify a textual selection bias, where label reordering triggers inconsistent diagnoses in up to 67.8% of cases despite identical visual inputs. Negative-control tests further reveal diagnostic overcommitment: models generate categorical diagnoses in up to 76.1% of cases after expert-annotated lesion slices are removed. These results demonstrate that high accuracy can overestimate clinical reliability, masking sensitivity to sequential presentation and textual framing that is not captured by aggregate accuracy. Our findings highlight the necessity of stability-based metrics for the deployment of VLMs in safety-critical clinical applications. Our evaluation data and code will be made public upon acceptance.

We introduce a variational approach to a finite-temperature continuous-spin perceptron trained on a Gaussian mixture. The model allows for a broad class of concave utilities and log-concave separable prior measures on the spins. By combining the interpolation method with log-concavity and concentration estimates, we derive lower and upper minimax variational bounds for the limiting quenched pressure. Remarkably, the two bounds differ only in the order of optimization of two variational parameters, while all remaining extrema are controlled by the concave--convex structure of the variational potential. Whenever the two optimizations commute, the two bounds match and identify the solution of the model. The same potential yields the fixed-point equations as stationarity conditions and provides a unified route to the computation of the ground-state energy, training loss, and generalization error.

Vision Transformers (ViTs) achieve strong image-recognition performance, but their parameter count grows linearly with depth when each block is independently parameterized. Single-block recurrent ViTs (bViT) remove this growth by repeatedly applying one shared block. Rather than proposing a new architecture, we fix a bViT and provide a controlled empirical characterization of three training and inference regimes under a common CIFAR-100 protocol, asking: (i)~when does recurrence beat independently parameterized depth---at matched FLOPs or at matched parameter memory? (ii)~when a residual recurrent block is trained through an ODE solver, does solver order act as numerical refinement or as an architectural bias? and (iii)~what does robustness beyond the training horizon cost in nominal accuracy? We find that standard ViTs remain preferable when FLOPs are the primary constraint, whereas recurrent ViTs offer a better accuracy--parameter trade-off under memory constraints. Consistent with the standard view of residual networks as Euler discretizations of ODEs, the continuous-time analogue of a residual recurrent block is the state-subtracted vector field $\dot{z}=F_θ(z)-z$; although known in principle, this distinction is easy to violate when the block is wrapped as a black-box vector field, and we qualify the cost at few accuracy points. Because the vector field is learned jointly with the solver, higher-order solvers act as a solver-induced architectural bias rather than a numerical-accuracy improvement, and their gains are not uniform. Finally, stage-wise deep supervision traces an accuracy--robustness frontier: it does not improve nominal accuracy, but degrades gracefully far beyond the training horizon, where naive recurrence collapses to near-random performance.

Symbolic regression aims to discover closed-form equations from data, but existing LLM-guided methods often rely on a unified proposal loop that compresses heterogeneous search failures into a scalar score and a single prompt. We propose A-SR, a self-evolving agentic framework that shifts the control unit from expression edits to role-conditioned evidence views. A-SR coordinates formula discovery through routing among coordination protocols, an online evaluator-reward role policy, and state-routed process memory. During search, evaluator feedback characterizes reliability and productivity, updates role-level utilities, and routes elite motifs, failure traces, and validity diagnostics to different agents. The framework self-evolves at two timescales: within a run, it adapts the search process without updating LLM parameters; across runs, recorded trajectories can be distilled into open-source LLMs as role-conditioned proposal priors. Averaged over the four LSR-Synth scientific domains in LLM-SRBench, A-SR improves Acc@0.01 over baselines from 25.79% to 48.30% with Llama3.1-8B, while A-SR-LoRA improves the corresponding Qwen3-4B result from 24.58% to 38.29%. On four real-world scientific discovery tasks, A-SR obtains the best in-distribution or out-of-distribution normalized mean squared error on 7 of 8 reported metrics.

This study examines the grammatical status of preverbal uninflected and underived roots in Mapudungun, with particular focus on wuno 'return/re-'. Through a critical review of scholarly classifications--auxiliaries (Smeets, 2008), modal prefixes (Longkon, 2011), and preverbal particles/complex verb stems (Zúñiga, 2006)--we demonstrate the limitations of existing frameworks. A diachronic corpus analysis spanning four centuries (1606-present) reveals that these elements exhibit three distinct profiles: stable V1 compounds (kim, shinge), volatile V1 rates reflecting orthographic shift (pepi, wuno), and a true particle (kalli). The discovery of V2 attestations for kim and kupa confirms their status as full lexical verbs. We propose a prosodic-orthographic hypothesis: apparent "variable binding" results from the fossilization of prosodic pauses transcribed by early missionaries as spaces, a convention later reanalysed by speakers as syntactic boundaries. The evidence supports Zúñiga's radical concatenation as the correct grammatical model, with implications for the study of languages with no pre-contact written tradition.

We introduce the neural echo as a tool for understanding the behavior of neural networks. It generalizes the model-based concepts of impulse responses, diffusion echoes, and filter echoes to learning-based methods. It provides local, space-adaptive impulse responses and filter kernels for a neural network, its so-called echoes. These echoes depend on the input image and can be visualized to understand the learned dynamics of the network via an affine mapping. Neural echoes build a bridge from classical signal processing to modern explainable AI. They are very general and can be applied to both image-to-image and classification networks, with convolutional or fully connected structure, of feedforward or recurrent type, including modern transformer networks. Network differentiability is not required. In the differentiable case, neural echoes comprise concepts based on the network Jacobian, such as saliency maps and the analysis of adversarial perturbations, as special instances. As a simple blueprint to explain our framework, we derive neural echoes for the denoising convolutional neural network (DnCNN). Our experiments suggest that this network weights pixels based on their spatial and gray value distances. This not only clarifies its behavior, but also shows that it can reproduce key concepts of classical model-based denoisers such as bilateral filtering.

This paper develops procedures for nonparametric goodness-of-fit testing under covariate shift, where labelled data are drawn from a source population but goodness-of-fit is evaluated for a target population. The distribution mismatch is quantified by either a bounded moment condition or a sub-exponential tail condition on the target-to-source density ratio. Our method combines truncated importance-weighting kernel ridge regression with a multiplier bootstrap to construct confidence sets for the regression function. The truncation stabilizes the importance- weighting kernel ridge regression as well as the bootstrap calibration, making our approach applicable even when the density ratio has heavy tails. We prove nonasymptotic validity and sharpness of the resulting confidence sets under suitable operator compatibility conditions, and establish explicit error rates for coverage probability under specific conditions on the target- to-source density ratio and on the spectral decay of the kernel integral operator. Numerical experiments corroborate our theoretical findings.

Modern software systems rely on cryptography for data protection, authentication, and trust establishment, yet organizations often lack a structured view of the cryptography deployed across source code, configuration, dependencies, and cryptographic files. This lack of visibility complicates security governance and post-quantum migration planning. This paper presents a static approach for discovering and assessing cryptographic assets in software systems. We introduce a classification of Crypto-Material, Crypto-Artifacts, and Crypto-Invocations, derive an extensible scanner-independent rule repository from it, and implement a static scanner that applies these rules to produce CBOM-oriented output. We evaluate the approach on a synthetic benchmark with known ground truth and on a real-world infrastructure of ten deployed services. The scanner achieves an F1 score of 0.75 for asset discovery and correctly annotates 91% of expected weaknesses and vulnerabilities. In the realworld setting, it processes 57 610 files in under six minutes and discovers 370 cryptographic assets, including six CVE-linked vulnerabilities and 52 post-quantum migration candidates. Real-world coverage is assessed against a manually compiled reference list rather than an exhaustive one. These results show that classification-driven static discovery can provide practical cryptographic transparency for governance and post-quantum migration planning.

Despite its cultural relevance and diffusion, queer slang remains underrepresented in Natural Language Processing research. Towards addressing this gap, we introduce Slang-Q, a manually curated dataset of naturally user-generated English sentences paired with queer slang terms and reference definitions, built upon a newly constructed taxonomy of 118 queer terms. We use this resource to conduct a first exploratory evaluation of language models on their ability to understand and define queer slang under varying prompting conditions. Slang-Q is intended as a basis for studying how current models handle sensitive, community-specific language and whether they can provide accurate and reliable information about such forms of identity and linguistic expression.

Verifying the authenticity of satellite imagery has become increasingly critical given advances in generative artificial intelligence. Highly realistic synthetic imagery produced for malicious purposes (deepfakes) can have major consequences in the remote sensing domain, where this data is a fundamental source of information for science applications, planning, logistics, and monitoring. The remote sensing community lacks high-quality, fine-grained manipulation datasets suitable for training and evaluating detection and image forensics algorithms. Existing datasets are lacking and those that do exist either provide no ground truth masks for evaluating manipulation localization, or consist of entire images generated by GANs or diffusion models, which are inadequate for measuring localization performance. To address this gap, we describe a preliminary dataset construction process and prototype benchmark dataset for satellite image manipulation detection and localization. The dataset contains 60 images total, with 30 images carefully manipulated using three manipulation types including copy-paste splicing and diffusion model inpainting, and 30 authentic images. Each image is accompanied by a ground-truth mask and acquisition metadata, enabling both pixel-level localization metrics, image metadata studies, and analyses of how manipulation detection performance relates to image collection parameters. We describe the dataset construction process and present this initial release to support further research in image forensics and geospatial deepfake detection. The prototype dataset can be downloaded at https://huggingface.co/datasets/geodf/fmow-fake-small.

Control policies optimized in simulation can perform poorly in the real system when the parameters $x$ of the simulator are estimated from limited data but the resulting parameter uncertainty is not represented inside the simulation. A common way to incorporate such ambiguity is to simulate each trajectory of the system under a randomly drawn value for $x$. Since the policy cannot observe the drawn value, it must initially choose controls that perform well across many possible parameter values. However, if the policy progressively observes the system, it can often gradually infer the value of $x$, so that ambiguity vanishes. Over time, the policy then specializes to its estimate of $x$ and loses its robustness. This is undesirable in many real systems, where latent factors are expected to shift. In financial markets, for example, a policy hedging a derivative payoff should remain robust to changes in the volatility regime. To induce such continual robustness, we propose training policies in simulators where ambiguity varies with the system's state but does not systematically decay over time. We formalize this requirement as stationary ambiguity: the simulator should induce a stationary filter process over the latent state. We show how to construct such simulators and demonstrate, on hedging problems, that policies trained under stationary ambiguity preserve robustness to latent factors over time, leading to strong performance on real market data. As a modeling principle, stationary ambiguity informs many simulator design decisions: which models make realistic simulators, how their parameters should be randomized, and how simulator and policy should be initialized. While our experiments focus on hedging, stationary ambiguity may also be useful for other sequential control problems driven by exogenous stochastic processes with shifting latent structure.

Memory is essential as language agents move from isolated tasks to long-horizon, stateful workflows, yet existing evaluations often reduce it to retrieval or question answering. We introduce ContextWeave, a longitudinal benchmark that evaluates whether recalled experience improves downstream agent performance in realistic office-work streams. ContextWeave reconstructs privacy-preserved, multi-month workflows of 14 participants into 1,005 executable tasks, including 568 core evaluation tasks, with instructions, containerized environments, trajectories, and task-specific rubrics. It measures workspace quality and alignment with participant-specific preferences, complemented by diagnostics of relevance, continuity, solvability, and robustness to misleading recall. Across six memory components under a fixed model, the strongest configuration raises Workspace Score from 68.08 to 78.20 and Preference Score from 41.50 to 70.60. With a fixed memory component, recall improves both outcomes for all five tested base models, although gains vary substantially. Our analysis shows that actionable, experience-rich memory supports workflow continuation and reduces redundant exploration more effectively than compact summaries, while it can also be more susceptible to misleading recall. These findings motivate memory systems that optimize not only retrieval relevance but also reliable use during execution.

Large language model (LLM) agents increasingly rely on skills, structured documents that specify when to act, which procedure to follow, and which tools are allowed. Existing evaluations mostly judge the quality of a skill or its contribution to task success, leaving unexamined whether an agent can recognize a relevant skill and apply it on its own. We introduce Skill-Use, a benchmark that evaluates skill use under progressive disclosure, where an agent sees only a skill's name and short description and must retrieve the full procedure before following it. Skill-Use separates three facets of skill use. Trigger measures whether the agent invokes the relevant skill, Compliance measures how faithfully it follows the prescribed procedure, and Boundary measures whether it avoids forbidden operations. A Skill-Use (SU) score combines the three and credits execution only after the skill is triggered. Skill-Use pairs 79 real skills with 177 executable tasks across nine domains, each grounded in real files, run in an isolated Docker sandbox, and scored by a trajectory-based rubric. Evaluating eight LLMs under two agent harnesses, we find that reliable skill use remains out of reach, as the strongest configuration reaches an SU of only 0.613. Triggering and procedural compliance fail as independent bottlenecks, and both scores and model rankings shift with the harness, so skill use behaves as a capability conditioned on the harness rather than a fixed property of the model.

We introduce the Intrinsic Hybrid Latent Diffusion Model (ILDM), a generative framework that integrates probabilistic dimensionality reduction with geometry-aware diffusion on unknown manifolds. While diffusion models (DMs) have achieved state-of-the-art results in high-dimensional data synthesis, they rely on large training datasets and ignore intrinsic geometric structure. Latent diffusion models (LDMs) address the high dimensionality by learning a latent space, but they typically impose a Euclidean structure, failing to capture the underlying manifold geometry, especially problematic in data-sparse regimes. ILDM addresses these limitations by interpreting the latent space as a chart of an unknown Riemannian manifold, with geometry and uncertainty quantified through a probabilistic decoder. The forward process is a hybrid diffusion that switches between Riemannian and Euclidean dynamics based on local uncertainty, where the Riemannian component is governed by a probabilistic metric tensor derived from the decoder. To learn the generative dynamics, we introduce an approximate denoising score matching method tailored to the hybrid diffusion setting, enabling a backward process defined by hybrid Langevin dynamics. Experiments on COIL-100, MNIST, and cardiac MRI datasets demonstrate that ILDM significantly improves generation quality, achieving lower FID and LPIPS scores compared to standard diffusion and latent diffusion models.

Part-of-speech tagging for low-resource languages remains challenging due to limited annotated data, especially for linguistically complex languages. Gaidhlig (Scottish Gaelic) is a morphologically rich and endangered language with limited digital resources, making it suitable for examining a lightweight language processing approach. This paper describes using the modular spaCy Natural Language Processing framework to build part-of-speech taggers for Gaidhlig using the Annotated Reference Corpus of Scottish Gaelic. We train two models with minimal pre-processing and configuration: one using a fine-grained tagset and another using a reduced coarse-grained tagset. Both models are trained without external embeddings or pre-trained language models, using only supervised learning from the available corpus. The fine-grained model achieves 88.6% tagging accuracy, while the coarse-grained model achieves 93.7%. The results are comparable to those of the two previously published Gaidhlig taggers, indicating that simple, off-the-shelf language processing pipelines can demonstrate good performance in low-resource and morphologically complex linguistic settings.

Leak detection models for multiphase pipelines often degrade when deployed under flow regimes that differ from training. Existing evaluations typically assess performance under in-distribution operating conditions, masking failures caused by regime transitions such as bubble-to-slug flow. We propose the Manifold Gated Signature Bias (MGSB), a regime-aware architecture combining regime-conditioned feature fusion, a TT-RoughPath encoder, and Mean-Teacher consistency regularization to improve robustness under distribution shift. Under leave-one-group-out evaluation, MGSB achieves a detection F1 of 0.930 and an OOD F1 of 0.783, substantially outperforming CNN-LSTM and fully connected baselines under severe feature corruption. Ablations show the proposed architecture, not the training procedure, is the primary contributor to OOD robustness, while Mahalanobis-distance analysis confirms the held-out conditions are genuinely out-of-distribution. These results show that explicit regime-aware modelling is a practical path toward robust, sensor-agnostic leak detection in industrial multiphase pipelines.

Frontier language models can resolve repository-level software issues, but each attempt is expensive, and existing routers select a model from the issue text alone. We present SuperScout, which routes after scouting the repository: a 7B searcher, SuperScout-7B, first explores the repository and produces a structured handoff whose reproduction claims are sandbox-verified, with false claims stripped before delivery. The searcher's hidden states, together with the task text, then feed a resume-based router that dispatches the task to one of four frontier fixers. Adding a new fixer requires no retraining. On the full Python slice of SWE-bench Pro (266 tasks) under the benchmark's official capped budget tier, SuperScout matches the best single model's solve rate (159 of 266 for SuperScout, 158 for the best model) at about a fifth of the total cost per solve, and the reported configuration sits above the random traffic-splitting baseline. A no-router ablation, always the cheapest fixer with the handoff, ties the routed system on this benchmark, so the handoff rather than the routing decision carries the result. A paired calibration study points to the mechanism: the handoff appears to redistribute rather than add solving ability, lifting the three cheaper fixers while slightly hurting the strongest, though at $N=99$ the per-fixer effects are directional only; the searcher's hidden states improve cost routing on the calibration labels while the handoff's own text does not. The searcher's compute adds less than half a cent of GPU time per task.

Self-distillation (SD) has emerged as a compute-efficient alternative to reinforcement learning with verifiable rewards: a self-teacher, conditioned on privileged information (PI) about the answer such as a reference solution, supplies dense per-token supervision to a student that never sees it. Reported gains, however, come almost exclusively from narrow, low-difficulty settings, leaving open a basic question: as a lone objective, with no reward term, does SD teach anything? We reproduce SDPO's reported gains in its easy setting, then apply the identical setup to difficult tasks and find that it does not. Across question answering, mathematics, coding, and multi-turn agentic tool use, across reasoning modes, model sizes, and forms of PI, and under both the SDPO and OPSD recipes, the per-token loss falls steadily while validation accuracy does not improve and typically degrades. We explain this failure through a single causal chain from the loss to the model it produces. The chain begins with PI bias: having seen one particular reference solution, the teacher's per-token target is pulled toward that trajectory rather than toward correctness in general, an effect we quantify with a PI Bias Score. Trained to match this target everywhere, the student's objective becomes nearly blind to whether a rollout is correct, and the loss it assigns falls mostly on low-information tokens like stopwords, punctuation, uncertainty markers, rather than those that determine the answer; within correct rollouts the exploratory tokens incur the highest divergence, so it penalizes the hesitation that reasoning requires. The result is a flatter, less decisive student that is no better at reasoning: as a lone objective, SD optimizes a signal decoupled from task success.

Accurate estimation of Above-Ground Biomass (AGB) from satellite imagery is essential for the large-scale monitoring of carbon stocks, yet it remains a challenging regression task at global scale. Geospatial Foundation Models (GFMs) have recently emerged as a promising machine learning paradigm to derive general-purpose representations from Earth observation data, but their utility for quantitative regression tasks like biomass estimation remains largely unexplored, as most benchmarks emphasize classification and segmentation. Here, we present a comprehensive benchmark of GFMs for global-scale AGB estimation using the AGBD dataset, a machine learning-ready benchmark spanning diverse biomes and geographies. We distinguish two ways in which GFMs reach practitioners: (i) models distributed as weights to be run by the user, which we evaluate as frozen encoders within the PANGAEA benchmarking framework; and (ii) models distributed as ready-to-use, pre-computed embedding products, for which we evaluate AlphaEarth Foundations (AEF) and TESSERA. We compare 11 GFMs available on PANGAEA and both embedding products against a fully supervised state-of-the-art (SOTA) model, assess their geographical and temporal generalization abilities, as well as agreement with the ESA CCI biomass product on independent reference data. Our results show that GFMs run as frozen encoders substantially underperform with respect to the supervised SOTA model, whereas pre-computed embedding products prove highly effective. An MLP trained on AEF embeddings outperforms the supervised SOTA model trained on AGBD features, and the same SOTA model trained on AEF embeddings (optionally augmented with selected raw features) achieves the best overall result, while also generalizing better across space and time.

Large language model agents are commonly trained through reinforcement learning with sparse trajectory-level rewards, which offer limited guidance on how strongly individual tokens should be updated. On-Policy Self-Distillation (OPSD) addresses this by re-scoring generated tokens under a privileged replay view to obtain dense, token-level supervision. However, we identify a confounding issue: the resulting support may reflect both the privileged information contained in the replay view and score shifts induced by the replay scaffold, making it difficult to attribute the support specifically to that information. This issue is especially pronounced when future environment observations serve as privileged information, since replaying them requires reconstructing an extended scaffold that itself perturbs token scores. To resolve this confounding, we propose Observation-Calibrated Self-Distillation (OCSD), which contrasts two structurally matched replay views, Full and Observation-Ablated, differing only in whether the actual future observation is present, to derive an observation residual that discounts score changes shared by the replay scaffold. OCSD then applies this residual to modulate token-level GRPO updates at high-uncertainty steps, while preserving the trajectory-level update direction. Experiments on ALFWorld, WebShop, and Search-QA across three Qwen3 model scales show that OCSD consistently outperforms strong baselines. Diagnostic analyses further confirm that the calibrated residual aligns better with local environment feedback. Our code is publicly available at https://github.com/yiy1x/OCSD.

We settle the exact complexity of the reachability problem in (stateless) vector addition systems (VAS) in fixed low dimension. In dimensions 2-4 it has only been known to be sandwiched between NP and PSPACE. We prove PSPACE-hardness of the reachability problem for symmetric vector addition systems in dimension 3 (3-VAS), a restricted fragment of general 3-VAS. Combined with previously established PSPACE upper bounds, our result settles the complexity of the problem to be PSPACE-complete in 3-VAS and 4-VAS, as well as in their symmetric fragments.

The integration of Large Language Models (LLMs) into software engineering has shifted the focus from function-level generation to repository-scale assistance. However, existing benchmarks largely rely on bug reports from GitHub Issues, which often allow models to bypass genuine understanding via pattern matching on error logs. This misalignment under-measures Edit Bias, which refers to premature generation, where models prematurely propose code modifications instead of understanding the existing repository architecture. Furthermore, current LLM-as-a-Judge scalar scoring suffers from high variance and low interpretability. This work introduces RepoProbe, a novel benchmark for evaluating repository-level code understanding through open-ended Q&A using GitHub Discussions, which focuses on open-ended architectural inquiries rather than defect reporting. To ensure rigorous evaluation, we propose a Checklist-Based Verification Protocol that decomposes answers into atomic, verifiable facts, thereby replacing subjective ratings with objective verification. Our evaluation of state-of-the-art (SOTA) LLMs reveals a persistent gap between high clarity and evidencegrounded technical correctness. It also quantitatively confirms the prevalence of edit bias, in which models prioritize code generation instead of architectural analysis. Finally, we demonstrate that our verification protocol significantly improves evaluation reliability compared to traditional evaluations with scalar scoring.

Physics-informed neural networks (PINNs) incorporate governing equations into neural-network training and can approximate PDE solutions without requiring large observational datasets. Parameterized PINNs (ParamPINNs) further take physical parameters as inputs, allowing a single model to represent a family of PDE solutions over a parameter domain. Existing ParamPINNs, however, still face inefficient training, uneven accuracy across parameters, and overfitting to a limited set of sampled parameter tasks, which can impair generalization to unsampled parameters. To address these issues, we propose a continual-learning physics-informed neural network (CL-PINN), which treats PDE instances at different parameter values as related tasks and learns them sequentially. CL-PINN combines Bayesian-optimization-based active parameter selection, task-wise dynamic loss weighting, sparse physics-constrained replay, and an optional parameter subnetwork to improve task allocation and knowledge retention under bounded active-task capacity. It requires no observational data and is designed to solve parameterized PDEs over relatively broad parameter domains under limited computational resources. Multi-seed evaluations on five benchmarks, including one continuous function and four parameterized PDEs, show that Bayesian selection substantially reduces objective-loss queries relative to grid-greedy search, while sparse replay mitigates forgetting of earlier tasks. Under the prescribed within-case resource protocols, CL-PINN generally provides higher and more balanced solution accuracy than fixed-sampling and grid-greedy baselines. CL-PINN offers a practical route toward learning PDE solutions that generalize across physical parameters and has the potential to support reusable physics-informed surrogates for large-scale engineering parameter studies.

Oscillatory signals, such as vibration, carry class-discriminative information in specific frequency bands; perturbing them in raw feature space for counterfactual analysis easily destroys their temporal structure and produces physically implausible results. In this work, we introduce IMFACT (IMF-based counterfACTuals), a model-agnostic framework for generating plausible counterfactual explanations for time series classifiers that operates in the decomposition space of Empirical Mode Decomposition. An input signal is split into Intrinsic Mode Functions (IMFs), and selected IMFs are progressively substituted with those of a Nearest Unlike Neighbour (NUN) until the classifier flips to the target class. We evaluate six IMF-selection strategies and a multi-NUN cycling extension on two UCR benchmarks (FaultDetectionA, FruitFlies). The variance-based strategy with three NUNs outperforms two prominent baseline techniques on reliability and plausibility metrics, while cycling across three NUNs yields the best proximity across both datasets.

The ability to accurately assess and anticipate risks in safety-critical scenarios is crucial for autonomous driving systems. While existing research has made progress in collision prediction, accurately quantifying risk levels from monocular vision inputs remains challenging due to the complex dynamics of multi-agent interactions and the inherent uncertainty in real-world environments. To address these challenges, we present NSF-HRPT, a novel framework that combines learning-based perception with structured reasoning for quantitative risk assessment. Our approach features a Neural Semantic Field (NSF) that learns to model scene semantics, trajectory predictions, and probabilistic Time-to-Collision (TTC) distributions from simulation data. During inference, the pre-trained NSF serves as a prior for our Hierarchical Risk Perception Tree (HRPT), which enables efficient parallel computation and spatial reasoning about multi-agent risks. Additionally, we introduce a Sim2Real enhancement strategy that improves real-world applicability without retraining by incorporating priors from foundation models. Extensive evaluations demonstrate that our framework achieves state-of-the-art performance on synthetic benchmarks and delivers competitive, near-state-of-the-art results on real-world datasets for both TTC estimation accuracy and risk localization precision. The proposed method provides an effective solution for real-time risk awareness from monocular camera inputs.

Scaling supervision for multi-turn medical agents is difficult because expert dialogue annotation is costly and clinical conversations are privacy-restricted. We introduce Guideline-as-Oracle (GAO), which compiles American Academy of Ophthalmology guidance into a 70-row operational rule table and uses it as the sole source of instance-level supervision for 3,000 training dialogues, reserving human labeling for evaluation. Because converting rules into dialogues is itself a design problem, we catalog eight construction strategies, including cited-row tier assignment, one-fact boundary pairs, metadata-only repair, and label repair, and characterize the evidential status of each: labeling mechanism, null, confounded, or evaluated only as a package. Fine-tuning a 9B backbone on this corpus yields GAO-Triage, improving agreement with a 201-case operational reference from 61.7% to 74.1% (exact McNemar p=0.0046) and emergent-case recall from 9.5% to 69.0%; the gains persist across a second seed and patient simulator. None of the seven general-purpose systems we test dominates GAO-Triage on both metrics, and GAO-Triage requires no frontier model at inference time. Permuting label-dialogue assignments collapses the model to a constant-routine predictor, indicating that the signal lies in guideline-derived assignment rather than dialogue surface form. Label repair coincides with the disappearance of a late-training safety degradation.

Large Reasoning Models (LRMs) excel on complex tasks through long chain-of-thought (CoT) reasoning, but their lengthy intermediate steps cause severe overthinking that inflates inference cost. KV-cache compression is a common solution, yet existing reasoning-oriented methods apply a uniform policy across the trajectory and judge compression only by what it removes from the cache. Two observations point the other way. First, a reasoning state's tolerance to context loss varies along the trajectory, and process reward tracks it: deleting tokens at high-reward steps preserves accuracy far better than deleting the same budget at random. Second, compression is not free on the generation side, since a smaller cache leads the model to generate more tokens, partly canceling the saving. Together these motivate coordinating both sides under a single process reward. We propose ReCo (Reward-Coordinated Compression), a step-wise framework in which a lightweight process-reward estimator scores each completed step and drives three components: (1) reward-adaptive KV-cache compression that shrinks the retained cache harder at high-reward steps and less at low-reward ones, (2) a reward-banded penalty on reflection tokens that curbs redundant generation, and (3) confidence-based early stopping that triggers when the reasoning is reliable. Across three reasoning models and six benchmarks, ReCo reduces generated tokens by 37%-65% and end-to-end latency by 2.08x-2.35x over Full CoT, all while largely preserving accuracy.

A large number of infants with congenital anomalies are born each year globally, especially in areas with underdeveloped medical resources. Currently, fetal ultrasound screening is the most common modality for early pregnancy anatomy detection. This modality can detect anomalies earlier and provide opportune treatment advice. However, the lack of an ultrasound dataset on early fetal gestation has slowed down the development of automated assisted diagnosis. In this work, we present a benchmark dataset for Fetal Ultrasound Screening in Early Pregnancy to facilitate intelligent ultrasound examination and assisted diagnosis called FUSEP. Our dataset consists of two ultrasound views recommended by the international guideline, i.e., Crown-rump Length (CRL) and Nuchal Translucency (NT) views in three hospitals, totaling 4,017 ultrasound images, with 45,820 box-level expert-level annotations. Our dataset and baseline present the following three contributions: 1) Our medical experts annotated a total of 14 key anatomical structures in two views using a box-level format; 2) Our data is collected extensively from different sonographers, devices, scanning angles, hospitals, etc; 3) We report the performance of the semi-supervised learning, fully supervised learning, unsupervised domain adaptation (UDA), and source-free UDA in ultrasound images multi-object detection. To the best of our knowledge, this is the first publicly available dataset and benchmark for fetal early pregnancy ultrasound screening. We believe that FUSEP and benchmark can contribute to the medical community in the development of multiple tasks such as standard plane recognition, quality control on ultrasound images, automated assisted diagnostics in early fetal pregnancy, medical multi-object detection, domain adaptation for object detection, etc.

Vision-language-action (VLA) models provide a unified paradigm for connecting visual perception, language understanding, and robotic control. However, existing VLA models still face major challenges in long-horizon tasks: sparse expert demonstrations constrain cross-task compositional generalization; the non-Markovian nature of long-horizon tasks makes it difficult for policies conditioned only on current observations to maintain temporal consistency; limited closed-loop error correction allows execution errors to accumulate; and end-to-end action fine-tuning may weaken the high-level semantic representations of vision-language model (VLM) backbones. To address these issues, we propose a hierarchical long-horizon VLA architecture with an explicit language-memory module. The central idea is to convert discrete temporal observations into a coherent textual memory sequence with temporal logic. The system is decoupled into a high-level VLM and a low-level VLA: the high-level VLM performs semantic reasoning through a visual question answering training paradigm, while the low-level VLA executes precise continuous control conditioned on subtask instructions and visual observations. The high-level VLM recursively updates both language memory and subtask instructions using the previous memory as a contextual anchor, enabling persistent temporal tracking and dynamic correction during long-horizon execution. We evaluate the proposed method in multiple simulation environments and conduct sim-to-real experiments on a real robotic platform. The results demonstrate that explicit language memory improves the success rate and robustness of VLA models on complex long-horizon tasks while providing an interpretable semantic account of the decision process.

Self-improving agents accumulate reusable insights from prior trajectories, making retrieval increasingly important for turning accumulated experience into actionable guidance. At each decision step, retrieving the right insight can help the agent progress toward its goal, a setting we refer to as agentic insight retrieval. However, existing retrieval methods primarily model semantic similarity, while overlooking whether a retrieved insight resolves the agent's current decision bottleneck. We propose InsightEmb, a contrastive embedding framework that learns transferable progress-oriented retrieval geometry using only mathematical reasoning data. InsightEmb jointly learns to align concrete situations with abstract heuristic rules and to cluster reasoning trajectories with similar progress structures. We evaluate InsightEmb on dynamic agent tasks and a static skill-retrieval benchmark. Without any environment-specific training, InsightEmb improves over all these evaluations, surpassing the performance of existing reasoning embedding models. These results suggest that the geometry of state-insight matching can transfer across domains, enabling effective training from publicly available reasoning data without expensive environment-specific supervision.

Although Multimodal Large Language Models (MLLMs) have made substantial progress, their spatial reasoning may still produce intermediate judgments inconsistent with the input image, allowing errors to propagate through the reasoning chain and affect the final answer. Existing methods mainly improve spatial reasoning through training or additional spatial information, without considering whether the reasoning process itself is faithful to the model input. Our study shows that unfaithful reasoning chains significantly reduce final-answer accuracy. To address this issue, we propose a modular and training-free framework for spatial reasoning verification and correction. The framework constructs a Spatial Evidence Graph (SEG), which associates atomic spatial evidence extracted from Chain-of-Thought reasoning with visual entities, spatial relations, source steps, and visual evidence. Spatial Evidence Reliability Assessment (SERA) evaluates the reliability of visual evidence based on object existence, localization, and geometric measurements. The framework then identifies the earliest spatial evidence unit contradicted by reliable visual evidence and guides the original MLLM to revise the subsequent reasoning and final answer. Across 15 model-dataset settings, our method achieves an average accuracy of 68.94%, outperforming the compared baselines by 8.55 percentage points on average. Our code will be open-sourced.

In Retrieval-Augmented Generation (RAG), post-retrieval conflict resolution arbitrates among noisy or contradictory retrieved passages. However, the robustness of this safeguard against knowledge poisoning has not been adequately studied. Existing black-box poisoning methods all assert the target answer in frontal contradiction with what the resolver treats as settled, the very signal these methods are built to detect. We propose PURPOSE, a strict black-box poisoning attack that reframes the injection as an update that minimizes conflict, rather than as a counter-claim. PURPOSE extracts query-related facts approximating the resolver's possible reference, then grounds a pivot event in them to keep the injection consistent with what the resolver might verify while steering the generator toward the target answer. Across three QA benchmarks, five generators, and three conflict-resolution methods, PURPOSE attains the highest attack success rate (ASR) in 35 of 45 settings and exceeds the strongest prior attack with +9.7 mean ASR points. These results show that our poisoning method is effective against conflict resolution in RAG and identify non-contradicting injection as a practical mode to enhance poisoning attack.

Attention layers are the backbone of today's most powerful and impactful models. Models with multi-million and billion parameters rely on contextual knowledge provided by attention layers. However, their use goes well beyond just being the core component of large language models. One particularly interesting application is in Memory Augmented Autoencoders (MemAE), specifically for unsupervised representation learning in outlier detection tasks. It was shown that attention helps these models be more effective in centralized learning scenarios. Our work aims to address the lack of specialized aggregation techniques in Federated Learning (FL) when it comes to MemAE models. In this paper we analyze the intricacies of the architecture behind Memory Augmented Autoencoders, and propose novel, guided approaches to effectively aggregate these models in federated scenarios. We demonstrate our approach on non-IID datasets and show that these novel aggregation schemes are more robust when dealing with numerous edge nodes in environments with unbalanced datasets, specifically for unsupervised anomaly detection scenarios. This approach improves the performance of even very shallow autoencoders, allowing them to be used in resource constrained environments.

Similes provide a compact and expressive way to describe visual characteristics in text prompts. Recent text-to-image models (t2i models) can produce visually compelling outputs from simile prompts, yet even frontier models frequently misinterpret the metaphorical vehicle and confuse it with the object. These systematic failures reveal a gap between figurative language and object-level visual grounding in t2i models. To investigate this issue, we propose a scalable evaluation framework for simile understanding. Our framework includes (1) a controlled simile dataset in which metaphorical vehicles are drawn from a predefined set of object-detectable categories and combined with diverse templates, (2) automatic grounding metrics based on YOLO (You Only Look Once) detection, and (3) text encoder layer analysis using Diffusion Lens to track how metaphorical vehicles emerge during generation. Experiments across architecturally diverse t2i models reveal consistent literalization failure patterns. We further discuss potential mitigation strategies for improving simile grounding in t2i models.

LLM agents that persist across sessions accumulate stored memories whose validity varies enormously by content type, yet existing memory architectures treat all memories as equally persistent and systematically contaminate retrieved context with outdated facts. We show that per-memory, type-conditioned temporal decay, a property of western scrub jay episodic memory, can be operationalized as an auto-classified coefficient $π_i$ in an external LLM-agent memory store, yielding ScrubJay-MEM: each memory is encoded as a jointly-bound What--Where--When tuple with an estimated perishability $π_i$ and utility horizon $τ_i$, retrieved by query-adaptive scoring, and revised retroactively at $O(1)$ LLM calls per update. We introduce the Temporal Generalization Test (TGT), a benchmark with held-out retention intervals and a Generalization Gap (GenGap) metric. On TGT, ScrubJay-MEM is the only retrieval-based system with substantially positive GenGap ($+0.108$); on MemoryAgentBench EventQA-64k it improves F1 by $+2.66$ over Mem0 and $+3.09$ over Qwen3-Embedding-4B under a llm backbone. A decay ablation collapses GenGap by $5.7\times$, establishing type-conditioned decay as necessary for the result. Gains narrow under stronger backbones and reverse on fact-consolidation tasks, scoping the contribution to temporal reasoning over perishable facts.

Autonomous research agents can generate hypotheses, execute experiments, and draft manuscripts, yet their outputs often contain unsupported claims and inconsistencies between research questions, experiments, results, and conclusions. We argue that this problem is partly architectural: existing systems organize research as sequential pipelines but do not explicitly maintain or validate the evolving claim-evidence structure across stages.In this paper, we introduce EviGraph, an autonomous research framework that represents the research process as a typed evidence graph containing Problem, Gap, Hypothesis, Experiment, Finding, and Claim nodes. The graph serves as the operational state of the agent rather than a post-hoc record. EviGraph inspects evidence chains for missing dependencies, semantic misalignment, and result-claim inconsistencies, localizes the earliest weak node, and regenerates its affected downstream subgraph. Graph checkpointing prevents unsuccessful repairs from corrupting previously validated evidence. Manuscripts are generated only after every retained claim is grounded in a validated evidence chain.Experiments on ARC-Bench-ML and NanoResearch-20 show that EviGraph outperforms the compared end-to-end research-agent baselines in overall research performance, improves Claim Support Rate by 40.19% over the strongest baseline, and achieves 87.73% Experimental Data Consistency. These results demonstrate the value of explicit evidence-state maintenance for reliable autonomous research.

Chain-of-thought (CoT) monitoring is increasingly treated as an important safety layer for frontier reasoning models. Most monitorability evaluations study explicit-influence settings: setups where the prompt directly incentivizes the model to hide something, e.g., by instructing it to perform a hidden side-task. A complementary axis for CoT-monitor evaluations is implicit-influence settings, where the prompt contains no instruction to hide, but the model's behavior is still shaped by features of the task or context, e.g. an irrelevant detail about a candidate that biases a hiring rating. We introduce the first benchmark that directly compares CoT monitorability under the two regimes. We test how model behavior changes in the presence of a nudge to choose a particular option. The nudge is delivered either as a casual aside (implicit), or as a direct instruction to act on the nudge and to conceal having done so (explicit). The benchmark spans four task formats (binary choice, numeric rating, multiple-choice QA, open-ended coding) and seven frontier extended-thinking models. Under explicit influence, a CoT monitor detects 60-94% of behavior shifts: even models instructed to conceal it leak the instruction into their CoT. Under implicit influence, the same factors still shift behavior, but detection falls by 41-46 percentage points in two of our four settings. Realistic system-prompt additions (of the kind a developer might deploy to reduce off-topic bias) lower implicit detection further, to as low as 5%, while preserving the behavioral influence itself. These results suggest that monitorability estimates obtained in explicit-influence settings may over-estimate monitorability, and that monitorability can be further decreased by well-intentioned deployment choices. Our benchmark and code are available at https://github.com/agatha-duzan/implicit-vs-explicit-influence

Safe actor-critic control often treats barrier filtering, uncertainty estimation, and experience replay as separate modules, even though each changes the data used for learning and control. We develop an integrated architecture in which the uncertainty estimate updates the obstacle geometry used by a control barrier function, filter interventions and estimation residuals determine replay priority, and the critic learns from the executed rather than nominal action. We instantiate the architecture on a two-dimensional robot-navigation task with corrupted obstacle measurements and compare six component-matched configurations under common training budgets, random seeds, sensor streams, exploration, and disturbances. Evaluation includes a moderate post-training test, an eleven-level perception-noise sweep, and an exploratory extreme-stress test at multiplier $6.0$. In the extreme test, the integrated configuration recorded no contacts and reached the goal in all five evaluation seeds. Its mean cost was $7.63\pm0.44$ and its obstacle-belief root-mean-square error was $3.52\pm0.55$ cm. The uncertainty-estimation ablation also recorded no contacts but reached the goal in four of five seeds, with mean cost $8.96\pm2.08$ and belief error $11.08\pm1.23$ cm. A finite-training bound clarifies replay exposure, and a robust barrier condition states the required estimation-error and feasibility assumptions. The results support coupling estimation, safety filtering, and replay on this benchmark; broader safety and convergence claims require further study.

Multimodal large language models increasingly reason over screenshots and documents where the task itself may be written in pixels. Yet benchmarks usually place questions in text, leaving it unclear whether models use the same instruction equally well across channels. We introduce Visualized Task Semantics (VTS), a controlled intervention that moves the question into the image while keeping the source problem and answer fixed. Across six MLLMs and four benchmarks, accuracy drops in all 24 model-task pairs, by 17.8 points on average. Models often transcribe the visual question correctly yet fail to use it, exposing a semantic channel gap beyond OCR. To reduce this gap, we present prompt-region grounding, whose core design aligns the question region with typed semantics and recovers its clean representation from a masked view. At matched training cost, our method raises four-benchmark VTS accuracy from 58.0 to 66.3 while preserving accuracy on the original interface, and requires no OCR or region metadata at inference. Reading task-bearing text and grounding it as an instruction for reasoning are distinct capabilities.

Agent evaluations tell us that a model picked the wrong tool, but rarely why. We introduce canary tools: diagnostic probe tools planted in an agent's Model Context Protocol (MCP) tool set, each engineered to probe one specific tool-selection weakness. A six-type taxonomy (semantic decoys, parameter traps, capability mirages, prerequisite blindness, temporal decoys, and granularity traps) turns a single "wrong tool" outcome into a multi-dimensional profile of how a model reasons about tools. We evaluate eight models -- six hosted and two 8B open-weight -- spanning three capability tiers, on 120 tasks across three canary-density conditions and three seeds (8,640 runs), plus a 2,880-run subtlety ablation. Task success is graded by a provider-independent judge, corroborated by a second independent judge (Cohen's kappa = 0.75). We report three findings. First, susceptibility drops sharply as models get more capable: the per-task canary susceptibility rate (CSR) ranges about 36x across models, lowest for Claude Opus 4.8 and highest for Llama 3.1 8B. Second, capability tier alone does not predict safety: the most susceptible hosted model is mid-tier, and within a provider the cheaper model can be the safer one. Third, the taxonomy is capability-stratified: capability mirages most reliably trap frontier models, while the other types are largely inert on strong models but fire on small open models, so they discriminate by capability rather than being weak. Softening each canary's give-away phrase leaves frontier CSR essentially unchanged, evidence that the probes measure reasoning, not phrase-spotting. Susceptibility also predicts task failure (Spearman rho = -0.34), while the most robust models are not significantly degraded by canary pressure. We release the framework, canary schemas, tasks, and logs.

Benchmark scores are reported as properties of a model, yet the inference framework used to produce them, such as HuggingFace, vLLM, or Ollama, are considered non-influential and their names and versions are almost never disclosed. In this work we investigate how much this choice can influence the model output. In a fully-crossed study (three instruction-tuned models x five inference frameworks x six benchmarks x four generation modes) we investigate how different tools (wrappers/backend) influence benchmark scores and how their score changes is influenced by generation hyper-parameters. We find backend to be a non-negligible factor where even under greedy, sampling-noise-free decoding, changing the backend can significantly alter models performance and this effect is structural and strongly model-dependent. Decomposing the variance according to generation mode reveal that considerable portion of the variability (roughly 39\%) a practitioner sees out-of-the-box can stem from the backend, while the remaining stems from sampling noise and each framework's default generation parameters, both of which are avoidable by disclosing and matching the generation configuration. These divergences are more pronounced on factual than on social-bias benchmarks. Overall, benchmark numbers are not backend-agnostic therefore, we recommend disclosing the backend, its version, and the full generation configuration, also using deterministic decoding for cross-backend comparison.

Integrated Sensing and Communication (ISAC) combines sensing and communication to efficiently utilize wireless resources and is emerging as a key paradigm for next-generation wireless networks. By leveraging the wide bandwidth, high frequencies, and massive antenna arrays of 5G-Advanced and 6G systems, ISAC enables physical-layer sensing using Channel State Information (CSI). The 3rd Generation Partnership Project (3GPP) Release 19 identifies 32 potential ISAC use cases, with particular emphasis on detecting and tracking moving objects. In this work, we address the Sensing for Railway Intrusion Detection use case, where intruders, including wildlife, entering a railway track can pose serious collision risks. We generated 22,695 CSI matrices with corresponding ground truth using a 3D-rendered railway environment and the Sionna radio simulator. We developed a machine learning model combining a three-dimensional Convolutional Neural Network (3D CNN) and Bidirectional Long Short-Term Memory (BiLSTM) network to detect intruders in the track danger zone and estimate their real-time position relative to the train, velocity, and time to collision. On synthetic CSI data, the model achieves 99.57% intruder-detection accuracy on a balanced test set and a combined Mean Absolute Error (MAE) of 0.4240 for position, velocity, and time-to-collision prediction. These results demonstrate the potential of CSI-based ISAC sensing with machine learning for reliable railway intrusion detection. The complete codebase for CSI generation, preprocessing, and model development is publicly available at https://github.com/EdgeIntelligenceLab/6g-isac-railway-intrusion-detection.

This paper presents EmpaAva, to our knowledge the first open-source, agentic 3D-avatar empathetic chatbot, which carries empathetic response generation (ERG) from text-only exchanges into live, face-to-face interaction. Through a video-call-like interface, a user speaks to a 3D digital human that reads their affect from speech and optional vision, and replies with emotional speech, lip-synced facial motion, and photorealistic 3D Gaussian rendering. At its core, an LLM coordinates a Tri-Agent Architecture, in which perception, empathetic response planning, and embodied rendering form a closed loop, paired with a Response Planning layer that compiles each reply into an executable multimodal plan, keeping voice, expression, and rendering on one empathetic intent. Building on strong open-source modules, EmpaAva supplies the intelligence that binds them into one controllable, inspectable experience. In automatic and human evaluations, EmpaAva surpasses text-only, 2D talking-face, and multimodal avatar baselines in emotion understanding, response quality, and audio-visual consistency. We open-source EmpaAva with an online live demo.

Monitoring of offshore wind energy infrastructure life cycles, especially during the deployment phase, is an important contribution for stakeholders to make informed decisions in a phase of increasing deployment activities. ESA's Sentinel-1 Synthetic Aperture Radar (SAR) mission produces large data archives that enable the global monitoring of offshore wind infrastructure. Turning these high-volume archives into information requires algorithms that automatically extract single event labels from dense time series at a global scale. In this study, we present a structured comparison of ten deep learning model-training variants for the dense classification of Sentinel-1 based offshore wind infrastructure time series, aiming to advance rule-based event classification of this task. We trained LSTM, Transformer, and fully connected model variants with monotemporal, unidirectional, and bidirectional context awareness, each with and without self-supervised pretraining. Among these, the supervised BiLSTM performs best, raising the target AUC score from 0.7853 for the rule-based baseline to 0.8509, and the perfect match rate from 0.3508 to 0.5063. Combining the BiLSTM predictions with the existing baseline labels in a label-transition-minimising ensemble further improves agreement with the test data. Using these improved labels, we isolate the deployment phase of individual turbines at a global scale and conduct a regional and subregional analysis covering 2016-01-01 to 2025-03-31, reporting median deployment durations of 84 d (China), 242 d (EU), and 258 d (UK). Deployment-related drivers, including legal regulations such as subsidies, and environmental conditions, emerge clearly from the analysed results across multiple spatial scales.

Benchmarking deep learning (DL) models for multi-label classification (MLC) of remote sensing images (RSI) typically yields rankings that do not generalize beyond the evaluated datasets. In this work, we move beyond rankings by employing functional analysis of variance (fANOVA) to systematically quantify the contributions of individual design choices and their interactions to performance variability. We conduct two empirical analyses covering 48 and 20 DL models, respectively, spanning design choices such as network architecture, fine-tuning strategy, learning strategy, and initialization. By applying fANOVA across seven MLC RSI datasets, we construct dataset meta-representations that capture design-choice sensitivity profiles. Hierarchical clustering of these meta-representations reveals that datasets naturally group according to how they respond to design decisions, with patterns strongly linked to intrinsic dataset properties such as scale, spatial resolution, and label space complexity. Our findings show that for large-scale datasets, fine-tuning strategy and architecture are dominant factors, while in data-limited regimes, initialization becomes decisive. For intermediate regimes, the interaction between architecture and learning strategy governs performance.

Large language models (LLMs) are increasingly used for question answering, education, and research, including in religious and cultural domains where answers depend on specialised source traditions. Yet in Islamic Studies, key concepts, methods, and debates preserved in the authoritative scholarly tradition, known as turath, lack high-quality annotated resources. We introduce IslamicTurathBench (ISTB), a multi-task, multi-discipline dataset for evaluating LLMs on classical Islamic scholarship. Developed and reviewed by domain experts, ISTB contains 3,465 question-answer items drawn from 35 recognised source works spanning more than 12 centuries of scholarship across seven key fields of Islamic Studies. To enable comprehensive profiling of model capabilities, ISTB is structured along two axes: scholarly demand (Beginner, Intermediate, and Advanced) and task format (multiple-choice questions, passage-based comprehension, and open-ended knowledge questions). ISTB includes aggregated scores from a scholarly human reference panel and zero-shot baselines from ten systems. The dataset supports reproducible evaluation of language-model behaviour across source works, disciplines, scholarly-demand levels, and question formats in a historically layered scholarly domain.

We tackle the challenging yet underexplored task of Generalized Referring Expression Comprehension (GREC), which requires a model to localize the object described by a textual expression when it exists (positive sample) and to refuse output when it does not (negative sample). Although Multimodal Large Language Models (MLLMs) excel at localizing existing objects, they often fail to reject nonexistent ones due to the absence of negative samples during training, producing hallucinated bounding boxes. Existing post-training approaches such as supervised fine-tuning (SFT) and reinforcement learning (RL) enhance refusal behavior but usually degrade localization accuracy on positive samples, undermining the model's core competence. To address this, we propose Refusal-Calibrated Group Relative Policy Optimization (RC-GRPO), a calibrated RL strategy that strengthens the refusal ability of MLLMs while preserving localization performance. It enforces "None" outputs in rollouts for valid advantage estimation on negative samples and applies a penalty to prevent over-refusal on positives, achieving a balanced trade-off between accuracy and reliability. A second-stage reasoning reinforcement further consolidates causal understanding and interpretability. Experiments on three GREC benchmarks demonstrate that RC-GRPO attains superior localization accuracy while maintaining strong refusal capability.

Operational hazard analysis of aviation system operations must consider interactions among weather, ATC actions, airspace constraints, aircraft operations, and human factors - distinct from the functional hazard assessment applied at the aircraft-system level. We present an AI-assisted approach that generates candidate hazard scenarios from NASA's Aviation Safety Reporting System (ASRS). Given a target adverse outcome, it produces a structured hypothesis as categorical factors and a narrative scenario describing an operational event sequence consistent with the structure. Each scenario includes by a plausibility score from historical co-occurrence evidence and traceability to the most similar held-out ASRS reports. We then propose a hybrid variant, conditioning narrative generation on a structured hypothesis produced via evolutionary abduction, improving correctness and reducing variability. We evaluate multiple large language models, zero-shot versus few-shot prompting, and optional fine-tuning, measuring how prompting and model choice affect the validity and realism of the generated structures and narratives.

Federated adaptation of time-series foundation models (TSFMs) is attractive for building energy forecasting because meter data are private, distributed, and highly non-IID. However, a single parameter-sharing strategy is unlikely to serve all pretrained TSFMs or building clients: fully shared adapters can suppress building-specific temporal behavior, while fully local adaptation discards cross-building transfer. We propose a personalized federated sparse adaptation framework with a heterogeneous temporal mixture-of-experts (MoE) adapter placed after the pretrained TSFM representation. A sequence-level router maps each 168-hour context window to a top-$k$ subset of experts specialized for periodicity, long-range interactions, local variation, trend-residual structure, and multi-resolution behavior. We compare global FL, local training, and personalized FL variants with globally shared or client-private expert banks. Across 50 buildings and three TSFM backbones, personalization consistently outperforms Global FL-MoE and Local MoE, while the best sparse-adaptation strategy varies by backbone and metric. Routing behavior further reveals client-level expert specialization, expert concentration, and near-uniform routing across backbones, showing that federated TSFM adaptation should be both client-aware and backbone-aware.

Task-vector arithmetic offers a closed-form way to modify a model, yet its behavioral locality remains unclear in closed-loop robot control. We present a target-and-control audit of per-skill task-vector subtraction from multitask vision-language-action (VLA) policies. Across all ten LIBERO-Goal skills, subtraction produces three qualitatively different regimes: target-control separation for five skills, resistance for three, and global collapse for two. On held-out initial states, the five suppressible targets remain at 0% success; however, mean baseline-normalized control retention is only 52%, and each target-suppressing edit materially harms at least one nominally unrelated control. Additional Goal panels show separation across tested policies with continuous-regression, discrete-token, and flow-matching action heads, whereas we observe no clean separation on Spatial and control collapse on the tested Object and Long-horizon panels. Mean task-vector cosine does not account for this variation. A matched-norm control identifies a local sign asymmetry around one Goal anchor, while multi-vector outcomes vary with anchor and scale. Retain-aware gradient baselines provide data-dependent comparators but require removal-time data and optimization; subtraction is data- and gradient-free only at edit time, assuming precomputed expert deltas. Finally, a single-skill relearning probe is consistent with behavioral masking, not certified unlearning. These results characterize task-vector subtraction as a fast but brittle intervention and underscore the need for closed-loop target-and-control evaluation when assessing locality in embodied model editing.

We study distributionally robust PAC learning for the $0$--$1$-loss, where adversarial perturbations of the data distribution are constrained by a Cressie--Read divergence of order $k>1$ and radius $ρ\geq 0$. For hypothesis classes with VC dimension $d$, we establish realizable and agnostic sample-complexity bounds tight up to constant and logarithmic factors, respectively; ordinary empirical risk minimization attains both rates up to logarithmic factors. For target accuracy $\varepsilon\in(0,1)$ and confidence $δ\in(0,1)$, their respective orders are \[ \max\!\left\{\frac{1}{\varepsilon}, \frac{ρ^{\frac 1{k-1}}}{\varepsilon^{k_\star}} \right\}\cdot(d+\log δ^{-1}) \qquad\text{and}\qquad \max\!\left\{\frac{1}{\varepsilon^2}, \frac{ρ^{\frac1{k-1}}}{\varepsilon^{k_\star\vee 2}} \right\}\cdot(d+\log δ^{-1}), \] where $k_\star={k}/{(k-1)}$. For every fixed $ρ>0$, robustness changes the realizable $\varepsilon$-dependence from $\varepsilon^{-1}$ to $\varepsilon^{-k_\star}$ as $\varepsilon\downarrow0$. In the agnostic case, for $1<k<2$, robustness changes the $\varepsilon$-dependence from $\varepsilon^{-2}$ to $\varepsilon^{-k_\star}$, whereas for $k\geq2$ the exponent remains the classical $2$, with nontrivial $ρ$-dependence. Building on the known scalar reduction of robust $0$--$1$ risk to ordinary classification error, our analysis reveals a scale-sensitive interaction between the statistical estimation of classification error and its amplification by robustness, sharply explaining the transition in the agnostic rate. We extend the previously studied $χ^2$-divergence case to every Cressie--Read order $k>1$, close its upper--lower gaps, and recover standard PAC learning rates as $ρ\to0$, unlike previous bounds that fail to interpolate correctly in this limit.

Coding agents powered by large language models (LLMs) are increasingly adopted in software engineering (SWE) scenarios, capable of fixing a specific bug in large-scale codebase. However, existing SWE benchmarks typically assume that high-quality issue reports with detailed information are always available, which is easily violated in practice due to the complexity of report acquisition and curation. To address this, we introduce Active-SWE, a benchmark for evaluating coding agents on proactively discovering and fixing multiple bugs without report guidance, covering 1,663 tasks across six bug categories and eight languages. Beyond shifting the focus from existing reactive bug fixing to proactive bug fixing, Active-SWE enables a more in-depth evaluation by expanding the scope from fixing a specific recorded bug to multiple-bug fixing and potential bug discovery scenarios. To construct Active-SWE, we propose a novel difficulty-aware task formulation pipeline with a dual-track evaluation framework, facilitating comprehensive evaluation of proactive bug-fixing capability. Extensive experiments reveal that most state-of-the-art coding agents struggle with proactive bug-fixing tasks, demonstrating limited performance in locating and resolving recorded bugs, handling multiple bug fixing scenarios, and discovering valid potential bugs.

Papers 1-2 of the Kathleen series showed that a byte-level, attention-free architecture built from a wavetable encoder and multi-scale reverberant state can match strong baselines on classification at ~450-700K parameters, without pretraining. We ask whether the same ingredients can generate. (1) Scaling: on byte-level language modeling (WikiText-103, raw UTF-8, no tokenizer), the reverberant model beats a parameter-matched transformer at every dataset scale measured (2-512 MB), e.g. 1.84 vs 2.04 bits/byte at 512 MB with ~0.5M parameters; the transformer needs more than 512 MB to match what the attention-free model learns from 32 MB. (2) Measurement: we introduce FORM DISTANCE, a non-parametric, gaming-resistant instrument for "reads like text": nine statistical axes of human text define a reference cloud, and five constructed fakes are all rejected. (3) Generation: decoding policy dominates architecture -- widening the sampler halves the same model's distance (3.17 to 1.52), and a retrieval-augmented decoding scheme takes the frozen model further (1.52 to 1.14) with no training step involved; the ablation attributes the gain to the sparse phrase dose itself, not the selection gate. The gain has a sharp boundary condition: the phrases must come from the model's own training corpus -- a 40x larger foreign library helps not at all, an effect the attention twin shares, consistent with in-context integration being a capability of scale. We also report four architectural additions that did not help, and a computed lexicon reaching 94% of a learned table's top-1 accuracy at one fifth of the parameters. Everything runs offline; all experiments are reproducible on a free Kaggle T4.

Algorithmic recourse seeks to help individuals reverse unfavorable automated decisions by recommending actionable changes that achieve a desired outcome. As an individual usually has several distinct routes to a favorable decision, and different people can act on different ones, a recourse system should offer multiple realistic alternatives rather than one. Existing approaches formulate recourse as an optimization problem that constructs one or a small set of counterfactuals rather than modeling the underlying space of feasible solutions, and in practice each sacrifices diversity, plausibility, or feasibility to secure the others. We propose Tractable Recourse Distributions, a probabilistic framework that represents the space of feasible alternatives for a given factual instance as a probability distribution over favorable outcomes. For commonly used cost functions based on proximity and the number of feature changes, we show that this distribution admits an exact representation as a probabilistic circuit, obtained by exponentially tilting the circuit; each individual's distribution is therefore available in closed form, without retraining the model. Sampling from these distributions naturally produces diverse and plausible recourses, while the tilting parameters provide explicit control over their proximity and sparsity. Experiments on standard algorithmic recourse benchmark datasets demonstrate that the proposed framework attains diversity, plausibility, and feasibility simultaneously, while retaining sufficient probability mass over feasible counterfactuals for rejection sampling to be practical. A visual study on MNIST illustrates how the tilt strength trades proximity against validity.

Large Language Models (LLMs) have transformed computational linguistics and achieved remarkable performance across numerous natural language processing tasks, yet significant gaps persist in understanding how these systems process culturally embedded linguistic expressions. This paper introduces ProverbIT, a novel Italian benchmark comprising 100 multiple-choice questions designed to evaluate LLMs' ability to complete Italian proverbs. We assess 13 frontier models, including Large Reasoning Models (LRMs) and traditional LLMs, across three tasks: proverb completion, multiple-choice selection with correct answers, and multiple-choice selection without correct answers. Our evaluation reveals surprising results: while nearly all models demonstrate knowledge of the proverbs through successful completion tasks, performance drops dramatically when transitioning to multiple-choice formats without correct answers, with even state-of-the-art reasoning models showing substantial degradation. Through detailed Chain-of-Thought analysis of two LRMs, we uncover that models exhibit a strong bias toward selecting literal synonyms and frequently mention correct proverb endings during reasoning without successfully identifying their absence from the given options. These findings suggest that current LLMs rely heavily on memorized patterns rather than deeper semantic understanding of culturally grounded expressions, highlighting important limitations in their reasoning capabilities for figurative language comprehension.

Many social services assign scarce resources, such as housing assistance or hospital interventions, to people who arrive one at a time: each arrival must receive a decision immediately, and the long-run usage of every resource must stay within its capacity. We study how to learn such an assignment policy from logged observational data. The standard pipeline is decision-blind: fit one outcome model per arm by regression, price each capacitated resource from the fitted models, and assign each arrival the arm whose predicted outcome minus price is largest. We instead train the outcome models end-to-end, differentiating an off-policy estimate of the deployed policy's value through the dual prices themselves. We study two formulations: an exact nonconvex one, and a convex relaxation whose optimum always satisfies the capacity constraints in expectation and which is suboptimal by at most a term linear in the smoothing temperature and logarithmic in the number of arms. Every method is evaluated in a queueing simulation with resources replenished at their capacity rates. Across six datasets, the two end-to-end variants take the top slots on a deployment-adjusted value index at every delay cost, including zero; when capacities are binding, decision-blind baselines frequently violate them and incur much longer queueing delays. On the largest dataset, a hospital cohort of seventy thousand patients, end-to-end training also achieves significantly higher policy value, a margin that survives a capacity-matched neural baseline. Flexible decision-blind regression remains the stronger pure predictor where ground truth is measurable; end-to-end training is best suited to settings where resources are genuinely scarce and feasibility matters.

Selective inference (SI) provides statistically valid $p$-values for hypotheses selected by applying an algorithm to the data, correcting for the bias that arises when the same data are used both to select and to test a hypothesis. Developing an SI procedure for a new algorithm, however, has required an expert to derive, and then implement, the selection event, i.e., the conditions under which the hypothesis is selected. Repeating this specialized effort for every new algorithm is why exact SI has so far been available for only a narrow class. We propose AutoSI, a framework that removes this barrier in two ways. First, AutoSI constructs the selection event automatically from the algorithm's individual operations, so the user only writes the algorithm as ordinary NumPy-like code and derives nothing by hand. Second, AutoSI broadens the class of selection events SI can handle: existing exact methods are limited to selection events characterized by linear or quadratic inequalities in the data, whereas AutoSI covers any algorithm expressible through rational functions of the data (ratios of polynomials). We prove that the $p$-values computed by AutoSI are exactly valid in finite samples. We demonstrate AutoSI on three feature-selection methods, each written in a few dozen lines of code. One of these methods, the lasso with its tuning parameter selected by cross-validated $R^2$, cannot be handled within existing exact SI frameworks and is made possible by AutoSI. Experiments on synthetic and real datasets show that the resulting $p$-values control the type I error rate (i.e., the false positive rate) at the nominal level while retaining high power.

Cooperative multi-agent reinforcement learning often adds social terms to individual rewards, yet the scale of those terms is usually chosen by hand. We ask whether a guilt signal can instead be calibrated from human neural and behavioural data and transferred to artificial agents. Using the public SoDec responsibility fMRI dataset (40 participants), we fit a subject-fixed-effects regression of momentary-happiness changes on outcome-type counts and recover a guilt weight as the Partner-negative minus Social-negative contrast ($\hat{w}=1.118$, Cohen's $d=0.214$). We embed this weight in a two-agent Social Lottery environment and train independent Proximal Policy Optimization actor-critics under four shaping regimes: neurally calibrated, uniform constant, zero (selfish), and a unit-coefficient oracle. Across 1{,}000 evaluation episodes per condition, the calibrated agents track the human Social safe-choice rate most closely ($0.459$ vs.\ human $0.484$; $\mathrm{KL}=0.0012$), while the other three conditions deviate by one to three orders of magnitude in KL. Human neurobehavioural priors can therefore act as quantitative constraints on prosocial reward shaping.

Repository-level configuration artifacts allow developers to provide guidance for agentic AI coding tools, such as Claude Code, Gemini, etc. Although prior research has examined repository-shared context files that capture project-level instructions and conventions (e.g., AGENTS.md files), little is known about more task-oriented artifacts such as Agent Plans. We present an exploratory study of Agent Plans in open-source software repositories, examining how plan files are preserved, which development activities they support, and what information they provide to guide agent execution. We screened 36,710 GitHub repositories belonging to engineered software projects and identified 85 Markdown plan files from 10 repositories. Within this highly concentrated corpus, Agent Plans supported several kinds of software engineering work, including maintenance, design, construction, quality-related work, and process support. They also provided task-oriented execution guidance, most commonly through implementation steps, concrete files and locations, and testing and validation information. Overall, repository-preserved Agent Plans under these tool-specific directories appear to be a narrow but informative artifact for studying task intent and execution guidance in human-agent workflows.

Curvilinear structure analysis is an important and fundamental task in multimedia. However, the controllable generation of images with precise curvilinear structure objects remains an open challenge. To address this, we propose CSGen, a hierarchical multimodal diffusion model that synthesizes high-fidelity images precisely aligned with multiple control conditions. The CSGen is built upon three key innovations: 1) We construct a multi-domain and multimodal dataset, including over 24K samples from 5 domains and 7 different types of annotations, to train the unified generation model. 2) We propose a novel hierarchical progressive control strategy that decouples topology clues from visual context by a phased signal injection, mitigating semantic drift while ensuring the topological integrity of sparse structures. 3) We design a sparsity-aware loss re-weighting mechanism to address the extreme sparsity of curvilinear structures, significantly enhancing the attention on thin and fragile structures during optimization. Extensive experiments demonstrate that CSGen generates images with superior structure accuracy and visual realism, significantly improving downstream segmentation performance while maintaining robustness across diverse prompts. Our results confirm CSGen as a scalable, data-centric paradigm for the analysis of complex curvilinear structures in diverse multimedia applications. Code and dataset are available at https://github.com/ShanZard/CSGen.

Image mixup is a widely adopted data augmentation strategy, yet it is ill-suited for ordinal classification tasks such as medical disease grading, where labels encode a progression of severity. By indiscriminately blending disease-severity cues (ordinal) with appearance-level variation (non-ordinal), standard mixup produces samples that distort the very ordinal structure that underpins clinical severity grading. We introduce DisMix, an order-aware mixup framework for ordinal classification. DisMix disentangles ordinal and non-ordinal features via a dual-codebook VQ-VAE, allowing each subspace to be mixed independently: ordinal codes are interpolated to produce meaningful intermediate ranks, while non-ordinal codes are varied to introduce appearance diversity without corrupting the ordinal signal. Across four medical imaging datasets, DisMix shows the best aggregate performance among six image mixup baselines paired with six ordinal classifiers and remains effective under data scarcity and clinical grading variability.

Advanced materials discovery increasingly relies on machine learning and Bayesian optimization to explore large discrete design spaces under limited evaluation budgets. However, conventional Bayesian optimization (BO) can become inefficient as candidate spaces grow, often evaluating low-value regions before reaching informative areas. We propose an active-learning (AL)-guided adaptive search-space refinement framework combined with multi-objective BO to accelerate materials optimization while preserving Pareto-relevant regions. We evaluate the approach on CH4/N2 separation in covalent-organic frameworks and pressure-vessel design with material-direction stress components and thickness objectives. Results show that the AL-guided refinement reduces the candidate space by approximately half while preserving more than 99 percent of the original hypervolume. The reduced-space strategy improves early convergence and cumulative Pareto-front discovery from the BO, demonstrating efficient large-scale materials optimization across constrained autonomous materials discovery settings.

Large language models (LLMs) have recently shown strong performance on Theory of Mind (ToM) tests, prompting debate about the nature and validity of the underlying capabilities. At the same time, reasoning-oriented LLMs trained via reinforcement learning with verifiable rewards have demonstrated notable improvements across a range of benchmarks. In this work, we examine the behavior of such reasoning models in ToM tasks using novel adaptations of machine psychological experiments together with results from established benchmarks. We observe that reasoning models consistently exhibit increased robustness to prompt variations and task perturbations. Our analysis suggests these gains come at least partly from models being more robust at reaching the correct answer under prompt and task variation. We read this as evidence for a robustness-based account rather than for a new ToM-specific ability.

Generative AI is transforming legal translation by introducing opportunities alongside linguistic, technical, legal, ethical and cognitive risks. This chapter examines the implications of AI for professional legal translation and proposes an AI literacy framework tailored to the profession. It argues that AI does not change the fundamental objectives of legal translation but requires an extension of professional competence through AI literacy. The proposed framework comprises four mutually reinforcing dimensions, foundational, procedural, critical and strategic, and conceptualises AI literacy as a transversal component of legal translation competence that fosters digital resilience. It further discusses the pedagogical implications of this framework by proposing classroom activities designed to develop AI literacy in legal translator education, enabling future translators to integrate AI critically, responsibly and in accordance with professional standards.

Industrial recommendation strategy iteration heavily relies on large-scale A/B experimentation. Traditional tuning requires experts to repeatedly design strategies, configure experiments, analyze results, and adjust parameters, making the process labor-intensive and time-consuming. Meanwhile, valuable knowledge from historical experiments is often fragmented, making systematic reuse difficult through manual expert effort alone. Existing RAG agents partially alleviate this burden by retrieving prior strategies, but typically organize experience in a flat manner, overlooking the hierarchical relationships among business scenarios, recommendation stages, optimization objectives, and experimental contexts. This often results in mismatched retrieval and limited cross-scenario transfer, while preventing agents from continuously refining strategies and parameters through sequential A/B feedback. % To address these limitations, we propose A/B Agent, a closed-loop A/B agent for industrial recommendation strategy optimization. The framework comprises three tightly coupled core components: Historical Strategy Knowledge Organization, Autonomous Target-Aware Strategy Generation, and Experiment-Guided Strategy Self-Evolution. It organizes historical strategies into a hierarchical experience tree, retrieves transferable evidence through multi-path Tree-RAG to generate executable strategies, and continuously analyzes online A/B feedback to guide autonomous tuning and update the experience tree for self-evolution. Extensive offline and online evaluations demonstrate its effectiveness, including a 4.829% improvement in GMV in a real-world short-video e-commerce recommendation system while maintaining positive gains across all guardrail metrics.

Current neural networks for beat tracking generate invalid outputs, such as consecutive downbeats and erratic tempo changes, even when these are not present in the training data. Heavy post-processing techniques can alleviate these problems, but the original cause of this inconsistent behaviour remains unknown. We hypothesise that it stems from inadequate modelling of multiple plausible output beat grids, resulting in an invalid mixture of competing interpretations. We propose a masked diffusion approach that properly models multiple outputs and enables the model to build coherent predictions through iterative inference. We devise three modifications to standard masked diffusion that enable its application to beat tracking: independent masking of beats and downbeats during training and inference, a balanced masking scheduler for inference, and peak-picking across inference steps. Our approach reduces erratic behaviours and improves beat-tracking performance.

Heterogeneous language-model ensembles expand the space of candidate responses, yet they lack a principled criterion for when a newly generated answer should supersede an already supported one. We decouple candidate headroom from replacement authority, rendering the latter as an explicit, auditable object. Our proposed method, Agreement-Before-Diversity (ABD), is a frozen, label-free decision rule: an anchor answer is retained if two additional trusted samples corroborate it under a fixed equivalence relation; otherwise, it is replaced by a heterogeneous synthesis. For this gating mechanism, we prove two exact identities. The first shows that the accuracy gap relative to unconditional synthesis is determined jointly by the agreement coverage and the anchor's advantage on the protected subset. The second shows that the gap relative to never synthesizing reflects a contrast between authorized recovery and authorized destruction. Neither identity assumes independence or calibrated confidence, and the expected inference cost is approximately eight minus five times the coverage in number of calls. Under blind, exact-ID evaluation, ABD achieves 59.43% on the complete LiveCodeBench-v6 (vs. 52.57% for Single9 and 52.00% for HAC; n = 175) and 75.00% on an untouched GPQA-Diamond split (both controls at 72.78%; n = 180). Furthermore, these identities localize every aggregate difference to an enumerable protected stratum: no discordant items occur among the 3 protected cases on LiveCodeBench, where coverage bounds the gate's contribution to 1.71 points a priori; 13 versus 8 discordant cases among 132 on GPQA-Diamond; and 12 versus 0 among 71 under a frozen anchor perturbation. Diversity supplies potential; verification structure supplies authority.

An information theory-based framework is proposed in attempt to explain insistence on sameness in autism as an instance of a general behavior pattern in which an individual tries to reduce surprise and uncertainty. It offers a new definition of autism as an impairment in which cognitive functions are restricted to discrimination, memorization and prediction of tangible properties of the environment. An analogy between insistence on sameness and constrained minimization of the entropy metric is observed and examined for a set of assumptions that describe cognitive limitations of a person with autism. The metric is given by the formula $D_H(R, M) = H(R|M) + H(M|R)$, where $R$ represents sequences of random stimuli, $M$ is a memory that stores and retrieves them, and where $H(.|.)$ denotes their conditional entropies interpreted as surprise and uncertainty, respectively. It is first inferred that to minimize the metric an individual can learn about $R$ (and store that knowledge in $M$) or can restrict $R$ to the already known $M$. Then, it is concluded that insistence on sameness is a manifestation of the latter. Moreover, it is shown that the proposed framework: (1) Helps to quantify the concepts of surprise, uncertainty, sensory overload and deprivation, anxiety, comfort zone, disappointment, disorientation, pedantry, rigidness, observance or aberrant precision. (2) Leads to a list of guidelines for learning therapies and daily care routines, and allows them to be defined as optimization algorithms and implemented as programs for robotic live-in caregivers. (3) Can be validated with the help of a Turing test-like approach that requires no experiments involving individuals with autism. The framework-if positively validated-will provide formal foundations and design guidelines for therapies aimed at improving self-reliance of individuals with autism in basic activities of daily living.

Anomaly detection is a safety-critical machine learning problem with applications ranging from fraud detection to network intrusion prevention and industrial monitoring. Despite the large number of proposed anomaly detection algorithms, many novel methods claim state-of-the-art performance. However, many authors do so under benchmark settings that are not aligned with one another. This lack of comparability raises concerns regarding the reproducibility and reliability of anomaly detection benchmarks. In this work, we study the impact of common benchmarking choices on the stability of algorithm rankings. Using seven representative anomaly detection algorithms and 690 datasets from the OddBench benchmark suite, we analyze how rankings change under varying dataset selections, evaluation metrics, hyperparameter configurations, and random seeds. To quantify this effect, we introduce a rank instability metric measuring the variability of algorithm rankings across benchmark settings. Our results show that algorithm rankings in anomaly detection are highly unstable. In many cases, almost every competitive algorithm can appear as the best-performing method under some benchmark configuration. Among the studied factors, dataset selection and hyperparameter choice contribute most strongly to ranking uncertainty, while random seeds and evaluation metrics have a comparatively limited impact. We also observe that reliable benchmarking requires substantially larger and more diverse dataset collections than the ones commonly used in prior work.

Frontier coding models now match or exceed strong human reference points on programming benchmarks, yet benchmark success does not imply maintainable software. Prompt-driven "vibe coding" is additive: new branches, guards, and fallbacks accumulate faster than obsolete logic is removed. We study the inverse problem-how an Al system should remove code when execution-verification capacity is finite. We formulate redundant-code reduction as proposal scheduling: a ranker orders single-statement deletion candidates, an execution suite accepts the first candidate that passes, and a budget bounds how many candidates may be tested. Our central observation is that candidate order, not model confidence, is the control surface a deployment can reason about. DELSCOUT instantiates two schedules. Given representative target-domain validation, a five-slot budget spends three slots on deterministic shortest-first candidates and two on complementary learned candidates; across nine MBPP replications with 0.5B, 0.6B, and 8B rankers this raises verified-deletion coverage by 9.5% relative (+6.7 accepted tasks) while consuming slightly fewer verifier calls than the matched static baseline. Without such validation the same rankers can lose coverage under shift, so we instead evaluate the complete static prefix first and append learned candidates only afterwards; for a deterministic verifier this makes coverage and character reduction non-decreasing by construction, at a measured 4.8-62.5% increase in verifier calls. MBPP+ then erases the in-domain advantage, showing that scheduling governs search while the test suite alone governs what "preserving behavior" means. The result is an auditable division of labor: models widen the search for removable code, order bounds the damage a mis-ranked proposal can do, and execution retains authority over every committed deletion.

Stochastic gradient descent (SGD) optimization methods are the standard instruments for the training of deep neural networks (DNNs). In many relevant artificial intelligence (AI) systems - such as popular large language models (LLMs)-not the standard SGD scheme is used as the optimization method but instead suitable accelerated variants of SGD are employed. One of the most popular methods of such accelerated SGD variants is the momentum orthogonalized by Newton-Schulz (MUON) optimizer proposed by Jordan et al. in 2024. The MUON optimizer exploits the special matrix structure of the weight parameters in the training of the DNNs and, in its original form, employs five Newton-Schultz (NS) matrix steps in each MUON iteration. In this work we propose and study a generalized variant of the MUON optimizer involving an arbitrary number of generalized NS steps with polynomials of possibly arbitrary high degree. The considered optimizer covers MUON with the original NS polynomial as well as MUON combined with the recently proposed Polar Express method as special cases. For a simple class of stochastic optimization problems (SOPs) we show for almost every mini-batch size that MUON fails to converge to the solution of the SOP as the number of gradient steps converges to infinity. We also establish an error analysis for MUON with the generalized NS steps that provides convergence rates in terms of the number of gradient steps and in terms of the size of the mini-batch. We illustrate our general error analysis for MUON in the case of several concrete examples including quadratic stochastic optimization problems (SOPs) as well as $\ell_2$ regularized logistic regression for binary classification.

As robotic systems become increasingly complex, generating control firmware from structural description files has emerged as a promising paradigm for reducing development complexity and improving maintainability. Existing robot description formats naturally represent robotic systems as hierarchical tree structures, where devices are recursively composed into functional subsystems and eventually into the complete robot. However, such tree-structured organization also introduces structured data dependencies that affect perception-to-decision latency and, consequently, control performance. In this paper, we propose FineMote, a control firmware generation framework with a scheduling mechanism tailored for tree-structured device models. The framework objectifies heterogeneous low-level control logic and exposes unified scheduling units and execution entry points. Based on the resulting object hierarchy, the scheduling mechanism exploits compile-time information to statically determine execution order with minimal runtime overhead. We prove that the proposed mechanism satisfies deadline and precedence constraints, and further derive an upper bound on intra-tree decision latency. We implement the proposed framework and evaluate it on real robotic control platforms. The experimental results show improved timing behavior and runtime responsiveness, demonstrating the practical effectiveness of the proposed design.

Echo State Networks (ESNs) offer an efficient framework for temporal prediction, but their randomly initialized reservoirs are often over-parameterized and dynamically redundant. Existing pruning methods largely rely on static connectivity or activation statistics, which may overlook neurons that shape input-driven state transitions. We propose Dynamical Mode Pruning (DMP), a reservoir pruning method that ranks neurons by their contribution to dominant transition modes obtained from a trajectory-averaged Jacobian Gramian. DMP removes low-impact units and retrains only the readout. Experiments on chaotic and real-world time-series benchmarks show that DMP improves or preserves forecasting accuracy while reducing redundant reservoir components. Our results suggest that dynamical influence is a useful criterion for reservoir refinement beyond static structural importance alone.

Large language models (LLMs) are often asked whether something is absent from a record, list, or retrieved context. Yet non-observation licenses a negative answer only when evidence completely covers the query scope; otherwise, the answer should remain unknown. We call this completeness-sensitive negative reasoning. We introduce CROWN-QA, comprising CROWN-Synth, a controlled paired core that fixes the question and observed facts while varying only query-relative coverage, and CROWN-Real, a real-document contrast-set evaluation with controlled coverage variants. Across three LLM families, models show unstable closure judgments and substantial over-closure, failing to reliably distinguish a justified negative answer (Certified-Negative) from insufficient evidence (Unknown). The dominant CROWN-Synth failure is asymmetric: models often recognize implicitly complete evidence yet treat implicitly partial evidence as query-covering. Prompting redistributes errors between over- and under-closure rather than consistently resolving them. Structured certificate elicitation traces many errors to evidence-coverage mischaracterization. CROWN-Real shows that the core partial-coverage asymmetry persists on real-document content, while its strength and the balance between over- and under-closure vary by model, prompt, and source.

The growing deployment of delay-tolerant networks (DTNs) has made store-carry-forward (SCF) communication indispensable under sparse connectivity. However, intermittent contacts, finite buffers, and limited message time-to-live (TTL) often give rise to sparse delivery and congestion, leading to substantial end-to-end performance degradation. To address this challenge, this study explores the joint optimization of decentralized opportunistic routing and controllable unmanned aerial vehicle (UAV) flight, aiming to enlarge future contacts through discrete UAV headings while enabling per-node replication under contact-limited observations. Building upon this architecture, we study cooperative factored routing--UAV control under centralized training and decentralized execution (CTDE) and propose JUROR (Joint UAV flight and Opportunistic Routing, based on the proximal policy optimization (PPO) framework. In our design, we first cast the problem as a factored partially observable Markov decision process with sequential motion--routing coupling and a per-step team reward; subsequently, decentralized actors act on local observations while a training-time critic uses global statistics, and an optional multi-horizon hotspot predictor provides auxiliary supervision. Simulation results over four traffic modes demonstrate effective gains over PRoPHET and MaxProp, while retaining contact-limited decentralized execution.

EgoCross is a cross-domain egocentric video question answering benchmark designed to evaluate whether multimodal large language models can generalize beyond common daily-life scenarios. The first EgoCross Challenge was hosted at the Third EgoVis Workshop at CVPR 2026 and evaluated models on first-person videos from four target domains: surgery, industrial assembly, extreme sports, and animal perspectives. Each test example consists of an egocentric video clip, a question, and four candidate answers, from which the model must select the correct option. This technical report introduces the challenge task, benchmark resources, and two official Codabench tracks. The Source-Limited Track restricts participants to the official baseline model and a small support set, whereas the Open-Source Track permits broader choices of models and training data under rules that prohibit the manual construction of target-domain training data. In total, the challenge received more than 1,500 submissions from over 130 participants, with 19 teams participating in the Open-Source Track and 38 teams in the Source-Limited Track. We further present the official leaderboard results and summarize the winning solutions from both tracks. We hope that this report will serve as a useful technical reference for advancing cross-domain egocentric video understanding. All resources, including the challenge data, baseline implementation, and code released by the winning teams, are made publicly available.

Agentic systems have emerged as a promising paradigm for solving complex tasks by coordinating specialized LLM-based agents. However, most existing systems primarily optimize task success while giving limited consideration to execution efficiency under practical constraints such as executor capability and computational cost. Existing router-based methods have limited ability to reason over rich, evolving task contexts, multi-step dependencies, and intermediate execution feedback, and often generalize poorly to unseen executors. We propose EASy, a trainable agentic framework that jointly optimizes task performance and computational efficiency through reinforcement learning. EASy equips an LLM-based orchestrator with explicit knowledge of the capability and cost profiles of heterogeneous executors, enabling context-sensitive coordination beyond performance-only routing. It further introduces a milestone-plan-act workflow that decomposes complex tasks into manageable milestones, constructs dependency-aware execution graphs, assigns suitable executors, and parallelizes independent steps while adapting subsequent decisions to intermediate outcomes. To train the orchestrator, we develop a tree-structured rollout procedure that explores alternative milestone decompositions and execution plans, together with multi-component rewards that capture task correctness, execution efficiency, and trajectory completeness. Extensive experiments on mathematical reasoning, embodied decision-making, and deep research benchmarks show that EASy consistently achieves stronger performance-efficiency trade-offs than strong agentic baselines.

Multimodal large language models (MLLMs) have achieved significant success in speech-to-text translation (S2TT). However, when processing multilingual speech inputs, a single speech encoder shared across all languages suffers from the curse of multilinguality: languages at different resource levels compete for limited representation capacity, leading to strong high-resource performance but substantial degradation on low-resource speech. To address this problem and improve multilingual consistency, we propose MSRT, a novel framework built around a resource-aware Mixture of Speech Encoders (MoSE). MoSE uses an explicit language router to assign each utterance to an appropriate expert encoder. A frozen expert preserves high-resource language capabilities, while a trainable expert adapts to and specializes in medium- and low-resource languages. We further introduce a five-stage curriculum learning strategy that substantially reduces data dependence, requiring only 10 hours of paired S2TT data per language for effective alignment. We conduct extensive experiments on 45 languages, systematically evaluating all $45 \times 44$ translation directions. Our 4B-parameter model achieves state-of-the-art performance, outperforming substantially larger baselines. Empirical analyses show that MoSE improves high-, medium-, and low-resource languages simultaneously, with the largest gains on low-resource speech, thereby breaking the curse of multilinguality without compromising high-resource performance. To support future multilingual S2TT research, we release our code and models.

Humanitarian reports are long, noisy, and multi-topic, making it difficult to consolidate decision-relevant causal evidence. We present a ReliefWeb study (2000-2024) and a two-stage Large Language Model (LLM) pipeline that extracts structured intervention-outcome records with direction and strength attributes. Query-conditioned extraction restricts output to a specified intervention class, reducing retrieval-induced over-extraction, while snippet grounding links each relation to supporting text for auditability and classification. In an expert-annotated dataset of 100 reports, the best closed-source LLM achieved a weighted F1 score of 90.73% with strong cost-efficiency, while Llama-3.1-8B with supervised fine-tuning reached 94.15% weighted F1 score. We further propose context-preserving triangulation that aggregates strength-weighted evidence within disaster$\times$source cells, applies Laplace smoothing and equally weights cells to quantify cross-context convergence via a Level-of-Evidence score. Applied to cash assistance, food-related outcomes show strong positive convergence (LoE=0.865) and stable long-horizon trajectories.

Reliable physical reasoning from video requires understanding how objects move, interact, and respond to interventions. Existing vision-language models (VLMs) often struggle to interpret these dynamics and reason reliably about future and counterfactual outcomes. We introduce PhysMind, a training-free agentic framework that constructs one reusable, question-agnostic executable world per video. PhysMind recovers a temporally consistent dynamic scene through object segmentation, mesh reconstruction, and 6D pose tracking, then fits analytic continuous-time dynamics and latent physical parameters without unrolling a time-stepped simulator. Given a question, it inspects, continues, or edits the world and answers from the resulting trajectories and interactions. Relative to direct chain-of-thought (CoT) reasoning with the same VLM, PhysMind improves accuracy by 38.23 points on CLEVRER and 8.08 points on Physion++. On counterfactual questions, it exceeds the strongest evaluated VLM baseline, GPT-5.5, by 19.25 points.

Memory-augmented VLM agents act on persistent spatial knowledge, yet that knowledge silently goes stale as the environment changes. We ask what happens when an agent must reconcile a confident memory claim with a contradicting observation, and whether current models can catch the conflict before it becomes a safety-relevant mistake. Using a dynamic FrozenLake testbed, we pair a staleness-detection task with a downstream navigation task across three closed-source models and three open-weight VLMs under both text and image inputs (1,800 detection runs, and 12,000 text-mode navigation episodes over four LLM navigators at a shared 50-seed scale). Three findings emerge. First, text solvability does not imply visual grounding: models that flag stale entries reliably from text nonetheless span vision F1 from 0.887 down to 0.067 on the identical grids, and the weakest keeps making fluent, confident decisions that ignore the image. Second, consuming stale memory without an audit is a safety liability: in our primary GPT-4o setting, an agent that trusts raw memory dies more than twice as often as the same agent given no memory at all. Third, auditing helps but does not close the gap: a transparent read-time filter removes much of the safety cost in text mode, yet even oracle stale labels bring no further significant gain on the current grid size, and when visual auditing is unreliable, filtering yields no consistent benefit. Together these results frame spatial-memory staleness as a safety failure mode and isolate reliable visual grounding and action selection under memory--observation conflict as the central open challenges for memory-augmented agents.

Personalized LLMs with persistent memory are increasingly deployed, yet the faithfulness of their user models remains unexamined. We study over-inference (OI): the phenomenon where LLMs fabricate user attributes beyond what evidence supports. We introduce MirageBench, comprising 150 personas balanced across stereotypical, counter-stereotypical, and neutral profiles, 6 personalization tasks spanning an ``imagination gradient'', a four-way faithfulness taxonomy operationalized by an independent judge (validated against a blind human annotator on 400 claims: Cohen's kappa = 0.863 four-class, kappa = 0.900 binary), and a leaderboard of 12 models across 7 families on 143616 judged claims. We find that over-inference is pervasive: every one of the 12 models over-infers 35%--49% of its claims (cross-model mean 41.6%; claim-weighted 41.8%), with no model in this evaluation escaping it. Most strikingly, we surface a Self-Monitoring Inversion: at the model-selection level, models' self-assessed OI is negatively rank-correlated with their judge-measured OI (rho = -0.60, p = 0.044; exploratory, wide bootstrap CI [-0.90, +0.06], n = 12). The models that report the least over-inference tend to be flagged as fabricating the most, so self-reported confidence is a misleading signal for comparing models, even though within a single model self-audit still ranks that model's own claims moderately well (AUROC 0.58--0.83). We further show that OI is task-dependent (27%--59%) and that, in a multi-turn pilot, inferred attributes accumulate approximately linearly with little revision. MirageBench positions external verification, rather than model self-report, as a more reliable foundation for trustworthy personalization.

Hard prompt compression reduces long-context inference cost by independently scoring tokens, sentences, or chunks and retaining the highest-scoring units under a budget. We identify a structural failure in this procedure: independent selection can split dependent evidence pairs, retaining one member while deleting the other. When retained text contains an answer but deleted text defines the entity needed to interpret it, we call the result referential dangling. At a compression ratio of 0.30, Beaver, which ranks coherent chunks using Qwen3-0.6B embeddings, leaves the answer path incomplete in 34-54% of bridge examples across three multi-hop question answering datasets. On a shared HotpotQA bridge set, all six hard compressors we test exhibit dangling at rates up to 60%, and every document in LongBench-v2 Single-Document QA contains at least one dangling reference. On dangling examples evaluated with Qwen3-8B, reinserting the missing supporting paragraph while removing nonsupporting paragraphs to maintain the token budget improves accuracy by 29-34 percentage points (p < 0.0001), recovering at least 88% of the gap to contexts retaining both supporting paragraphs. Stronger answer models do not absorb the loss: on MuSiQue, GPT-5.5 is 8.8 points less accurate on compressed contexts than on contexts retaining both supporting paragraphs. Finally, we train a compact classifier to rank omitted sentences by whether they are needed to interpret retained text and reinsert the top-ranked candidates without support annotations at inference. On HotpotQA with Qwen3-8B, this automatic restoration improves accuracy by 4.7 points while changing the compression ratio only from 0.30 to 0.31. Hard compressors should optimize both relevance and referential completeness.

Event knowledge concerns who does what to whom. Psycholinguists use event-plausibility judgments to examine how this knowledge supports human language processing. To isolate plausibility effects, these studies require controlled event sets in which one event slot varies across plausibility levels while all other event features remain fixed. Constructing such sets manually is labor-intensive. We therefore introduce STRIVE, an LLM-based framework for jointly generating and evaluating controlled event sets crossing plausibility class (plausible vs. implausible) with intended classification difficulty (easy vs. hard). Given a verb, STRIVE constructs a shared event frame, then produces one event per condition by varying one slot while holding all others fixed. In experiments with six models across 60 verbs, GPT-5.1 produced high-quality sets only 16.7% of the time using the baseline generation prompt. Adding a global reasoning scratchpad and evaluator-guided refinement raised this rate to 75.0%. Greater reasoning effort also improved evaluator--human agreement. Nevertheless, events near the plausibility boundary remain most difficult. They elicit the greatest human disagreement, and the best evaluator reaches only 57% accuracy on the implausible-hard condition, indicating a need for human input. Overall, STRIVE offers a scalable approach to reducing manual effort by automating initial event-set generation and evaluation for psycholinguistic studies.

LLM-based search agents are widely used for information-seeking tasks, but their reliance on external tool returns introduces a critical security risk: web content retrieved during execution is untrusted, exposing agents to prompt injection and goal hijacking. Prior work on search-agent safety primarily focuses on static web-content injection, but modern agents issue follow-up queries and cross-check competing sources, so a single injected page is often diluted or rejected. We show that the channel delivering search and page observations is a fragile security boundary: beyond exposing the agent to a single poisoned page, a mediated search interface can repeatedly steer how the agent gathers evidence and forms its final answer. Under a constrained tool-intermediary threat model, appending only one controlled result per query can substantially increase attack success when the evidence is coordinated across the agent's trajectory. We study this setting with a strategy-driven long-horizon attack system and introduce Authority-Chain Hijack (ACH), an expert-refined strategy that turns isolated search-result and page-content manipulations into a coherent evidence chain across seemingly corroborating sources. ACH achieves the highest Overall ASR among all baselines, reaching 55.9% / 83.3% ASR / MaxN ASR on the full SafeSearch test split. We further introduce Trace-Guided Strategy Evolution (TGSE), which automatically improves attacker strategies from execution traces, replacing manual redesign with trace-driven refinement; its strongest single setting reaches 71.4% / 95.0% in held-out evaluation.

Agent skills are increasingly optimized by automated feedback loops, producing long structured artifacts whose internal value remains unclear. We study skill valuation: assigning credit to the internal units of a fixed skill, such as rules, examples, scripts, and heuristics, under a fixed agent and held-out task distribution. Skill valuation differs from data or prompt-span valuation because skill units are structured: they may depend on other units, belong to a document hierarchy, trigger agent behavior, and consume limited prompt context. We introduce SkillSV, a structure-aware Shapley-style framework for skill valuation. SkillSV compiles a skill into units, dependencies, and hierarchy, so that only valid counterfactual skills are evaluated. It uses paired deletion and length-neutral padding to separate content value from context cost, and estimates the resulting values with a rollout-budgeted estimator for noisy agent evaluations. On four agentic benchmarks, we assess the faithfulness, actionability, and explanation of SkillSV: it recovers unit interactions, preserves aggregate skill lift, and guides safe pruning and compression.

Predicting item difficulty from content can provide an initial estimate for newly developed questions before sufficient student responses are available. Existing approaches typically represent the question stem and answer choices as text. When mathematics items contain visual components, a common pipeline first textualizes that evidence and then applies a text predictor. We ask: how should visual evidence be represented for item difficulty prediction? We compare question text alone, visual textualization, which expresses visual evidence in language, and image-native modeling, which retains the original image. Using Eedi items with difficulty calibrated from student responses, we train large language models (LLMs) and vision-language models (VLMs) directly for difficulty regression. Both visual interfaces achieve the lowest point estimates, although the leading systems cannot be reliably ordered. Open-VLM textualization yields lower RMSE point estimates for all evaluated LLMs, while broader adaptation does so for all image-native VLMs. Test-time interventions show dependence on the paired full-item image, but do not isolate the additional visual component. The two visual interfaces also make partially complementary item-level errors and differ substantially in computational workflow. Thus, textualization should not be treated as the only practical interface: image-native modeling is a competitive alternative whose effectiveness depends on how the VLM is adapted.

Black-box language-model reliability is commonly pursued by sampling, prompting, voting, verifying, or iteratively revising individual answers. We ask a prior question: \emph{what determines whether a collection of black-box responses is recoverable at all?} We represent responses to typed transformations of a query as a \emph{relational response field} (RRF). Edge transports encode how valid responses must change under paraphrase, scaling, decomposition, refactoring, or other task symmetries; anchors encode independently trusted evidence such as execution or a verifier. For relation operator $D$, anchor operator $A$, and at most $k$ corrupted response nodes, we identify $γ_k(D,A)$ as the intrinsic difficulty of black-box response recovery. It is positive exactly when every $k$-node corruption is identifiable; it gives a deterministic stability bound proportional to $1/γ_k$; and a matching two-point minimax lower bound shows that no estimator can improve this dependence. Thus consistency is not truth: relation-only methods are blind to null directions, including shared hallucinations. We derive sparse field-repair algorithms while separating information-theoretic identifiability from the stronger null-space conditions required by convex optimization. Controlled theorem tests and black-box mathematics/code experiments evaluate four theory-fixed consequences: consistency--truth separation, anchor phase transitions, redundancy saturation, and cross-model, cross-task prediction of repair difficulty. The results support $γ_k(D,A)$ as a measurable property of a response-recovery instance, rather than a score attached to one repair heuristic.

Frontier LLMs are increasingly put to use on open-ended complex questions, different in nature from the ones they are typically evaluated on. We dedicate more than 4,000 human expert hours to evaluate a selection of six frontier LLMs on a member of this class of problems: EuroExec, our introduced human expert-based benchmark composed of 413 open-ended long-form European executive tasks authored by 47 vetted domain experts, each question drawn from experience in a real case. Every response is manually evaluated through a multi-attribute rubric, an item-specific checklist of requirements, and a preference rank ordering, extracting an aggregate metric "Solve Rate". The strongest model solves only 56.9% of tasks, while expert-written reference answers judged blindly are solved at near-ceiling levels and are preferred over every model response in 74% of direct rankings, placing frontier generative systems well below the professional standard of work they are already used for. We see that the best way to extract this kind of conclusion is by employing human evaluators, carefully checking their consistency through rigorous statistical analysis, and observe that automatic measurements also fall short when evaluating on this case of real-world open-ended problems with a subjective ground truth.

Multimodal large language model (MLLM) unlearning methods have been proposed to remove private, sensitive, or proprietary information from well-trained models. However, most existing MLLM unlearning methods are designed for one-shot requests and fail to adequately address continual scenarios, as repeatedly applying one-shot operations leads to cumulative utility degradation, unlearning rebound, and retention drift. We introduce Merging for Continual Unlearning (MCU), an approach that dynamically merges multiple one-shot unlearning adapters into a unified adapter upon receiving each new unlearning request.Through a leave-one-out merging analysis, we reveal that these unlearning adapters exhibit strong cross-task dependencies. Such dependencies have two contrasting effects: they can facilitate cross-task unlearning transferability, but they can also introduce severe interference that degrades unlearning effectiveness and compromises retained knowledge. To address this challenge, MCU projects the adapters into a shared representation space, preserves their dominant directions, suppresses over-concentrated coordinates, and reconfigures cross-task dependencies to mitigate interference while enhancing transferability. Experiments on ICU-Bench and MLLMU-Bench demonstrate that MCU achieves superior unlearning effectiveness while preserving both retained knowledge and general multimodal utility.

We study the problem of learning compact rule-based compressors for structured network traffic. Each packet is a record of header fields that are highly redundant within a flow, and a compressor is a small set of rules matching such records and replacing predictable fields with short codes. We cast rule learning as a two-stage problem: (i) an unsupervised structure-discovery stage that recursively partitions training packets using a normalized entropy-ratio criterion robust to small samples, and (ii) a constrained selection stage that uses dynamic programming to pick the rule subset maximizing expected compression gain under a hard budget on the number of installable rules. We instantiate the framework on Static Context Header Compression (SCHC), the IETF standard for rule-based header compression in constrained networks, and evaluate it on four real-world Internet-of-Things and 5G core-network datasets. Our method, Robust Entropy Clustering for Adaptive comPression (RECAP), surpasses expert-engineered rule sets with a small number of learned rules and removes the need for manual rule design.

Functional flow matching is posed on distributions of functions but implemented from finitely many coefficients or point values. Under scattered or adaptive refinement, the resulting conditioning sigma-algebras need not be nested, so martingale convergence does not justify the sensor limit. We prove strong $L^2$ convergence of finite conditional velocity targets for every strongly consistent sequence of finite-rank reconstructions, with quantitative bounds for orthogonal projections and a point-sensor extension through a regularity space. For learned flows, coupling directly to a population superposition path yields an end-to-end Wasserstein bound without assuming uniqueness of the population finite-dimensional ODE. We verify sensor-independent constants for a normalized quadrature neural operator, including globally Lipschitz activations through an explicit magnitude recurrence. A noncommuting trace-class Gaussian example gives boundary multiplier $0$ under projected restriction and $0.72$ under exact conditioning. A spatial regularity--cubature certificate closes the operator-realization term, a Bernstein argument gives a $\widetilde{O}(n^{-1})$ excess-risk term for fixed model dimension and envelopes, and an exactly realizable clipped Gaussian scaling specialization yields an explicit end-to-end rate.

Synthetic generation of Cognitive Behavioral Therapy (CBT) sessions is challenged by two competing demands: adhering to strict therapeutic structure while modeling the resistant, unpredictable behavior of real patients. Existing script-based methods fail to capture dynamic therapeutic interactions, while multi-agent approaches struggle to adhere to CBT's sequential structure; both suffer from sycophancy, producing overly compliant patients that misrepresent real clinical settings. In this work we introduce ODRA, a novel framework for synthesizing therapy dialogues through a Chain-of-Thought (CoT) strategy grounded in foundational CBT guidelines (Beck, 2020). ODRA further incorporates a resistance orchestrator to solve patient sycophancy, which employs steering techniques to elicit behaviors aligned with their resistance level. Automated and expert evaluations show that ODRA significantly outperforms existing methods across therapeutic skills, CBT alignment, and patient behavioral fidelity, with licensed psychologists preferring ODRA sessions across 12 of 13 clinical metrics. Furthermore, models fine-tuned on our dataset demonstrate superior therapeutic performance against both cooperative and resistant patients, validating that explicit resistance modeling in synthetic training data directly translates to downstream clinical robustness.

Hardcoded secrets in source code pose critical security vulnerabilities which can be easily exploited by malicious adversaries. Existing regex-based detection approaches suffer from fundamental limitations, as secrets often lack identifiable patterns, resulting in poor precision and recall. Recent studies have explored context-aware detection methods, as surrounding code can reveal the purpose of candidate strings. However, these methods confront three key challenges: (1) obfuscation robustness where models over-rely on easily obfuscated identifiers, (2) cross-language generalization difficulties due to uneven training data distribution, and (3) lengthy and noisy context that introduces excessive irrelevant tokens and slows inference. We observe that strings serve as a critical information source for code semantics, offering superior contextual density, obfuscation robustness, and language independence. Based on this insight, we propose StringGroup, a novel context extraction algorithm that mines strings surrounding potential secrets. By introducing a relatively simple modification to existing patterns that narrows the analysis specifically to string literals, the method achieves significant gains. With only 33.2% of the original context, it preserves over 80% of semantic information and significantly improves the signal-to-noise ratio for secret detection. We further design a context-aware secret detection tool, Secretron, based on StringGroup methods and Transformer model. Evaluation on the SecretBench dataset demonstrates high accuracy with 98.74% F1-score and strong robustness under obfuscation and cross-language scenarios, outperforming state-of-the-art LLM-based baselines. We deploy our tool in real-world environments and successfully detect 48 previously unknown secret keys from 26 applications, demonstrating the practical effectiveness of our approach.

Benchmarking machine unlearning methods is critical to understand whether sensitive knowledge is removed from large language models (LLMs) or not. Current unlearning benchmarks include mainly single-hop questions and a narrow set of multi-hop questions. Although effective, they still face two challenges. (1) Knowledge is not isolated, whereby diverse multi-hop reasoning paths can potentially induce knowledge leakage than normal queries. (2) Unlearning may be fragile: unlearned knowledge can be partially recovered through recovery attacks such as lightweight post-unlearning adaptation, making static evaluation insufficient. Therefore, in this paper, we introduce \unlearning as a novel benchmark to understand robust LLM knowledge removal across diverse reasoning paths and recovery attacks. We experiment with this benchmark on 3 models, 6 unlearning methods, and 2 carefully curated datasets. Results show that existing methods are vulnerable to multi-hop reasoning paths and recovery attacks. We further explore the trade-off among forget quality, robustness, and model utility for LLM unlearning.

Slice-based MLLMs leverage mature 2D encoders by representing 3D volumes as sequences of 2D slices. However, this slice-wise formulation produces thousands of visual tokens that burden the LLM backbone, many of which capture overlapping visual evidence across adjacent slices. To understand how effectively a growing visual token budget improves performance, we perform scaling analyses on two 3D medical VQA benchmarks and find diminishing returns: cost keeps rising while accuracy saturates, and improving in-plane resolution is more effective than adding slices at comparable budgets. The budget should therefore be allocated more selectively rather than simply enlarged, yet most token compression methods are designed for 2D images or videos, where redundancy arises from spatial layout or temporal motion rather than from near-duplicate content along the depth axis. We present CARVE, a training-free framework that compresses visual tokens prior to LLM inference and casts token reduction as budget-constrained 2.5D allocation. CARVE partitions the depth axis into coherent windows and allocates tokens non-uniformly according to normalized cross-slice evidence. Under a shared budget, CARVE builds spatial anchors on representative slices and retrieves locally varying evidence from the full volume, then merges remaining eligible tokens into nearby anchors within each window. Removing roughly 80% of the visual tokens on Hulu-Med-7B, CARVE leads all compression baselines on every AMOS-MM report-generation metric, with 6.2 points higher retention of full-token quality than the strongest baseline, and preserves 98.1% of full-token performance across three VQA benchmarks.

Background: Respiratory specialty care requires multimodal interpretation, longitudinal risk assessment, guideline-concordant intervention, and whole-course management, which are poorly represented by examination-oriented medical benchmarks. Objective: To develop RESPClinBench, a real-world scenario-based benchmark for respiratory clinical decision-making, and evaluate seven contemporary large language models across AECOPD-PIM and PNBIM. Methods: RESPClinBench cases were adapted from de-identified respiratory clinical data. Three attending-level respiratory physicians revised cases, reference answers, and atomic clinical-action points, while one senior respiratory specialist performed cross-review and final adjudication. AECOPD-PIM comprised 427 open-ended COPD cases, and PNBIM comprised 196 multimodal pulmonary nodule cases combining chest CT with structured clinical information. Seven models generated 4,361 responses through standardized API inference with temperature 0 and a maximum output length of 8192 tokens. An automated framework calculated the final score as the arithmetic mean of atomic-action recall and rubric-based LLM-as-a-Judge assessment. Results: Across 623 cases, the mean final score was 68.58. Qwen3.6-27B ranked first overall at 71.22, Qwen3.5-397B-A17B led PNBIM at 72.48, and Qwen3.6-27B led AECOPD-PIM at 71.11. Imaging hallucination and serious medical risk occurred in 31.85% and 8.16% of PNBIM responses; medication-safety risk and serious medical risk occurred in 26.93% and 1.44% of AECOPD-PIM responses. Conclusions: RESPClinBench identifies task-specific limitations in multimodal pulmonary nodule assessment and longitudinal COPD management. Combining explicit clinical-action coverage, holistic evaluation, and independent safety flags provides a clinically grounded basis for model selection and prospective validation.

Symbolic execution plays a critical role in software reliability, as they are used to find bugs, generate test cases, and provide correctness guarantees, particularly for safety-critical systems. Yet their own correctness is rarely subject to formal scrutiny, as it is typically established empirically by evaluating tool behavior across many programs. This leaves open the possibility that the tools themselves introduce unsoundness, potentially invalidating the verification results they produce and undermining the very guarantees they are meant to provide. In this paper, we address this gap by providing the formal treatment of symbolic execution with path-merging, an optimization that improves path explosion by summarizing branching code regions into disjunctive constraints rather than exploring each path independently. Specifically, we target Java Ranger, a path-merging tool for Java programs that progressively transforms imperative Java code toward the language of formal logic through a series of code transformations. We formalize each of these transformations and prove their soundness with respect to a simplified version of the Java concrete semantics, establishing that Java Ranger's path-merging process preserves program semantics.

Diffusion-based vision-language-action (VLA) policies can generate plausible actions even when their predictions are weakly grounded in the visual and language evidence defining the task. We introduce GUARD, a test-time failure detection method that measures this grounding without modifying the pretrained policy. GUARD estimates the influence of token-indexed entries in the final vision-language model key-value (KV) cache, constructs counterfactual caches by ablating salient KV entries, and compares their denoising responses with the original conditioning. Based on the comparison, we derive GUARD diagnostic stream including sensitivity, attention entropy, modality bias, and grounding efficiency, which are calibrated online and processed by a lightweight temporal classifier. We evaluate GUARD under task-held-out splits across five policy-benchmark settings, using Pi0, SmolVLA, and Alpamayo-1.5 on LIBERO, SimplerEnv, MetaWorld, and PhysicalAI-AV. GUARD achieves the best ROC-AUC on four of five unseen-task settings and ranks second on the remaining setting, improving the average unseen-task ROC-AUC by 5.73 percentage points over the strongest competing runtime monitor while remaining within 0.19 points of the best seen-task average. These results show that directly probing action-head dependence on multimodal evidence provides a transferable failure signal across policies, tasks, embodiments, and domains.

Vision-language systems combine images with retrieved text, but these sources can disagree or jointly fail to support an answer. Reliable models must identify the trustworthy source and abstain when neither is adequate. Existing post-training objectives score instances independently and therefore do not enforce coherent behavior under counterfactual evidence changes. We introduce CARGO-VL, a group-relative framework that optimizes matched variants covering aligned, image-correct, text-correct, and both-wrong (A/V/T/N) evidence states as one bundle. Its objective couples condition-wise correctness with transition rewards for answer invariance, source equivariance, and answer-to-abstention switching, while a primal-dual controller balances unsafe answers against excessive deferral. We also contribute XMC (eXtended Modal Conflict), a four-condition conflict training resource, and evaluate transfer on CMC-Bench and Modality-Bias. Across multiple seeds, CARGO-VL improves conflict handling, unsupported-answer avoidance, and modality balance over pointwise baselines. Ablations identify complementary benefits from relational transition signals and adaptive risk control, supporting counterfactual consistency as a practical objective for reliable multimodal evidence arbitration.

This technical report presents K-EXAONE 2.0, an open-weight multilingual foundation model developed by LG AI Research as a step in our effort toward global frontier-scale foundation models. Rather than training from scratch, we upcycle K-EXAONE and expand its architecture, yielding a Mixture-of-Experts (MoE) model with 750B total parameters and approximately 37B activated per token---more than three times the capacity of its predecessor. K-EXAONE 2.0 supports context lengths of up to 256K tokens and expands multilingual coverage from six to ten languages. Its training pipeline combines continual pre-training, difficulty-focused mid-training, and post-training to strengthen reasoning, agentic coding, multilingual capability, and safety grounded in Korean sociocultural contexts. Across nine evaluation categories selected to reflect the conditions of practical use, K-EXAONE 2.0 improves over K-EXAONE and remains competitive with open-weight models, showing its largest gains in agentic coding and long-context understanding and its clearest strengths in long-context retrieval and safety. Released under the Apache 2.0 license, K-EXAONE 2.0 enables the wider AI ecosystem to evaluate, deploy, adapt, and build upon it, while marking the beginning---rather than the endpoint---of our challenge toward the global frontier.

Vision-language models excel in many multimodal tasks but remain prone to a subtle yet impactful failure mode: they tend to overestimate dominant visual-textual cues while underestimating sparse but decision-critical contextual variables. This issue, which we term Contextual Variable Overestimation (CVE), becomes particularly evident in real-world applications such as predicting advertisement image preferences across diverse geographic markets. For instance, when a VLM is asked to choose between two product images tailored for different countries, it often defaults to a consistent output, ignoring ground-truth regional variations. This collapse occurs because pervasive high-volume signals, such as product attributes and dense image patches, overwhelm the few but critical tokens that encode market-specific context. To address CVE, we first collect a new multimodal dataset of real advertising creatives and their click-through performance across multiple countries. We then introduce GeoReward, a reward model designed to predict ad image preferences across diverse geographic markets. GeoReward integrates three purpose-built mechanisms: (1) Market-Aware Retrieval Augmentation, (2) Context-Guided Visual Modulation, (3) Selective Sensitivity Loss. Furthermore, we demonstrate how GeoReward can guide the fine-tuning of RL for a VLM to generate background designs for text-to-image models, producing market-aware advertising creatives. Experiments validate that our framework mitigates CVE and outperforms existing baselines. This work not only diagnoses a systematic bias in VLMs toward dominant perceptual features but also delivers a targeted solution for applications where sparse contextual variables govern decision-making.

Attention--Feed-Forward Network (FFN) Disaggregation (AFD) is emerging as a promising architecture for serving Mixture-of-Experts (MoE) language models. While existing AFD systems improve the efficiency of disaggregated execution, they leave a deployment question unanswered: under the same model, workload, time-per-output-token (TPOT) service-level objective (SLO), hardware budget, hardware catalog, and runtime capabilities, does AFD provide higher throughput than the best collocated deployment? Answering this question requires jointly optimizing hardware assignment and deployment organization for both architectures, making exhaustive provisioning prohibitively expensive. We present AFD-Ledger, an offline analytical provisioning system that independently provisions AFD and collocated deployments using an analytical execution model and an evaluation-bounded hardware search. Across deployment spaces where exhaustive provisioning is feasible, AFD-Ledger reduces complete deployment evaluations by 68.8%--83.5% while still recovering the globally optimal deployment. On three physical LongCat 2.0 deployments, it preserves the correct architecture decision while predicting AFD-to-collocated throughput within 6.6%--9.6% of measurement. Using this validated framework, we show that homogeneous AFD improves fixed-budget throughput in only a minority of the studied settings, heterogeneous AFD requires deployment-level hardware complementarity rather than heuristic device selection, and role-specific hardware improvements matter primarily when they enable better deployment organizations by crossing deployment capability--price boundaries.

Advanced wireless sensor networks powered by artificial intelligence are increasingly required for applications demanding continuous monitoring, autonomous operation, reliable communication, and fast decision support. This systematic review examines recent work from 2023 to 2026 on energy-efficient, AI-enabled wireless sensor networks (WSNs) in mission-critical environments, with particular focus on power electronics, and urban infrastructure systems. The authors synthesise a corpus of 50 DOI indexed studies satisfying inclusion criteria that received qualitative thematic coding and comparative analysis. Other references were only cited to provide historical, methodological, or technical context and were not included in the systematic review corpus. As such, our results show that AI can improve WSN energy behaviour through routing and clustering, edge AI, reinforcement learning, fuzzy logic, metaheuristic optimisation, and AI-based security. At the same time, energy efficiency cannot be treated as an isolated performance target. In mission-critical systems, security, latency, and reliability are closely interlinked requirements. The review concludes that future work should move away from optimising protocols in isolation, and instead focus on building lightweight, explainable, secure, and field-tested AI-driven WSN architectures suited to real operational environments.

Visual inputs in vision-language models (VLMs) are often encoded into substantially longer token sequences than text, making visual tokens a major bottleneck for efficient inference. Abundant recent methods address this bottleneck by scoring token importance and pruning low-scoring tokens in a single pass. However, one-shot scoring is insufficient because a token's prompt-relevant usefulness depends on the evidence already retained. Motivated by this insight, we introduce DIVE (Dynamic Iterative Visual Evidence Construction), a training-free framework that recasts visual-token pruning as dynamic evidence construction. DIVE repeatedly selects the remaining token with the highest residual-conditioned score, updates the visual and prompt residuals to discount the evidence already explained, and re-evaluates the remaining tokens. This select-update-re-evaluate process builds a retained set of complementary, prompt-relevant evidence. Experiments across eight image-understanding benchmarks show that DIVE consistently preserves performance across token budgets. With an 88.9% reduction in visual tokens, DIVE retains 98.2% of the uncompressed model's average performance. Code is available at https://github.com/Zhong-Chenchen/DIVE.git.

Despite rapid advances in large language models (LLMs), deploying and personalizing them on resource-constrained devices remains impractical due to high VRAM, time, and energy costs. Parameter-Efficient Fine-Tuning (PEFT) of Small Language Models (SLMs) offers a promising alternative, yet few studies compare PEFT methods across architectures using both general and personalization benchmarks while accounting for energy consumption. We compare five fine-tuning approaches (Full Fine-Tuning, LoRA, LoRA+, QLoRA, and BitFit) on four SLMs from two families (Transformer-based: TinyLlama-1.1B, Qwen3-1.7B; SSM-based: Mamba-1.4B, Mamba-2-1.3B) across three GLUE tasks (SST-2, QNLI, STS-B) and three LaMP personalization tasks (LaMP-1, LaMP-2, LaMP-3). Each configuration is evaluated with the energy-focused NetScore-E and the memory-focused NetScore-M, the two variants that reflect the constraints binding on-device deployment. Methods are selected with a strict energy-first rule (highest NetScore-E, ties broken by NetScore#). LoRA+ achieves the highest NetScore-E in 19 of 24 configurations and the highest NetScore-M in 13 of 24, and is the selected method in 18 of 24. QLoRA, available only for the Transformer models, cuts peak finetuning VRAM by up to 3.9x relative to LoRA and therefore takes the best NetScore-M in 5 of the 12 Transformer configurations, although its de-quantization overhead leaves it selected in only one of them once energy decides. BitFit and full fine-tuning are almost never competitive on either variant, and TinyLlama-1.1B leads the energy-focused NetScore-E on five of the six benchmarks and the memory-focused NetScore-M on four. These results show that compact SLMs paired with PEFT provide a practical, energy-aware path to personalized on-device deployment, with the optimal method set by the dominant constraint: LoRA+ for energy and QLoRA for memory.

Text-to-audio (TTA) generation has recently achieved remarkable progress in synthesizing realistic audio from natural language descriptions. However, determining whether generated audio faithfully satisfies complex textual instructions remains challenging. Existing benchmarks mainly rely on global similarity metrics, providing limited insight into fine-grained semantic failures. To address this limitation, we introduce \textbf{AudioScape-TTA}, a structured and complexity-aware benchmark for fine-grained TTA evaluation. AudioScape-TTA represents realistic soundscapes through modality-aware semantic structures and characterizes generation complexity using event density and structural complexity. Based on these annotations, we propose a rubric-based audio-grounded evaluation framework that verifies event realization, acoustic attributes, and speech content through fine-grained semantic criteria. The benchmark contains 2,258 audio-text pairs with 25,707 binary QA rubrics, enabling scalable and interpretable analysis of TTA systems. Experiments on 13 representative open-source TTA models reveal persistent limitations in fine-grained attribute control, speech-content preservation, and compositional soundscape generation. Human validation further demonstrates that our rubric-based evaluation achieves stronger alignment with human semantic judgments than conventional global similarity metrics.

Cloud-based language model services routinely process prompts containing sensitive information. Obfuscation-based defenses---including ObfusLM, SentinelLMs, TextObfuscator, and DPNR---mitigate this risk by transforming prompt representations before transmission, offering a lightweight alternative to cryptographic solutions. We show these defenses provide far less protection than previously believed. We present DeepInvert, a semi-supervised embedding inversion attack that recovers original tokens from obfuscated representations with higher accuracy than prior methods. The key insight is that unlabeled obfuscated embeddings retain exploitable semantic structure despite perturbation. DeepInvert combines supervised training on labeled shadow data with a novel unsupervised consistency objective over unlabeled target embeddings, alternating between the two via a mixed training pipeline. Defense-aware adaptations further extend the attack to diverse obfuscation mechanisms across encoder-based and autoregressive architectures. Experiments on nine defenses, five tasks, and four model architectures show that DeepInvert outperforms prior attacks on most defenses. Against ObfusLM, DeepInvert achieves 73.5\% top-1 token recovery versus 26.2\% for the previous best. Our results reveal a task-dependent tension: obfuscation schemes preserving enough signal for utility also retain sufficient structure for inversion, while schemes resisting inversion collapse utility. On simpler classification tasks, some DP-based defenses can maintain both. We call for a re-evaluation of this defense class.

Data-driven decision pipelines combining predictive machine learning models with downstream optimization software are increasingly used to make high-stakes operational decisions. Certifying the safety, fairness, and reliability of these decisions is essential, yet traditional scenario generation methods rely on repeated random testing, which becomes computationally prohibitive when failure events are rare and offers little insight into why failures occur. We present a framework for local violation certification designed specifically for linear decision pipelines under input uncertainty. We mathematically demonstrate that standard sampling methods fail efficiently for rare violations, motivating a direct structural approach. By analyzing the fixed decision boundary of a deployed pipeline, we show that the local risk of failure can be calculated directly in closed form using a single optimization solve. Furthermore, we introduce an exact sampling procedure and closed-form risk statistics that provide feature-level attributions (identifying which input characteristics contribute most to potential non-compliance) without requiring repetitive random trials or complex sampling algorithms. We demonstrate our approach on an economic power dispatch system subject to emissions regulations, delivering precise, auditable risk assessments at a fraction of the traditional computational cost.

The development of foundation models (FMs) is crucial for advancing endoscopic image analysis. However, existing endoscopy FMs mainly rely on self-supervised learning from uni-modal images or videos, overlooking the rich semantic knowledge contained in clinical reports. Furthermore, effectively leveraging these records is hindered by a fundamental modality gap: structured anatomical descriptions are not naturally mapped to specific frames within the high-redundancy, uncurated visual streams. In this paper, we present EndoVLM, a novel vision-language FM pre-trained on over 348K endoscopic examinations, each pairing a clinical report with its corresponding image collection. An Anatomy-Guided Sparse Pooling mechanism utilizes textual descriptions as queries to drive sparse attention, efficiently aggregating semantically salient frames into anatomy-specific visual representations across redundant image-sets. Next, a Progressive Semantic-Aware Alignment strategy models clinical taxonomy (anatomy and pathological status) via structured soft targets, bridging the gap from global patient-level matching to fine-grained localized alignment. Finally, a Semantic-Concentrated Masked Autoencoder is applied exclusively to these semantic-rich frames, integrating low-level visual precision with robust high-level semantic representation. Extensive experiments across various downstream tasks demonstrate that EndoVLM outperforms existing foundation models and remains competitive with task-specific methods. Remarkably, EndoVLM also exhibits robust zero-shot generalization capabilities, highlighting its potential for broader clinical application.

Time series in real-world applications are often generated by nonlinear dynamical systems, making accurate forecasting challenging. Existing approaches that explicitly model system dynamics typically rely on linear assumptions or Koopman-based linearizations, which may inadequately capture complex nonlinear behaviors and lead to error accumulation in long-horizon prediction. To address this limitation, we propose the Neural Bilinear Dynamical Model (NBDM), which models nonlinear system dynamics through a bilinear latent dynamical formulation. Specifically, NBDM leverages Koopman theory to lift the original nonlinear dynamics into a higher-dimensional latent space, where a bilinear dynamical model is constructed to characterize state evolution. To mitigate the approximation error introduced by bilinear representations, we further incorporate a parameterized error compensation term. Within this formulation, control inputs are explicitly integrated into the dynamics, using auxiliary variables when available and learned feedback signals otherwise. To handle scenarios with missing control inputs, we design a memory-enhanced controller that infers latent controls through multiplicative interactions between historical states and control signals. Experiments on five real-world datasets demonstrate that NBDM consistently outperforms competitive baselines in both given-control and missing-control settings, particularly for multi-step and long-horizon forecasting.

Prior work on LLM conformity largely measures discrete answer flips under verifiable labels. Open-ended revisions require a different measurement strategy because answer quality is graded, latent, and judged imperfectly. We introduce an experimental protocol implemented across a pooled main peer-condition corpus and separately constructed decomposition corpora, allowing us to separate ordinary re-answering, candidate-content exposure, a bundled peer-presentation residual, and directional judge sensitivity to visible peer context. Across four open-weight generators and three benchmarks, all-wrong peer input produces the lowest-quality revisions in every generator-dataset cell. Blind and informed ratings of identical answers also differ by evaluator: one judge shifts toward the peer-endorsed position, two shift away, one is approximately neutral, and GPT-4o and GPT-5.4-mini audits are likewise non-neutral. Finally, an anchor audit shows that terse correct anchors can be misread often enough to destabilize the latent scale unless calibration is checked explicitly. These results support four conclusions: flip rates are insufficient as a complete measure of open-ended conformity, wrong peers harm open-ended revision, evaluators are not neutral, and anchor calibration is necessary.

The quantitative analysis of 3D neuronal morphologies requires capturing both graph topology and spatial geometry. Current message-passing Graph Neural Networks (GNNs) are bounded by the 1-Weisfeiler-Lehman (1-WL) test, limiting their ability to capture cycles induced by spatial proximities. To address this, we propose a training-free geometric prior based on tropical algebraic geometry. We apply the recently established tropical Abel-Jacobi transform and polarization distances to machine learning on tree-structured data. We introduce a structural transformation pipeline, comprising cycle space augmentation and quotient space construction, to convert spatial trees into cyclic metric graphs suitable for embedding into the Tropical Jacobian. Computing exact tropical polarization distances requires solving the NP-Hard Closest Vector Problem (CVP) on integer lattices. Instead of relying on explicit approximations with quantization errors (e.g., Babai's rounding), we adopt a continuous relaxation on the universal cover of the Albanese torus. We show that the discrete Arakelov-Green measure, computed in closed form via the graph Laplacian's generalized inverse, decomposes exactly into the intrinsic path metric minus the unquantized polarization distance on this cover, avoiding integer lattice searches. This metric yields two descriptors: eigenvectors provide node-level structural coordinates, and the permutation-invariant eigenvalue spectrum provides a graph-level signature. On the BREC benchmark, the eigenvector formulation demonstrates expressivity beyond the 1-WL limit. On 3D morphology datasets (ACT-4, JML-4, BIL-6), the spectrum seamlessly integrates into standard architectures (VAEs, GNNs, Tree-LSTMs) without additional trainable parameters, outperforming explicit lattice approximations and improving classification accuracy over existing spatial models.

Developers often need to adapt into their projects the code generated from LLMs or code snippets from online forums. However, integrating them into an existing repository remains challenging in a manual process. A successful integration typically requires more than copying code as a user must produce correct adapting changes at a designated location in the target repository. We formalize this as the code adaptation problem: given a snippet, functional intent, a target repository, and an adaptation location, generate a patch that adapts the snippet into the repository. We present AdaptAgent, a multi-agent, domain-guided reasoning framework for code adaptation. Rather than relying on single-shot prompting, AdaptAgent decomposes adaptation into specialized agents that communicate via typed artifacts: an Intent Summarizer extracts adaptation goals from Q&A text; a Policy Agent derives domain policies from six adaptation categories; a Domain Planner generates a self-ordered plan; a Context Miner distills sibling-method semantics from the target codebase; and a Code Adapter realizes the plan as a minimal unified diff, iteratively refined using a compiler-based Verifier. This division of labor enables robust, policy-aligned adaptations and supports adapting code snippets into a project. On a real-world dataset, AdaptAgent outperforms strong baselines in semantic correctness and produces patches that mirror developers' actual adaptation patterns. Our ablation study shows each agent's necessity, especially planning for code-hardening and exception-handling, and intent for logic customization.

Agentic AI is emerging in datacenters, but its architectural implications remain unexplored. We organize agentic workflows in a taxonomy and present its first architectural characterization with a production study at Microsoft Azure and a controlled study of open-source frameworks. We show that agentic execution is fragmented and heterogeneous. Requests expand into a workflow of LLM inferences, tool invocations, and orchestration decisions that repeatedly cross the CPU-GPU boundary. Our taxonomy explains how this fragmentation turns into resource demand. As orchestration and tools run on the host, the CPU sits on the critical path. Execution structure sets the load over time, which stays low with sudden spikes. Model composition sets how evenly the workflow uses the GPUs. Diversity in tasks and tools widens this range even further. These characteristics expose architectural mismatches of conventional uniform servers. Fragmented execution strands CPU and GPU capacity despite bursty demand. Different software roles make homogeneous CPU provisioning inefficient. Finally, multiplexing many agents onto shared cores degrades microarchitectural locality. Guided by our findings, we derive implications for agentic servers and examine them through Agora, our prototype for commodity servers. Agora dynamically harvests idle CPU cores for co-located throughput work, while protecting agentic tail latency against tool spikes. It oversubscribes GPU memory by placing more agents on each GPU, prefetching the next agent's state to hide swap latency. To match the machine to the heterogeneous roles, Agora pools cores by role and applies affinity-aware scheduling to restore locality. It automatically tunes mechanisms to the workload. Agora improves utilization and server throughput while preserving agent tail latency. Our insights also identify key directions for future server architectures for agentic AI.

As "AI Scientists" emerge to drive research via the Model Context Protocol (MCP), systems relying on ephemeral scripts will fail. The sheer scale of stateful, interconnected evidence requires a machine-walkable warranty grounded in a purpose-built database architecture. Eigenius is an open-source, typed knowledge-graph DBMS built on a single premise: answering the audit question ("what do you know, and what is your warranty?") requires a unified kernel. By tightly coupling the type system, storage engine, and integration protocol, Eigenius turns data provenance into a structural invariant rather than a property reconstructed across subsystem boundaries. The kernel rests on three pillars: a dependent type theory woven through the core, institutions acting as strongly typed integration boundaries, and a content-addressed immutable storage layer. On this foundation, epistemic status (declared/observed/derived/verified) is enforced as a strict commit-time invariant. Cross-system translations (comorphisms) are checked at commit and materialized directly into the graph as durable, first-class resources. To eliminate O(N^2) polystore bottlenecks, shared on-chain intermediate representations (IRs) collapse multi-system translations to identity. Crucially, this architecture unifies both domains of scientific epistemology: it relies on justification logic for empirical science, while embedding a fast, in-process term checker to safely evaluate formal mathematical proofs (via Lean 4) without IPC overhead. In an end-to-end recomputation of a published Nature study from fragile scripts to a materialized evidence graph, all 52 derived conclusions hold from pinned data, surfacing four machine-checked discrepancies in the original study.

One of the most challenging problems entertainment live-streaming services face in recommendation systems is that user behaviors are sparse and delayed, and interaction data exhibits bias for different user segments. Unlike e-commerce applications where user actions follow linear sequences, live-streaming viewers engage in multiple concurrent behaviors of watching, chatting, following, and spending, each occurring with varying delays. We address these challenges through three key contributions: 1) a delayed window approach that extends feedback collection beyond immediate responses, 2) a multi-model architecture that combines fresh and delayed signals, and a segment-aware targeting module that optimizes ranking scores differently across user lifecycle stages, and 3) Multi-gate Mixture-of-Experts (MMoE) integration that jointly models correlated targets while reducing model parameters by 41.9% compared to independent models. Online A/B testing demonstrates significant improvements, including a +0.09% increase in Daily Active Viewers (DAV), generating millions more annual active viewer days, and +0.56% increase in highly engaged viewers' capped Average Revenue Per User (ARPU). Viewer-segment targeting achieved an additional +0.15% DAV improvement for newer and less engaged viewers, while MMoE enhancement added +0.08% overall DAV and +0.27% new follows. The proposed system processes ranking requests with low latency, providing a scalable approach for balancing multiple business objectives across diverse user populations. In addition, we tested the multi-model architecture on the Twitch mobile live feed and achieved a +1.12% increase in positive user-channel interactions (clicks, follows, and likes), demonstrating applicability beyond the primary use case.

Mixture-of-experts vision-language models (MoE-VLMs) increase model capacity with sparse expert activation, yet deployment requires storing the full expert pool. Training-free expert merging reduces this burden, and many routing-based methods aggregate routing statistics across all tokens to determine merge compatibility. However, MoE-VLM inference is phase-structured: image-context tokens carry visual content, question tokens specify the query, and answer tokens produce the output, with different counts and routing distributions. Because image-context tokens are far more numerous, global aggregation can overemphasize image-context processing and obscure phase-conditioned expert roles, making experts serving different phases appear interchangeable and degrading model performance. We therefore argue that MoE-VLM expert merging should preserve phase-conditioned expert roles, judging compatibility by how experts serve different phases rather than globally aggregated routing statistics. Based on this view, we propose RoleMerge, a training-free method that constructs each expert's Routing Role Profile (RRP) from phase-normalized routing statistics, capturing its relative phase preference. Guided by expert-phase information loss, RoleMerge merges experts with compatible profiles and their corresponding router entries while preserving answer-decoding expert distinctions. Experiments on three models and multiple benchmarks show that RoleMerge preserves more of the full model's performance than alternative expert-merging methods at matched expert-retention ratios, with relative improvements of up to 9.6 percent in six-task macro-average performance. These results validate phase-conditioned expert roles as a more effective basis than global routing aggregation for MoE-VLM expert merging.

Online HD map construction is critical to prediction and planning in autonomous driving. We find that existing physical attacks against online map construction are limited by a cross-boundary compensation effect: after the target boundary is perturbed, another visible boundary may retain sufficient geometric cues for the model to recover the original road geometry. Based on this observation, we propose TwinIR, a new mechanism-guided physical attack methodology for online map construction. TwinIR jointly optimizes attack effectiveness and point sparsity, seeking the minimum number of attack points needed to suppress compensating geometric cues from surrounding boundaries. To reduce the perceptibility of multi-point attacks, TwinIR models camera responses to near-infrared illumination and maps optimized attack points to feasible physical placements, producing camera-visible interference with minimal visible-spectrum changes. Experiments on nuScenes across state-of-the-art online map construction models show that TwinIR reduces mAP by 8.18-8.96 percentage points under RSA and 2.84-5.62 points under ETA, while increasing the unreachable-goal rate by 25-28 points and the unsafe-planned-trajectory rate by 19-20 points over clean inputs. These attacks are also validated on a real-world testbed AV, where TwinIR successfully induces both road straightening and early-turn deformations while remaining inconspicuous in full-color views.

High-resolution pixels and crop or zoom tools give multimodal large language models the ability to inspect an image, but they do not provide a reliable task-conditioned policy for deciding where to inspect. Q-CueGraph makes this decision explicit. It maps a question and an image representation to budgeted, coordinate-level observations for a frozen reader. Text-rich images use a reusable OCR/layout graph; natural-image search instantiates query-conditioned visual nodes behind the same selection, composition, and budgeting interface. Optional utility refinement learns which candidate crops the frozen reader can use from training-answer correctness, without region-box supervision. With a frozen Qwen2.5-VL-7B reader, Q-CueGraph reaches 0.833 accuracy on V*Bench versus 0.696 for full-image inference from a 19% image-area budget, and reaches 92% of full-image ANLS on InfographicVQA from about half the image area. Across six benchmarks, explicit observation is most valuable when evidence is localizable, the question discriminates its location, and resolution limits full-image reading.

We give a negative solution to MAIS-O60. We first construct an example in which an initially active ReLU neuron becomes completely inactive in finite time and thereafter remains frozen at a limit whose Fourier energy is equally distributed among all nonzero real frequency classes. The counterexample holds on an open set of initial conditions and therefore occurs with positive probability under Gaussian initialization. An appendix prepared by GPT-5.6 Sol strengthens the counterexample by showing that the same failure can occur for every Clarke trajectory from an open set of initial conditions, under the convention $\mathrm{ReLU}'(0)=0$, for smooth dead-zone approximations of ReLU, and for fixed-step full-batch gradient descent. Thus, single-frequency alignment is not a general consequence of training a single neuron on modular addition.

Diffusion transformers deliver strong image generation, but their training cost grows superlinearly with resolution. Recent work justifies training or sampling at reduced resolution on a spectral premise: at high noise, a downscaled latent preserves almost the full surviving signal. Whether a downscaled step also preserves the native training gradient signal, however, has remained unresolved. We reduce how that signal changes under downscaling to two terms: a noise-dependent term governed by the downscale ratio, which decays at high noise as the spectral premise predicts, and a σ-independent floor governed by the target grid's absolute token count, carried by the compute graph itself and removed by no noise level. The measured (route, σ) map corroborates the account and uncovers structure the spectral picture cannot express: on the 1024->768 route, a window (0.65 < σ< 0.95), predicted by no spectral criterion at any tolerance, where the downscaled gradient stays within a small margin of the native one. Training LoRA adapters with downscaled steps restricted to the routes and noise windows the map validates reduces training time by 14.6% at a fixed step budget while remaining near-native in weight space. Code is available at https://github.com/sorryhyun/anima_lora.

Large language models (LLMs) often generate inaccurate answers due to their reliance on static internal knowledge. Retrieval-augmented generation (RAG) addresses this limitation by integrating external knowledge and excelling at single-hop queries. However, it struggles with multi-hop questions that require cross-document reasoning. Existing methods, such as graph structured RAG or question decomposition, often lack dynamic decomposition and effective filtering, which leads to lower efficiency and accuracy. To overcome these limitations, we propose Dynamic Decomposition and Filtering for Multi-Hop Reasoning-Augmented Generation (D2F-ReAG), a novel paradigm that adaptively controls reasoning depth by judging the reliability of the root-level reasoning. If the root reasoning is reliable, the model directly generates the answer. Otherwise, the question is logically decomposed into sub-questions, and the verified reasoning derived from these sub-questions is used to refine the root reasoning. Experiments on three multi-hop benchmarks demonstrate the effectiveness of our method in handling complex multi-hop questions.

Robustness to natural corruptions remains a fundamental challenge for deep neural networks. In this paper, we identify a robustness fading phenomenon where shallow layers spontaneously develop robust representations and flat loss landscapes in early training, yet these properties are not preserved during standard convergence. To address this, we propose a framework that performs strategic interventions on training dynamics to stabilize the empirically identified early-emergent robust priors. Our approach includes two parameter-free strategies: Early-Phase Stabilization~(EPS) and Asymmetric Weight Reversion~(AWR), which stabilize or recover robust shallow configurations without modifying the model architecture or introducing learnable parameters. Extensive experiments demonstrate the efficacy of our framework across various benchmarks and architectures, yielding significant gains in downstream transfer, dynamic adaptation, and diverse computer vision applications.

Large language models (LLMs) have made notable progress in code generation, but they still struggle on challenging tasks that require sophisticated algorithms or complex implementations. Recent methods increasingly use code execution as feedback, especially in self-correction pipelines that construct verification signals from generated code. However, these pipelines often depend on supervision signals whose reliability is unknown, which can introduce misleading feedback, unnecessary revisions, and incorrect final answers. To address this issue, we propose ExeCRE, an Execution-Consistency guided code Reliability Estimation framework. Instead of judging candidate code by tests or LLM feedback, ExeCRE estimates code reliability by statistically analyzing consistency patterns in execution outputs over a large number of randomly generated inputs. It collects execution outputs over generated inputs, projects them into consistency signals, and applies the Dawid-Skene model to infer latent code reliability. We integrate ExeCRE into self-correction for code generation. Experiments show that ExeCRE consistently improves both effectiveness and stability, while substantially reducing misleading correction signals. Under GPT-5.2 on LiveCodeBench, the average number of misleading feedback cases on already correct code drops from 113.2 with a representative self-correction baseline to 14.0 with ExeCRE. As an additional study, we apply the same reliability estimation strategy to code-based mathematical reasoning and observe similar benefits. These results suggest that ExeCRE enables more reliable use of generated code in execution-based pipelines.

We present MERaLiON-GR, a speech gender recognition system that performs binary classification (female / male) on English and Southeast Asian (SEA) languages. The model finetunes MERaLiON-SpeechEncoder-2, a large conformer based transformer pre-trained on a broad speech corpus, and applies parameter efficient fine-tuning via Low-Rank Adaptation (LoRA) to adapt the encoder to the gender recognition task, and appends a multi-scale ECAPA-TDNN down stream network with attention pooling and a lightweight linear classifier. Extensive evaluations across multilingual Singaporean and Southeast Asian languages (English, Chinese, Malay, Tamil, Thai, Vietnamese, Indonesian, and Khmer) show that MERaLiON-GR consistently surpasses the state-of-the-art gender recognition model Vox-Profile and a large Audio-LLM, in both full-utterance and segment level evaluation modes. The results underscore the value of dedicated speech models in achieving accurate paralinguistic understanding and strong cross-lingual generalization.

On two-sided content platforms, symmetric two-sided isolation (assigning matched fractions of creators and viewers to isolated treatment and control submarkets) is widely used for creator-side and cold-start experiments because it removes cross-arm marketplace interference. Isolation, however, thins each viewer's candidate catalog, and intuition suggests the resulting engagement cost should fade as the platform grows: a small fraction of a vast catalog is still vast. We show that, in an order-statistics model of engagement, whether this intuition holds depends on the upper tail of match quality. Extreme-value theory yields tail-class loss laws with a sharp dichotomy: for light or bounded tails the loss vanishes as the candidate pool grows, whereas under heavy tails it converges to a size-independent constant, so expanding the candidate pool, even by orders of magnitude, does not asymptotically eliminate the cost. Evidence from two production experiments on a platform with millions of active creators is consistent with this picture: a pure A/A traffic sweep reveals a measurable, depth-graded engagement cost; a one-sided catalog ablation independently shows that per-viewer thinning contributes to the loss; and a tail index calibrated on the small exploration pool predicts an effect consistent with the one observed in the far larger full-catalog ablation. Isolation thus carries a price that experimenters should budget for, like any other cost. We give practitioners a preflight procedure that estimates it before launch, sizes traffic accordingly, and recommends a fallback design when the predicted cost exceeds a chosen tolerance.

Vision-language-action (VLA) models have emerged as a key component in embodied AI. Among existing approaches, diffusion-based VLA models achieve superior motion quality and generalization. However, diffusion-based VLA models are compute-intensive and must run at high control frequency, e.g., 50-200 Hz. Thus, it imposes strict latency and energy constraints on edge devices. In this work, we present Deltoris, an algorithm-hardware co-design framework for efficient diffusion-based VLA inference. First, we exploit the temporal similarity of consecutive inputs and propose a \textit{temporal-aware bit-sparsity} algorithm that computes only the differences between consecutive inputs, eliminating redundant bit-level operations. To further address the extra off-chip traffic introduced by our algorithm, we propose a \textit{speculative inference} technique, which amortizes data loading across multiple control steps. Lastly, to support these techniques, we co-design a dedicated accelerator with customized 1D systolic bit-serial PE arrays that eliminate PE workload imbalance. Our evaluation shows that Deltoris achieves up to 34.2$\times$ speedup over mobile GPUs and 6.1$\times$ over prior accelerators, while maintaining comparable accuracy.

We introduce Predictive State Retrieval (PSR), a task in which a model observes a short video prefix and a temporal question about an object's future state, then retrieves instances from other videos or images that depict that state. Unlike action anticipation, which predicts a label, moment retrieval, which localizes an observed event within a video, or video generation, which synthesizes pixels, PSR combines anticipation with cross-instance retrieval across multiple temporal horizons. We construct a benchmark from four datasets with graded, human-validated ground truth, difficulty tiers, and an oracle ceiling. We also propose LFTR, a lightweight retriever with frozen encoders that predicts a question- and horizon-conditioned future latent and matches it in complementary semantic and visual spaces. A ceiling decomposition reveals a clear bottleneck: the true future state is highly retrievable once specified, whereas every predictor we evaluate, including a large multimodal language model with access to the prefix frames, remains far below the oracle. Thus, forecasting rather than perception is the central learnable challenge. LFTR narrows this gap at substantially lower inference cost, and ablations attribute its gains to cross-space fusion and hard-negative training rather than latent rollout. We release the benchmark, code, and evaluation scripts.

Incentivized advertising allocates monetary or virtual rewards to drive user engagement, where a key challenge is optimizing continuous incentive magnitudes under strict global constraints. This problem is complicated by high-frequency interactions, delayed feedback, and non-Markovian user dynamics such as fatigue, which limit the effectiveness of existing uplift modeling and constrained reinforcement learning approaches. To address these challenges, we propose GOAL, a constraint-aware generative framework that formulates incentive allocation as a conditional sequence generation problem. GOAL directly generates incentive magnitudes conditioned on user histories and system-level global pressure, and integrates a hierarchical causal state encoder to capture both local behavioral dynamics and long-range dependencies. To enable flexible constraint control, we introduce \textbf{S}afe \textbf{C}onstrained \textbf{P}olicy \textbf{O}ptimization (SCPO), which learns a single generative policy that generalizes across a spectrum of ROI constraints without retraining. Experiments on large-scale real-world data and a synthetic fatigue-aware environment show that GOAL improves long-term revenue and user retention while substantially reducing ROI violation rates compared to strong baselines.

On-policy distillation (OPD) provides dense teacher supervision on student-generated trajectories, but standard reverse-KL training can assign insufficient probability to other plausible continuations. Teacher entropy alone does not reveal whether uncertainty is concentrated among a few plausible next tokens or dispersed over a long probability tail, nor whether the student already represents those candidates well. Moreover, local teacher probabilities may not predict downstream success. We introduce Sparse Probing and Outcome-calibrated Targets OPD (SPOT), which addresses two coupled decisions, where to probe and what to distill, through an acquisition--exploration--exploitation procedure. During acquisition, a position-level score combines normalized teacher entropy, the probability mass captured by a small top-$k$ candidate set, and student--teacher mismatch to allocate a limited probing budget. During exploration, SPOT evaluates teacher-proposed candidates through verifier-scored student continuations. During exploitation, these outcomes produce a closed-form, KL-regularized target that favors candidates with better downstream outcomes while remaining anchored to the teacher distribution. Extensive experiments across multiple student models and reasoning benchmarks demonstrate the effectiveness of SPOT in improving reasoning performance while balancing solution quality and coverage.

Large language models (LLMs) are increasingly used to detect unsafe content. A common approach is to combine judgments from a panel of models to correct individual mistakes, but this benefit may disappear when every model sees the same misleading context before voting. We study this risk in a controlled two-round experiment. Each model first judges an item alone, then judges it again after six simulated peers either assert the wrong label or abstain. We combine the final judgments by majority vote. Across six open-weight LLMs and six datasets, we find that the wrong-label peer message raises the average reviewer false-alarm rate from 56.5% under silent peers to 87.5%, and majority voting raises the panel false-alarm rate to 100%. Without an asserted label, the same panel outperforms its average member. The effect is strongly asymmetric: reviewers follow pushes toward "unsafe" far more than pushes toward "safe" (about 75% versus 17%), so the panel's false-alarm rate rises sharply while its harmful-miss rate changes little. The proprietary-model probe shows substantial variation across models. These results identify susceptibility to shared social cues as a failure mode of safety panels and provide a simple pre-deployment diagnostic.

On-policy distillation (OPD) supervises student-visited trajectories, yet divergence-based rules cannot determine whether an erroneous prefix remains correctable. We formulate this decision as counterfactual recoverability and replay each error state through budget-matched teacher-continuation and rollback branches. Based on their relative success, states are categorized as recoverable, irreversible-but-avoidable, or ambiguous, and these labels guide whether training retains, rolls back, or conventionally supervises the corresponding trajectory. On AIME branch diagnostics, the mean continuation-minus-rollback effect is 0.185 for recoverable states and -1.000 for irreversible-but-avoidable states, demonstrating opposite intervention preferences. A branch-derived recoverability proxy achieves an AUC of 1.000, substantially outperforming divergence alone at 0.392. Across frozen evaluations, recoverability-aware control achieves the strongest recorded performance, reaching 0.578 success on held-out AIME2025 compared with 0.517 for the best baseline. It also improves AIME2024-2025 average@32 from 0.2656 to 0.3125 and GPQA-Diamond average@32 from 0.2702 to 0.3070. Component ablations further show that retaining teacher-correctable prefixes provides the largest individual contribution. These findings establish recoverability as an outcome-grounded decision variable for selective supervision in OPD.

Memory-efficient matrix optimizers such as Sinkhorn gradient descent remove most AdamW optimizer state for dense Transformer matrices, but direct application to Mixture-of-Experts (MoE) training is unreliable. We study this failure in a controlled 110M-parameter nanowhale DeepSeek-style MoE pretraining setting. A SAGE/Sinkhorn hybrid reduces optimizer state from 0.883GB to 0.331GB but degrades evaluation loss to 3.8265, far above the AdamW baselines observed in the same setup (3.58--3.64 across the seeds we study). We show that routed MoE expert matrices are the dominant failure point: their gradients are conditional, temporally varying, and poorly served by stateless Sinkhorn normalization. We propose MESH, a hidden-momentum Sinkhorn update for MoE experts. MESH restores a temporal first-moment signal through the gradient-buffer lifecycle, without storing the expert first moment as optimizer state. MESH is an optional block-preconditioned variant that adds a coarse neuron/block inverse-RMS multiplier. Across ablations, temporal smoothing before matrix normalization is the primary causal ingredient; block/neuron preconditioning can improve the memory-quality frontier, but is not established as universally necessary. In two additional seeds, MESH and MESH-B reduce optimizer-state memory by 62.5\% and peak PyTorch CUDA allocation by about 12.6\% relative to AdamW, with a modest evaluation-loss gap. Full-state diagnostic variants recover AdamW-like performance in ablations, supporting the conclusion that MoE experts need temporal smoothing, but not necessarily full coordinate-wise AdamW state.

Long-context large language models (LLMs) are increasingly deployed in real-world applications, yet self-attention remains a major efficiency bottleneck -- especially during decoding -- due to the necessity of repeatedly processing ever-growing key-value (KV) caches. Existing sparse attention reduce computation by attending to fewer KV pairs, but often suffer from substantial accuracy degradation, require additional training, or rely on expensive hashing. In this work, we present BinaryPC, a training-free, data-aware hashing-based sparse attention for long-context LLMs. BinaryPC constructs compact binary hash codes and corresponding hash function by computing binary principal components of data. Unlike Locality-Sensitive Hashing (LSH) with data-independent random projections or learned non-linear hashing methods, BinaryPC constructs binary codes that explicitly preserve the structural information of data without requiring gradient-based training. Comprehensive experiments across multiple model families and long-context benchmarks show that BinaryPC preserves accuracy relative to full attention while achieving superior performance among sparse and hashing-based baselines. On modern GPUs, BinaryPC improves end-to-end decoding throughput by 3.56$\times$ over the FlashAttention kernel. Our code is available at https://github.com/yudaohai666/BPC.

Mixture-of-Experts (MoE) models enable model scaling while maintaining low inference-time compute by activating only a subset of experts per token. However, conventional routing relies on a fixed top-k selection, forcing the model to spend the same compute regardless of how many experts are relevant. We introduce elbow-based routing, a training-free inference-time modification that dynamically adjusts the number of experts on a per-token basis. Our method examines the sorted router probability distribution and identifies an elbow point that separates high- and low-probability experts. We find that most router distributions exhibit clear inflection points suitable for this strategy, and we show both theoretically and empirically that elbow-based routing preserves expert load balance. Experiments on a state-of-the-art MoE model demonstrate an average latency reduction of 5.3% while maintaining accuracy across six benchmarks.

Robotic planning often involves multiple objectives with complex priority relationships, such as safety, efficiency, and regulatory compliance. Rulebooks formalize these relationships, allowing partial ordering of objectives that generalizes both Pareto and lexicographic dominance. Computing the full set of rulebook-optimal solutions, however, is computationally expensive. To address this challenge, we introduce the concept of epsilon-rule-dominance, a principled notion of approximate dominance under rulebooks, and propose RA*pex, a best-first search algorithm that efficiently computes a compact set of epsilon-approximate rulebook-optimal solutions. RA*pex leverages dimensionality reduction, a technique used to speed up existing multi-objective search algorithms, while respecting rule hierarchies by maintaining separate closed sets and performing dominance checks over truncated and residual rule sets. We provide a formal analysis of RA*pex, proving that every rulebook-optimal solution is epsilon-rule-dominated (a generalization of approximate dominance we introduce) by at least one solution in the returned set. Empirical results demonstrate that our approach achieves computation times over two orders of magnitude faster than existing methods.

We introduce NOLLI, a procedurally generated English-Korean puzzle benchmark designed to diagnose where Korean performance gaps arise. It comprises 15 puzzle types (25 tasks; 7,500 items), with every instance seed-regenerable, verified to have a unique solution, and scored deterministically. Rather than equating harder with bigger, we calibrate difficulty behaviorally, tuning each generator until a fixed reference model lands in target accuracy bands. Its three-level design combines matched direct translations, script adaptations over Hangul jamo (sub-syllabic letters), and Korean-only tasks grounded in Korean culture or orthography. We evaluate 15 frontier, open-weight, and Korean-developed models; among the 12 above a 3% overall-accuracy floor, matched English-Korean accuracy is statistically equivalent within a +/- 10 pp margin (TOST), suggesting little cost from presentation language alone. Writing-system-intensive tasks show sharper gaps: Korean Cipher falls behind English by up to 68.7 pp, whereas Cryptarithmetic over the same jamo shows no systematic penalty, and Jamo Composition accuracy predicts Korean Cipher accuracy. These contrasts are diagnostic rather than causal, consistent with difficulty in multi-step sub-syllabic execution. Korean-only tasks separate rule-application deficits, which vary in sign, from a Kinship deficit positive in all 12. Finally, a salient size measure fails to grow from Easy to Hard in 7 of 15 types, making structural size an unreliable proxy for empirical difficulty.

Scientific machine learning often relies on proxy targets computed from known domain factors when direct observations are limited. When those same factors are used as model inputs, however, high predictive accuracy may reflect reconstruction of the proxy-generating equation rather than robustness to degraded factor information. We study this problem in RUSLE-derived soil-loss proxy prediction under controlled degradation of the soil-erodibility factor $K$. We introduce a diagnostic framework that combines degraded-formula references, classical tree-based baselines, matched direct and formula-feature predictors, contextual ablations, tail-error analysis, and degradation robustness scoring. We then propose RASPL, a formula-preserving residual framework that retains the degraded formula estimate as the prediction anchor and learns an adaptively gated contextual correction. RASPL substantially outperforms matched direct prediction and provides stronger degradation and tail robustness than treating the formula estimate as an ordinary input feature. Within RASPL, a compact statistical encoder achieves the highest macro-averaged $R^2$ and lowest computational cost, whereas a convolutional encoder achieves the strongest degradation robustness and lowest Tail95 mean absolute error (MAE). These results establish formula preservation as the central design principle for robust learning from factor-derived proxy targets.

Large language models (LLMs) perform table-centric prediction through in-context learning, making demonstration selection critical to performance. Existing retrieval methods prioritize similarity to the query, but similar demonstrations often reinforce the model's likely prediction rather than reveal the distinctions needed for difficult decisions. We propose EdgeLM, a retrieval framework that instead selects edge evidence, demonstrations that are both relevant to the query and informative about the decision boundary. EdgeLM retrieves two complementary forms of edge evidence by selecting data edges, nearby examples with different ground-truth labels, and model edges, similar examples previously misclassified by the deployed model. EdgeLM requires neither model retraining nor task-specific engineering. Across five data wrangling tasks, fifteen datasets, and five open-weight and proprietary LLMs, EdgeLM consistently achieves the best or near-best performance in every setting, while ablations show that the two forms of edge evidence provide complementary benefits. Our code and datasets are publicly available at https://github.com/soroushomidvar/EdgeLM.

Decoding continuous motor trajectories from neural activity is essential for developing practical brain-computer interfaces (BCIs). However, current neural decoders are constrained by the limited scale and heterogeneity of neural recordings. In contrast, behavioral data can be collected more readily and at substantially larger scale from humans, animals, simulations, and robotic systems. Here, we introduce NeuroPB, a framework that scales neural decoding by transferring knowledge from pretrained behavioral representations. NeuroPB first pretrains a motor encoder on large-scale motor behavior data and then aligns neural activity with the resulting behavioral representation space using a limited set of paired neural-behavioral recordings. A neural encoder and lightweight motor decoder are subsequently optimized to reconstruct continuous movement from the aligned neural representations. Across multiple macaque motor datasets, behavioral pretraining improves trajectory decoding, including an 11% $R^2$ increase on center-out and 8% on random-target compared with training the motor encoder from scratch. Notably, pretraining on robotic trajectories achieves performance comparable to pretraining on macaque trajectories, demonstrating that transferable kinematic structure is shared across biological and artificial models. Moreover, decoding performance improves as the scale and diversity of robotic pretraining data increase, when the amount of neural data is fixed. Pretraining also enhances generalization across recording sessions, subjects, and motor tasks, with only 10% calibration needed to match training from scratch. Overall, these results establish behavioral pretraining as a scalable source for neural decoding and provide a promising route toward high-performance and calibration-efficient BCIs under limited neural data.

Neural PDE solver auto-design is fundamentally a search-space representation problem. In the space of unrestricted Python programs, valid solvers form an extremely sparse subset: most candidate programs are syntactically incorrect, semantically incompatible, or numerically unstable. Direct code generation therefore forces an LLM to spend most of its search capacity navigating implementation failures rather than reasoning about solver quality. ADSL-PDE addresses this challenge by introducing a structured search state between solver concepts and executable code. It represents the functional decisions that determine a neural PDE solver (architecture, physical constraints, objectives, sampling, and optimization) while abstracting away low-level implementation details. A deterministic compiler maps each valid search state to an executable solver. In effect, ADSL-PDE reshapes the search space: it removes large regions of invalid programs, increases the density of meaningful candidates, and preserves the compositional freedom needed to discover previously unseen designs. Solver evolution can thus operate over design decisions rather than code artifacts. Built on this representation, our evolutionary agent iteratively proposes, evaluates, and refines solver search states using empirical feedback. Across multiple PDE benchmarks, ADSL-PDE improves both search efficiency and optimization stability, achieving an improvement of more than 52% within the first ten evolution iterations. These results suggest a broader principle for LLM-driven auto-design: effective agents do not merely require stronger reasoning, but rather a search representation that concentrates exploration on valid and consequential decisions.

Gradient descent has been of particular interest in modern machine learning beyond sole focus on optimization. Implicit bias emerging from optimization, though not being encoded by the learning objective, often prevents from overfitting to spurious patterns. A typical instance is the max-margin implicit bias of a linear classifier, widely established for exponentially tailed loss functions. Even after having a given dataset separated, the parameter vector continues to evolve towards the max-margin direction asymptotically along the gradient descent dynamics. This phenomenon corroborates a frequent empirical observation of "train longer, generalize better." However, the max-margin convergence is an asymptotic phenomenon, and what is worse, this asymptotic convergence rate is significantly slower than pure convex optimization. Even so, the parameter vector along gradient descent dynamics commonly correlates with the max-margin direction positively (though not exactly) within considerably fewer iterations than the asymptotic rate. By shedding another light on this classical problem, this work aims to understand the mechanism of this early-stage alignment phenomenon. Our theoretical results demonstrate that the parameter vector weakly aligns with the max-margin direction within $O(\exp(\exp(-δ)))$ iterations, where $δ>0$ is the permissible alignment error, which is shown to be tight. By tracking the radial and tangential flows, our proof operates on the alignment dynamics directly with dataset geometry and gets rid of the asymptotic expansion, which is a key insight to establishing faster weak alignment.

Self-supervised learning on graphs is largely shaped by contrastive methods that depend on carefully designed augmentations, and by generative methods that reconstruct node attributes in the input space. Both paradigms can entangle representations with low-level input statistics rather than with relational structure. Joint-embedding predictive architectures (JEPA) instead learn by predicting latent targets rather than reconstructing inputs. Recent work has explored this idea for graph-level representation learning, but how to design JEPA-style objectives for node-level tasks, and which structural signals the predictor should condition on, remains less clear. We present NodeJEPA, a joint-embedding predictive architecture for node-level graph self-supervised learning. NodeJEPA masks structure-aware k-hop ego-subgraphs and trains a context encoder to predict the latent representations of the masked nodes. These targets come from an EMA-updated target encoder with stop-gradient. A structure-conditioned predictor integrates spectral and centrality descriptors through cross-attention. Variance, covariance, and Laplacian spectral regularizers help stabilize the embedding geometry, and an optional curriculum gradually increases masking difficulty during training. Because prediction occurs in latent space, NodeJEPA does not rely on input reconstruction or hand-crafted graph augmentations. We evaluate NodeJEPA on standard node classification benchmarks under linear probing and fine-tuning protocols, and conduct ablations on masking, prediction, and regularization design choices. Our study offers a practical recipe for node-level JEPA-style latent prediction on graphs, and clarifies when structural conditioning helps representation learning. Code, configurations, and evaluation scripts are publicly available at https://github.com/OliverZ-dot/Node-Jepa.

We propose a method to optimize the correlation among convolutional neural network (CNN) features that are used as inputs to quantum neural network (QNN) to enhance image classification accuracy. Unlike prior approaches that employ orthogonal decomposition as preprocessing, we intentionally introduce correlated features that are more physically compatible with QNN. This design leverages the QNN's inherent ability to exploit quantum entanglement for representing correlated states-an advantage unavailable to classical neural networks. We hypothesize that aligning feature correlations with the entanglement structure of QNN improves binary classification performance. Based on a mathematical derivation of QNN outputs, Monte Carlo simulations indicate that an average correlation between features of 0.5 yields optimal classification accuracy. To validate this finding, we evaluate a quantum-classical hybrid model on three tasks: CIFAR-10 (automobile vs. truck), Fashion-MNIST (shirt vs. coat), and radar micro-Doppler signatures (robotic dogs vs. non-robots). To regulate feature correlations, we introduce a correlation-regularization term on the outputs of the CNN, driving the off-diagonal entries of the feature correlation matrix toward a target constant. Across all datasets, inducing intermediate correlation consistently improved accuracy compared to low, high, or unregulated correlations, while also reducing classification accuracy variance. These results demonstrate that imposing moderate feature correlations-without modifying the quantum circuit-enhances classification accuracy and stability by aligning feature statistics with the QNN's entanglement structure. This study highlights the potential of QNN to surpass the performance of classical classifiers as more qubits become available.

Collaborative music agents need internal representations rich enough to support both understanding and generation, yet flexible enough for a workflow where the human retains agency. We present a hierarchical self-supervised ``world model'' for symbolic music: a 2.55M-parameter Swin V2 encoder trained on MIDI piano-roll images with JEPA-style objectives (pitch- and time-shift equivariance, masked embedding prediction, and a distributional regularizer), using no labels and no music-theory vocabulary. Probing the frozen embeddings shows that the level at which a musical property becomes decodable tracks its musical time scale: phrase boundaries are read off the coarsest levels, note density and harmonic detail off the finest. Temporal and phrase structure emerge from the self-supervised objectives alone, while harmonic content must be asked for; a small chord-supervision head raises joint chord recovery from .18 to .54, and key detection, which is never supervised, from .16 to .70. Following the Representation AutoEncoder paradigm, a conditional flow-matching model stands in for a trained decoder, flowing in pixel space from PCA-reduced conditioning: it reproduces a target window at pixel F1 $0.996$, and the same per-level conditioning dropout that controls how far variations stray also enables graphical prompting for masked inpainting with no inpainting-specific sampler. The pipeline runs on CPU producing a suggestion in $2.8$ s, or $0.6$ s on Apple MPS, which we demonstrate in a live interactive demo. In concert with an LLM-based brain, these capabilities supply the core of a collaborative music creation agent in service of, rather than in place of, human agency.

Hypergraph neural networks (HGNNs) have demonstrated remarkable capabilities in processing complex higher-order relationships. However, their performance is highly dependent on labeled data, making them vulnerable to label noise. Despite advances in learning with label noise (LLN) and graph learning with label noise (GLN), noisy-label learning on hypergraphs remains underexplored. In this paper, we present a systematic study of hypergraph node classification under label noise. First, we adapt representative LLN and GLN methods to hypergraphs and evaluate them under a unified benchmark, revealing the limitations of existing robust learning strategies for hypergraphs. Building on this, we propose a new hypergraph robust framework, HyperTrust, which first estimates hyperedge trustworthiness through a pretraining-based, entropy-aware strategy, and then incorporates the HyperedgeBoost module to enhance reliable supervision by connecting unlabeled nodes to trustworthy hyperedges, as well as the HyperedgePrune module to suppress noisy propagation by removing untrustworthy node-hyperedge incidences. Finally, two modules work collaboratively to adjust the hypergraph structure and generate final predictions. Extensive experiments and theoretical analysis demonstrate the effectiveness and robustness of HyperTrust on multiple hypergraph datasets under various noisy settings. Our work provides a unified benchmark and an effective solution for hypergraph learning with label noise and lays a foundation for future research in this direction.

Large language models can produce fluent financial analysis, but fluency alone does not establish whether a report is suitable for institutional delivery. We introduce FinReportBench, an expert-grounded benchmark for measuring and improving institution-grade financial report generation. Expert review reveals recurring gaps in report identity, institutional components, source discipline, and visual delivery. We derive a 35-item rubric through expert partial orders, multimodal evidence, and audits of decision boundaries, covering deliverability, report identity, and institutional completeness. Starting from 10,000 balanced Chinese and English financial-research source records, we curate 244 bilingual tasks across three research objects and two input tiers. Each task separates the public query, reconstructed research trajectory, and hidden source packet. Three independent judge families reproduce the expert partial order at near-ceiling rates, showing that bounded, observable criteria support reliable evaluation. Across nine model families, basic deliverability is nearly saturated, while report identity and institutional completeness remain the primary bottlenecks. The largest cross-model gaps concern generation-trace control, information density, and data discipline rather than basic report framing. We then use benchmark-guided skill distillation to turn recurrent failures into reusable generation and self-review constraints. Across five model families, the evolved skill improves mean G1 by 33.85 points and mean G2 by 13.83 points over paired no-skill runs while preserving G0 for every pair. Code and benchmark artifacts are available at https://github.com/MisterBrookT/finreportbench.

Multimodal temporal data are inherently irregular and uneven in information density, yet most models rely on uniform discretization, leading to inefficient representations. We propose \textbf{EvtGraph}, a unified framework that aligns computation with temporal salience under explicit budget constraints. EvtGraph reparameterizes sequences into event-level tokens via event-adaptive compression (EAMC), selects a compact subset with a node budget (NBC), and performs temporally constrained sparse graph reasoning (T2SG). This transforms dense sequences into structured computation over salient events, reducing complexity while preserving critical transitions. We show that this design provides a practical mechanism for allocating representational capacity under a fixed budget, yielding a consistent performance--efficiency trade-off, where a small budget is often sufficient in practice. Experiments on multimodal clinical (MIMIC-IV + CXR) and cross-domain benchmarks demonstrate that EvtGraph outperforms both Transformer-based and recurrent baselines while significantly improving efficiency. These results suggest that budget-constrained event-centric representation provides a general paradigm for learning from high-redundancy temporal data.

While retrieval-augmented generation systems partially address the hallucination issues in large language models, it also introduces new vulnerabilities to knowledge corruption attacks. Adversaries exploit these vulnerabilities by poisoning documents provided by RAG system to manipulate LLM outputs. To counter this threat, we propose SecureCollaRAG, a Byzantine-tolerant collaborative RAG framework leveraging Multi-source Knowledge Validation Mechanism. Our approach enables agent system to securely verify document provenance through dynamic GNN-based credibility scoring, effectively preventing stealthy knowledge corruption attacks while preserving essential domain knowledge integrity. Through extensive evaluations and formal analysis, we demonstrate that SecureCollaRAG maintains robustness against attackers under non-IID data distributions.

Audits have emerged as a critical instrument for algorithmic governance, providing a mechanism for external scrutiny and governance of machine learning models. However, ensuring the integrity of such assessments remains a challenging issue. For instance in regulatory contexts, audits are typically declared or easily detected, thus enabling model providers to manipulate the process, whether intentionally or inadvertently. This vulnerability is particularly acute in the context of fairness evaluations, in which providers can often infer sensitive attributes and strategically equalize allocation rates between groups to satisfy fairness metrics. In this paper, we introduce a novel audit protocol designed to significantly increase the post-audit detectability of such manipulations by enabling the auditor to query the model in an oblivious manner. Our approach leverages a Private Information Retrieval mechanism to require the provider to label a large set of instances, while preventing it from knowing which subset will ultimately be used for the audit. The protocol is efficient, imposes minimal overhead on the auditor, and requires no modification to the audited model, its training procedure, or its inference pipeline. We provide theoretical guarantees showing that, under this protocol, a provider attempting to hide unfairness must falsify a significantly larger number of responses, thereby increasing both the difficulty and the likelihood of detection of manipulation. Experimental results across representative audit scenarios confirm the effectiveness and practicality of our approach.

Continual learning (CL) requires models to learn tasks sequentially, yet deep neural networks often suffer from plasticity loss and poor knowledge transfer, which can impede their long-term adaptability. Drawing high-level inspiration from global neuromodulatory mechanisms in the brain, we introduce Neuromodulation and Synchronization (NeuMoSync), a novel architecture that integrates dynamic, neuron-specific modulation into deep neural networks to enhance their adaptability and plasticity. NeuMoSync extends standard neural network architectures with learnable feature vectors for each neuron that track network-wide historical context and with a module operating at a higher level of abstraction. This module synthesizes neuron-specific signals, conditioned on both current inputs and the network's evolving state, to adaptively regulate activation dynamics and synaptic plasticity. Evaluated on diverse CL benchmarks, including memorization (Random Label CIFAR-10 and Random Label MNIST), concept drift (Shuffle CIFAR-10 and Shuffle Mini-ImageNet), class-incremental learning (Class Split ImageNet and Class Split CIFAR-100), and domain-incremental learning (Permuted MNIST), NeuMoSync demonstrates strong performance in retaining plasticity and achieves improvements in both forward and backward adaptation compared with existing methods. Ablation studies validate the necessity of each component, while analysis of the learned modulatory signals reveals interpretable coordination patterns across tasks. Our work underscores the potential of integrating global coordination mechanisms into deep learning systems to advance robust, adaptive continual learning. The code is publicly available at https://github.com/RoozbehRazavi/NeuMoSync.

Accuracy changes after language-model self-revision are usually interpreted as changes in reasoning. We show this can fail at the answer-extraction boundary, and test the failure causally rather than only observationally. Across Qwen3.5 (0.8B-9B), Gemma-4-12B, and two frontier models via API (Tencent Hy3, Nvidia Nemotron-3-Ultra-550B) in 29 primary cells plus a frontier arm, we decompose the always-revise accuracy shift into a content margin (both answers parseable) and format-recovery/loss margins (parseability changes). On 12 cells with meaningful unparseable-answer rates, format effects exceed content effects (Wilcoxon p=1.7e-3). To test this causally, we force already-generated reasoning through grammar-constrained decoding so every answer is parseable by construction: across 14 cells this closes a median 71% of the gap between the naive total effect and the content-margin estimate, with two cells converging exactly and a residual on the two largest-effect cells reported rather than dismissed. A clustered model confirms floor-scale (0.8B/2B) models have far higher odds of content-level change and harm than capable-scale models (p<1e-7). Replicating a cited confidence-gating protocol verbatim on Qwen3.5 does not reproduce its reported gain and shows the same near-zero content margin. A frontier check on much larger models shows format-dominance intensifying with scale: content margin is exactly zero in all 5 cells despite total effects up to +0.275, though this arm is lower-powered. The calibration-floor criterion on the content margin reveals a squeeze: floor-scale cells have headroom but insufficient signal, capable-scale cells have signal but little headroom; only one cell is marginally viable, with negligible sealed-holdout gain. Content is a minority share of what the field has measured as self-correction. We release the instrument, code, and derived results.

Large Audio-Language Models (LALMs) excel at general speech understanding; however, adapting them to fine-grained tasks like Speech Emotion Recognition (SER) remains a significant bottleneck. Current Parameter-Efficient Fine-Tuning (PEFT) methods typically operate in flat Euclidean space, and this geometry fails to capture the multi-granularity nature of emotion cues, which range from low-level prosody to high-level semantics. To address this, we propose HyPASE, a hyperbolic PEFT framework for LALM-based SER. HyPASE leverages the Poincare ball model, using the hyperbolic radius as an explicit proxy for representational granularity. The framework consists of two core components: a Hyperbolic Geometric Adapter (HGA) for layer-adaptive weight modulation, and an Emotion-aware Multi-capacity Cross-modal Aggregator (EMCA) that compresses multi-scale features into compact audio prefixes. Empirical results on standard benchmarks show that HyPASE outperforms Euclidean PEFT baselines across all metrics on MELD and achieves a notable Unweighted Accuracy gain on IEMOCAP, particularly in class-imbalanced emotion recognition, with the accompanying slight Weighted Accuracy trade-off reflecting hyperbolic space's geometric prioritization of minority-class representations; furthermore, HyPASE achieves robust zero-shot cross-dataset generalization within a constrained parameter budget. By grounding the adaptation process in hyperbolic geometry, HyPASE offers a highly efficient path for LALM fine-tuning.

Multimodal learning of images and tabular data is often impaired by ineffective representations, resulting in redundancy, dispersion, and generalization problems. To tackle this challenge, we introduce Graph-Enhanced Descriptor Sequencing (GEDS), a structured feature sequencing algorithm grounded in principles from the Column Permutation Problem (CPP). GEDS refines statistical descriptors of the features through similarity graph-based computations, systematically determining an effective feature sequencing. We incorporate GEDS within an order-aware efficient transformer framework, utilizing order-aware memory tokens that explicitly adhere to the derived feature sequencing via a dedicated loss function. Experimental results across multimodal benchmarks demonstrate that iStructTab effectively minimizes feature dispersion, improving predictive performance and robustness, and highlighting the significance of structured feature sequencing in multimodal learning.

Fine-tuning enables a source model to acquire desired capabilities and behaviors in a target domain while retaining much of its general-purpose competence. However, this adaptation process can also degrade alignment properties that were present in the source model. Recent work has shown that large language models can be trained using LoRA-based modules known as introspection adapters (IAs) to describe behavioral changes induced by fine-tuning. However, existing studies primarily consider settings in which the model is fine-tuned on datasets explicitly designed to implant a specific behavior and is then asked to explain the implanted behavior. This differs from practical deployment scenarios, where the central concern is often side-effect misalignment: unintended degradation of alignment caused by fine-tuning on tasks that are not obviously related to safety or alignment. To bridge this gap, we formulate a novel problem setting called \emph{side-effect introspection}, in which the target of introspection is not a behavior explicitly implanted through fine-tuning, but rather alignment shifts that emerge as unintended side effects, and we construct a dataset for this setting. Furthermore, to enhance sensitivity to internal model changes, we propose the Delta-Aware Introspection Adapter (DAIA), a novel mechanism designed to explicitly process both base-model activations and activation differences induced by fine-tuning. Our empirical evaluation shows that introspection learning generalizes to unseen fine-tuned models and safety categories, and that DAIA consistently outperforms existing introspection adapters.

Large language models are increasingly used for information seeking, yet semantically equivalent questions phrased in different ways can receive answers of considerably different quality. System prompts are widely employed to steer response behavior, but they are typically optimized for average-case quality, so some question phrasings may still receive incomplete or low-quality answers. To address this, we formulate a constrained mixed-strategy GroupDRO framework for system-prompt selection. Instead of optimizing the system-prompt text, the framework assigns weights to system prompts in an existing pool to minimize the worst-case information-quality loss across evaluation metrics and groups, while constraining the mean loss to stay close to that of average-based selection. Because pool generation and selection are decoupled, the method applies to any system-prompt pool and can leverage an ensemble of complementary system prompts rather than a single one. Across five LLMs on two bilingual medical and consumer-finance benchmarks, the constrained method reduces the Overall Mean, Worst 25% Mean, and Worst by 13.1%, 13.2%, and 13.7% on average relative to no mitigation while keeping overall quality close to Average selection. Its multi-prompt weights reveal complementarity across metric-group pairs. Code and data are available at https://github.com/Rainxu09/equitable-system-prompt-selection.

Code generation systems make each LLM call with a model, a prompt, and decoding settings. However, existing optimization methods usually tune only part of these choices or use one fixed configuration for all tasks: global optimizers search one configuration for all tasks, routers choose only a model, and prompt optimizers keep the model and decoding settings fixed. This leaves their joint, group-specific interactions unclear. We therefore examine how these choices interact and observe that prompts and decoding settings interact, tuning effects vary by model, and the best configuration varies by task difficulty. Guided by these observations, we introduce COMPAS (Code-generation Optimization over Models, Prompts, And Decoding Settings), a difficulty-aware method that learns group-specific quality-cost fronts through low-cost model selection and joint prompt-decoding search, then routes each test task to its matching front online without further search. Under a matched search budget on LiveCodeBench, COMPAS improves pass@1 from 45.9% for the best baseline to 52.8% while reducing cost from $36.57 to $4.92. This also transfers to repository-level code generation on SWE-bench, resolving 76.0% of tasks versus 70.0% for the best baseline. Code and the reproducibility artifact are available at https://github.com/gjz78910/COMPAS.

Contemporary model-free reinforcement learning algorithms can achieve very high performance, but have low sample efficiency and are not robust to changes in the environment. Model-based algorithms have much higher sample efficiency, but still fail when the environment shifts. This paper introduces Adaptive Topological Learning with Abstract Successors (ATLAS) to combat these challenges. ATLAS uses a Grow When Required network with Successor Features in order to achieve high sample efficiency while also robustly tackling catastrophic forgetting. We evaluate ATLAS in spatial navigation tasks, benchmarking its performance against common on-policy and off-policy algorithms. Our empirical results demonstrate that by structurally decoupling transition dynamics from the reward signal, ATLAS achieves near-instantaneous adaptation to new goals and can exhibit positive backward transfer, significantly outperforming baseline methods in non-stationary environments.

Large language model (LLM) configuration evaluation is challenging due to limited evaluation budgets, varying costs, and multiple competing objectives. In this paper, we formulate LLM configuration evaluation as a cost-aware multi-objective bandit problem, where each configuration evaluation incurs a configuration-dependent cost and yields a noisy vector-valued outcome. Under this framework, we study two fundamental problems: online configuration selection and Pareto configuration identification. For online configuration selection, we propose a hypervolume-based UCB algorithm that optimizes an optimistic hypervolume-per-cost index. We establish a budgeted regret bound of order $O\bigl(\sum_{i\ne i^\star}\frac{\log B}{Δ_i}\bigr)$, where $B$ is the evaluation budget, $i^\star$ is the optimal configuration in terms of hypervolume efficiency, and $Δ_i$ is the corresponding efficiency gap of configuration $i$. This bound retains the logarithmic budget dependence of classical single-objective budgeted bandits. For fixed-budget Pareto identification, we develop a cost-aware empirical gap elimination algorithm and prove that its error probability is of order $O\bigl(\exp(-\frac{B}{H_{μ,c}})\bigr)$, where $H_{μ,c}$ is a cost-aware Pareto identification complexity depending on configuration costs and Pareto classification gaps. This error probability decays exponentially with the evaluation budget and recovers the standard Pareto set identification guarantee when all configuration costs are identical. Experiments on LLM configuration evaluation tasks demonstrate that the proposed framework enables efficient online decision-making and accurate cost-aware Pareto identification under limited budgets.

Removing the left context from a causal language model reveals a useful kind of boundary: an edge where the model processes the same right-hand tokens with little change. We turn this observation into prefix-removal probing and introduce Right Reset (RR), which measures preservation of the right-hand hidden-state trajectory. A dynamic program converts RR edge scores into variable-length chunks. On flattened text formed by concatenating topically similar records after deleting their separators and layout, RR recovers 47.7% of the original records as clean units, versus 25.9% for a BGE embedding boundary, the strongest tested conventional baseline without task-specific model training. The gain persists after rendering and OCR. Passive scores from the same Qwen3-4B layer and direct prompting of a same-scale instruction model perform substantially worse on flattened records. Across six language models, RR-selected cuts also undergo consistently less local output disruption than unselected candidate edges. An observed-token likelihood-ratio readout is competitive in some architectures, indicating that the central contribution is the intervention: context dependence itself can provide a boundary signal when surface structure is weak.

Explicit onshore tsunami inundation forecasting can improve public risk awareness, but deterministically predicted inundation boundaries under highly uncertain conditions, such as near-field tsunamis generated by megathrust earthquakes, may falsely imply safety outside the boundaries. Consequently, current warnings primarily target coastal tsunami height, not onshore inundation. Although machine learning enables instant inundation predictions, they remain deterministic, lacking uncertainty quantification. Here, we develop a probabilistic ensemble model based on a conditional diffusion model (a type of generative AI) that reconciles accuracy with calibration. Validated with the 2011 Tohoku-oki earthquake data, our model faithfully tracks the postearthquake uncertainty decreasing over time while accurately predicting inundation depth and extent. Our framework shows that generative AI can shift tsunami forecasting from determinism to probabilism, providing a foundation for next-generation early warning.

This paper studies generalized low-rank matrix bandits with multiple prioritized objectives. At each round, the learner selects a matrix-valued arm and observes a vector-valued reward, whose components correspond to multiple objectives with different priority levels. Each objective is governed by an objective-specific generalized low-rank matrix model, and the learner evaluates arms according to a lexicographic preference order, prioritizing higher-level objectives before lower-level ones. We propose \textsc{Lexi-LowGLM}, an efficient online algorithm that first estimates objective-specific low-rank subspaces and then performs lexicographic learning in the reduced feature spaces. Unlike existing single-objective algorithms that repeatedly solve a batch generalized linear estimator using all historical observations, \textsc{Lexi-LowGLM} updates each objective-specific estimator via an online Newton step, reducing the estimator-update complexity over $T$ rounds from $O(T^2)$ to $O(T)$. We establish a regret bound of $\widetilde O\left(W_i^{\rm lex}\sqrt{m}\,(d_1+d_2)r\sqrt{T}\right)$ for each objective $i\in[m]$, where $r$ is an upper bound on the ranks of the objective-specific parameter matrices and $W_i^{\rm lex}$ characterizes the lexicographic trade-off effect. This bound depends on the effective low-rank dimension $(d_1+d_2)r$ rather than the ambient dimension $d_1d_2$. Numerical experiments further validate the effectiveness and computational efficiency of the proposed method.

Task-specific fine-tuning can improve the performance of large language models (LLMs) on downstream tasks. However, our study reveals that task-specific fine-tuning can also weaken the safety guardrails of aligned LLMs. A widely adopted strategy for preserving safety during fine-tuning is to incorporate safety data. Although previous studies have shown that randomly mixing safety data can alleviate safety degradation, the underlying principle determining why some safety examples are more effective than others still remains unclear. In this paper, we propose DataRx, a missingness-aware sampling method for selecting safety-critical examples. DataRx is based on the hypothesis that a safety sample is more effective when the selected examples provide safety signals that fill the missing parts of LLMs' safety capabilities. DataRx's key insight is leveraging high-dimensional hidden representations rather than discrete tokens to quantify the safety signal gap between the target model's native response and the safety reference response. The results show that, with only 1% additional safety samples from BeaverTails, DataRx reduces the average attack success rate of Llama3-8B-Instruct across seven downstream tasks from 59.23% under random sampling to 13.70%. In addition, DataRx can be combined with the existing safety data synthesis method to further enhance safety defenses during fine-tuning. We hope that DataRx will inspire more data-centric defense research.

Autonomous cyber defense systems based on Deep Reinforcement Learning (DRL) have attracted significant research attention, yet remain evaluated almost exclusively against static, heuristic red agents, leaving their robustness against adaptive threats critically understudied. Meanwhile, recent advances in Reinforcement Learning with Verifiable Rewards (RLVR) have improved LLM reasoning, but their integration into cybersecurity remains elusive due to the absence of suitable benchmark environments and interaction datasets. To bridge this gap, we introduce Trident, an agentic LLM red teaming framework comprising three components: a dynamic benchmark with isolated sandbox servers spanning CybORG CAGE 4 and CyberWheel, a dataset comprises over 13,000 high-fidelity red-blue interaction trajectories for RLVR, and a ``Code-as-Policy'' RLVR agentic architecture Trident Agentic). The latter reformulates red agent training as a contextual bandit via a tripartite Log Summarizer--Planner--Coder design, where a trainable Planner generates complete attack strategies from compressed execution logs, which a frozen Coder translates into executable Python policies deployed against live DRL defenders. Empirical evaluations reveal a fundamental brittleness in existing defenses: with a single trainable 7B planner, Trident reduces blue agent defensive performance by an average of 522% compared to static red agent baselines while autonomously discovering emergent behaviors such as decoy avoidance and adaptive state prioritization that static heuristics entirely fail to uncover.

Translating wordplay across languages has long challenged both professional translators and machine translation systems. We investigate three approaches to translating puns from English to French by combining large language models with linguistic constraints for wordplay generation. Our baseline uses a large language model with feedback from a discriminator prompted with positive and negative French examples. Our guided reasoning pipeline uses combined phonetic-semantic embeddings to retrieve lexical candidates for wordplay generation. Finally, our multi-agent framework iteratively evaluates and regenerates candidate translations using specialized feedback. Moving beyond literal translation, our objective is to preserve the linguistic creativity, ambiguity, and humor of the source-text wordplay rather than simply reproduce its vocabulary. The multi-agent and guided chain-of-thought systems ranked first and second, respectively, in the CLEF JOKER 2025 Task 2 competition under expert human evaluation, despite only modest improvements in BLEU and BERTScore. These findings suggest that both explicit phonetic-semantic guidance and iterative multi-agent evaluation can improve LLM-based wordplay translation relative to direct discriminator-guided generation, particularly when balancing semantic fidelity, phonetic similarity, and natural target-language expression

The Rashomon effect describes the phenomenon that many models can achieve nearly equivalent performance on the same learning task, with significant ramifications for robustness, feature importance, and customizability. These use cases motivate the computation of Rashomon sets: the set of all models whose regularized loss is near-optimal. Decision trees are one of the few model classes for which Rashomon sets can be fully enumerated, but this computation has always been conditional on a binarization of the original data, either restricting which splits each tree is allowed to make or substantially increasing the complexity of an already difficult combinatorial problem. We introduce the first algorithm that exactly enumerates decision-tree Rashomon sets while exploiting the ordered structure of continuous features. We further develop a relaxation for approximate enumeration and an anytime algorithm that progressively refines the set of candidate thresholds, producing increasingly detailed approximations that converge to the continuous-feature Rashomon set. Experiments show that coarse binarization can miss many trees, important features, and predictive multiplicity; our algorithms achieve orders-of-magnitude speedups over existing enumeration methods, with approximations providing further speedups while maintaining near-perfect recall.

Text summarization is deceptively difficult. While condensing information seems straightforward, real-world enterprise summarization of support tickets, legal documents, incident reports, and more, demands strict adherence to domain-specific guidelines, output formats, and organizational conventions. Crafting prompts that reliably satisfy these constraints is labor-intensive, requiring significant human expertise and continuous maintenance as requirements evolve. Existing automated prompt optimization methods reduce this burden through Large Language Model (LLM) critique-driven refinement, yet remain limited by static prompts that cannot adapt to the diversity of summary applications. We propose Multi-LLM Iterative Data-Adaptive Summarization (MIDAS), a multi-LLM framework that extends this paradigm with data-driven pattern learning and use-case-specific personalization, enabling automatic adaptation to different summarization requirements without manual prompt engineering. Applied to enterprise customer ticket summarization across five output formats, MIDAS achieves the strongest overall performance against state-of-the-art critique-driven optimization frameworks such as CriSPO and ZERA, improving ROUGE-1 by up to 11.0%, ROUGE-2 by up to 18.2%, and ROUGE-L by up to 8.0%, while consistently improving BERTScore F1 across all formats and output types. We additionally demonstrate cross-model and cross-domain generalization through multi-LLM configurations and finance-domain summarization benchmarks.

Risk-aware Q-learning (RaQL) provides a model-free, two-timescale estimator for dynamic risk objectives, but its finite-budget behavior remains fragile: fixed inner-loop hyperparameters can produce unstable value estimates, persistent Bellman residuals, and inefficient sample reuse. This paper proposes an adaptive training controller for Conditional Value-at-Risk (CVaR) RaQL and evaluates it on a daily Bitcoin trading task. The controller preserves the original CVaR estimator and Bellman fixed point; instead, it redesigns the training procedure through six coordinated mechanisms: per-cell inner-step sizing, outer-rate-matched decay synchronization, a short early correction for the VaR-like inner variable, a coverage-first-then-greedy sample allocation rule, progressive suffix aggregation of mature inner estimates, and data-driven calibration of key scales from online-observable quantities. Across 20 random seeds and 856,000 inner-transition samples, the controller reduces the mean empirical CVaR Bellman residual by approximately 85% relative to the fixed-parameter baseline (MeanBEQ: 1.2202 to 0.1854; MeanBEV: 1.1624 to 0.0535) and maintains stability across CVaR levels, discount factors, and training budgets. On the chronological out-of-sample test set, the learned policy attains a Sharpe ratio of 0.9281 with a maximum drawdown of 6.46% after transaction costs. Although buy-and-hold yields a higher cumulative return (35.43% vs. 23.61%), the adaptive policy achieves far lower volatility (9.57% vs. 47.93%), drawdown, and CVaR loss. These results demonstrate that adaptive finite-budget training design, applied solely to the training procedure without altering the risk objective, can materially improve the reliability and risk-adjusted performance of risk-aware Q-learning in financial applications.

Fifteen years ago, Low proposed that pun translators should stop searching for equivalent words and instead search for new points of contact between sound and meaning. In this paper, we investigate that idea computationally. We model pun translation as a process of discovery, exploration, and selection. A retrieval system searches semantic and phonological neighborhoods for target-language affordances: sound-meaning bridges that may support new wordplay. Multiple language models then explore these opportunities by generating competing translations, while a multi-perspective generate-and-rank architecture selects among them. Beyond system development, our primary contribution is an analysis of how retrieved affordances propagate through the translation process. We find that generators actively exploit retrieved opportunities, evaluators progressively concentrate around stronger sound-meaning bridges, and exact phonological collisions are selected at disproportionately high rates when available. At the same time, many puns still yield no usable affordances, suggesting that retrieval remains the central bottleneck in computational pun translation. The resulting picture is remarkably close to the process envisioned by Low. Successful pun translation emerges not from preserving source-language words, but from discovering new places in the target language where sound and meaning collide.

Long-horizon agents increasingly use persistent memory and tools to take actions with external side effects. A central failure mode is premature commitment: an agent acts before resolving whether its memory grounding is stale, conflicting, incomplete, or corrupted. We formalize this problem as safe commitment under memory uncertainty and introduce SafeCommit, a risk controlled layer between agent reasoning and external execution. The layer constructs a calibrated set of plausible latent worlds from memory, observations, tool outputs, provenance, and policy constraints. It permits a side effectful action only when a conformal action certificate shows that the action is safe in every retained world. Otherwise, it selects a low-side-effect probe that targets the worlds blocking certification, or returns a conservative fallback. Under calibrated world coverage, the probability of an unsafe certified commit is at most the target level α; with imperfect world proposal, the bound separates calibration and representation error. A dependency-free controlled simulator illustrates the safety-utility tradeoff and reproduces all reported results with one command. The goal is to offer a concrete approach for deciding not only what an agent should do, but when the available evidence is sufficient to safely do it.

Calibration requires a predictor to be unbiased after conditioning on its own predictions. Multicalibration asks for this guarantee simultaneously across a collection of groups. Many prediction tasks ask for several related features of the same conditional outcome distribution: variance is defined relative to the mean, skewness relative to both mean and variance, and conditional value at risk relative to a quantile. We study multicalibration for a sequence of $k$ properties in which each property is identifiable once the preceding properties are fixed. This framework includes Bayes pairs but does not require the properties to arise from a single loss. For every fixed $k\ge2$, we establish matching upper and lower sample-complexity bounds up to logarithmic factors under regularity conditions. Even with only polylogarithmically many binary groups, achieving multicalibration error $\varepsilon$ requires $\widetildeΩ(\varepsilon^{-(k+2)})$ samples. Conversely, for any finite group family $\mathcal G$, we give a randomized learner using $O(\varepsilon^{-(k+2)}+\varepsilon^{-2}\log|\mathcal G|)$ samples. Thus the sample complexity is $\widetildeΘ(\varepsilon^{-(k+2)})$ for polynomial-size group families. We instantiate the theory for three canonical examples.

Large language models (LLMs) are often used in conjunction with external knowledge sources to improve their factual accuracy and decrease hallucinations, through methods such as Retrieval-Augmented Generation (RAG). However, these systems remain susceptible to intrinsic hallucinations, where the model generates unfaithful or fabricated information that is not supported by the retrieved evidence. We propose a novel framework to assess model robustness against this phenomenon by stress-testing using natural, semantically equivalent variations of a user query found via adversarial optimization methods. We apply our framework, which enforces strict semantic equivalence constraints and an intrinsic hallucination objective, to a range of adversarial attack techniques across white-box, gray-box, and black-box adversarial settings. Evaluating these attacks on 5 open-source and 5 closed-source generator models across 3 datasets, we demonstrate that even state-of-the-art models are highly susceptible to meaning-preserving perturbations, which significantly degrade contextual faithfulness (by up to 50% for GPT-5-mini). Our findings indicate that faithful use of in-context evidence remains fragile even in state-of-the-art LLMs, motivating architectures and training objectives that enforce robust grounding independent of surface query form. Code is available at: https://github.com/atriviveksharma/intrinsic_hall

Neurosymbolic AI systems that integrate machine learning and symbolic reasoning are rapidly gaining attention. They complement the data-intensive statistical approaches of neural networks and language models with symbolic reasoning algorithms to function in high-stakes domains or in low-data regimes that characterize many real-world applications. We argue that the neurosymbolic combination of machine learning and formal reasoning is not a niche approach within AI, but rather includes many already successful techniques that are of crucial importance to the development of reliable, efficient and, ultimately, trustworthy systems. This perspective prompts a re-examination of the design of current AI systems. We show that many leading AI systems, including some that are not traditionally considered as neurosymbolic, can be analysed from the perspective of four principles of neurosymbolic AI design: Reasoning, Assurances, Interfacing and Learning (RAIL). Applying the RAIL framework offers a unified view of seemingly disparate AI systems, ranging from physics-aware machine learning to neuro-guided search (such as Google DeepMind's Alpha-* suite), causal learning and tool-augmented Large Language Models. Importantly, the RAIL principles will enable engineers to make better-informed and more principled decisions about the design and deployment of production-level AI systems. In this article, we introduce the RAIL principles, examine how they can be applied across major areas of AI, and illustrate how they may guide practitioners to integrate neurosymbolic methods into next-generation AI technologies.

Coding agents increasingly work across sessions, but prose notes can preserve a conclusion without the program state that supported it. After an upstream change, a repository may still build even though earlier verification claims are no longer valid. EA-Graph is an artifact-anchored memory for verification claims. It represents artifacts at sub-path granularity, resolves aliases to leaf definitions, anchors each claim to the content used to establish it, and keeps evidence strength separate from freshness. When replacement content is unavailable, the claim becomes unprovable rather than guessed. EA-Graph is evaluated on generated repositories whose behavior-to-artifact ground truth is known by construction. The task is to classify prior claims as unaffected, affected, or unprovable after value drift, logic drift, and deliberately withheld upstream content. The analysis covers 42 sessions across seven clean worlds, 14 model-world instances, three memory conditions, and two model tiers. In the Haiku round, artifact-anchored memory outscored prose notes and no persistent memory in all seven worlds; each exact paired Wilcoxon comparison yielded p = 0.0156. In the Sonnet round, the anchored condition was perfect, but frequent control ceilings left the preregistered contrasts non-significant. No session fabricated withheld content. These results support a bounded claim: artifact-anchored memory improved the smaller model's provability judgments in this testbed. An exploratory comparison further suggests that structured claim memory may narrow a capability gap by externalizing in-session re-derivation, but it does not establish cross- model equivalence. The study makes no claim about efficiency or repair quality.

Social media recommendation feeds often optimize for users' immediate impulses rather than preferences they would hold after deeper reflection. Some systems address this misalignment by incorporating users' explicit preferences via a configuration page or in-feed controls instead of just behavioral signals. However, users typically have evolving preferences, and their stated preferences and behavior naturally diverge, necessitating continuous reflection and feed realignment. But existing strategies require the user to take initiative and are often effortful; as a result, in practice they are rarely invoked. We present Compass, a system that aligns a user's feed with their reflective preferences by helping users reflect on and articulate their preferences given their behavior. To enable continuous reflection during everyday browsing, Compass surfaces in-situ reflections via lightweight notifications, while feed alignment is achieved by periodically simulating behavioral signals and directly manipulating feed content. We embedded Compass within YouTube Shorts and compared it against a baseline without continuous support through a 10-day field study (N=15). We found that Compass promoted more reflective and purposeful feed consumption, iterative preference adjustment, and stronger feed alignment, without sacrificing the casual nature of feed browsing.

Just like software and hardware, business processes are susceptible to vulnerabilities that can lead to product quality issues, delays, and increased costs. Business process vulnerabilities can arise from a variety of sources, including conflicting requirements, ambiguous documentation, invalid measurement spec-ifications, omission of quality checks, or implementations that differ from speci-fications. MIRABELLE is a system that identifies and characterizes business logic (BL) vulnerabilities from available business process representations, in-cluding ISO 9000/9001 documentation, user guides, work instructions, and pro-cess execution logs. MIRABELLE leverages recent advances in AI/ML to pro-cess available business process documentation and generate attributed graph rep-resentations of the business logic that can be processed using both graph and for-mal logic approaches for identifying potential vulnerabilities. However, extract-ing the business logic (e.g., operation execution sequences, decisions, input/out-put resources) from mostly natural language artifacts is challenging due to the required domain expertise, inherent process complexity, and the sometimes very large volumes of information. This paper focuses on our experimentation with Large Language Models (LLMs) and their role within MIRABELLE. We report on the performance of several LLMs across vital stages of vulnerability detection, from grammatical and technical error-flagging in short phrasings, to complete process structure recovery and extraction.

Agentic code generation has shown promise in automating and accelerating software development by utilizing Large Language Models (LLMs) to generate, test, and deploy code. For engineers and scientists, such systems have the potential to accelerate the development of applied and scientific workflows while reducing barriers to entry in domains that have yet to fully realize their benefits. However, a key gap remains: existing coding agents primarily focus on code generation and do not address the entire workflow lifecycle, including deployment and sharing. As a result, users develop and stitch modules independently while managing deployment on their own. To address this gap, we propose CURATE - Composition, User-in-the-loop, Reuse, and Automated Task Execution - a novel human-in-the-loop multi-agent system that uses LLM agents to manage and develop composable workflows across their entire lifecycle. A key feature of the system is a catalog that allows for the storage and reuse of modules across workflows. Module catalogs provide a foundation that can be expanded to support FAIR principles by facilitating the sharing and reuse of curated modules and subgraphs. We demonstrate the feasibility of our system with an initial prototype using Claude Opus 4.8, comprising 6 experiments: reproducing and adapting 4 workflows derived from the SeBS-Flow benchmark suite, and automating the development and scaling of a workflow that leverages a complex mechanistic model in environmental engineering used to simulate anaerobic digestion.

Generative models trained on artificially generated data have been shown to exhibit model collapse, resulting in significant performance degradation. As synthetic content increasingly contaminates the training corpora of language models, this raises critical concerns about the use of open data in continued pretraining. Although previous work has demonstrated model collapse in language models, it remains unclear whether exposure to synthetic data amplifies or attenuates the social biases already present in pretrained models. Because language models are known to reproduce and amplify demographic stereotypes, recursive training on self-generated data may create a self-reinforcing feedback loop in which biased associations become progressively stronger across generations. We call this hypothesized phenomenon fairness collapse. In this work, we construct controlled training regimes in which models are repeatedly trained on synthetic data using the Bias in Bios dataset. Across experiments, we observe a consistent and concerning pattern: fairness degradation emerges before substantial degradation is reflected by standard language-modeling metrics. This result highlights a critical risk associated with synthetic data contamination in language model training: bias can increase silently before strong indicators of model collapse become apparent.

Evaluations of LLM planning agents largely ask whether a task succeeds or a declared plan is followed. In strategic cyber-physical systems, a stronger question is whether the planning architecture remains appropriate after autonomous participants respond and physics constrains the outcome. We introduce a controlled, physics-grounded benchmark built around planning-induced control trajectories: the ordered planning operations and directives through which an execution architecture acts on other agents and the physical process. It implements predefined, sequential, hierarchical, and search executors in a smart-grid demand-response system with 40 heterogeneous prosumers and an independently simulated radial feeder. The LLM is bounded to typed policy declaration and short operator messages, while schedule construction, prosumer dynamics, and power flow remain explicit code. The protocol uses paired forced-mode counterfactuals, common random response draws, and event-level deadline feasibility. Three properties follow. Architecture materially changes outcomes: forced search is the oracle in all five baseline seeds. Execution fidelity needs more than mode agreement: objective substitution holds agreement at 1.0 while increasing voltage shortfall by 2.68x. A 144-scenario, 576-episode bank has feasible oracles from three of the four architectures. A prespecified stress-held-out ridge has mean regret 90.7 (95% interval [73.8, 108.6]) and no detectable value over fixed sequential; applying known deadline feasibility before quality prediction cuts regret to 29.0 and improves over fixed sequential by 61.1. An all-feasible ablation does not beat fixed search, localising the remaining challenge to within-feasible quality selection. A five-model extension separates stress-conditioned, state-blind, and invariant declarers; latency tails show that live feasibility should be treated probabilistically.

Emfrp is a functional reactive programming (FRP) language designed for small-scale embedded systems. Time-varying values are the primary abstraction mechanism in FRP and enable concise descriptions of reactive behavior. In practice, however, Emfrp programs are compiled into C and combined with platform-dependent input/output components written in C or C++. Consequently, developers must debug the resulting mixed C/C++ program using conventional debuggers such as GDB, even though the application logic is written in Emfrp. This situation creates an abstraction gap between the source-level FRP program and the executable system. This paper presents a multi-mode debugging framework for Emfrp-based embedded applications. The framework supports debugging at the level of Emfrp abstractions while also allowing inspection of platform-specific C/C++ I/O code. Our approach uses a source code mapping technique that relates Emfrp constructs to corresponding locations in the compiled program. A case study on an ESP32 microcontroller using representative debugging scenarios demonstrates improved debugging efficiency.

Metaphorical language remains a major challenge for multilingual natural language processing because successful interpretation and translation require reasoning beyond literal lexical meaning. Existing research has largely investigated metaphor detection, machine translation, and translation evaluation as separate tasks, while little work has explored how these components can be integrated into a unified computational framework. This PhD proposal aims to develop an end-to-end framework for multilingual metaphor processing consisting of three complementary research directions: (1) robust metaphor detection across languages, (2) metaphor-oriented translation evaluation for both human assessment and automatic quality estimation, and (3) joint modelling that connects metaphor detection with translation evaluation. The proposed research will combine linguistic theory with recent advances in large language models to develop new datasets, annotation methodologies, evaluation benchmarks, and automatic evaluation approaches for metaphor-aware machine translation. The expected outcome is a unified framework that improves both the development and evaluation of multilingual NLP systems when processing figurative language.

Blood-brain barrier permeability (BBBP) prediction is a critical screening task in central nervous system drug discovery, where candidate molecules must be assessed for whether they can cross, or should be prevented from crossing, the blood-brain barrier. However, this task remains challenging because of limited, class-imbalanced datasets and sensitivity to molecular structure. Recent advances in deep learning have established graph neural networks (GNNs) as a powerful approach for molecular representation learning, while pre-trained molecular GNNs provide transferable knowledge for downstream tasks. However, full fine-tuning is often parameter-inefficient and prone to overfitting, whereas existing parameter-efficient fine-tuning (PEFT) methods mainly adapt node features or the two-dimensional covalent graph, limiting their ability to capture three-dimensional geometry and second-order interactions. To address these limitations, we propose BBBP-GeoPEFT, a geometry-informed PEFT framework for pre-trained molecular GNNs. BBBP-GeoPEFT constructs distance-based graphs at multiple cutoffs and their corresponding line graphs from molecular conformers to capture spatial atom and second-order edge interactions. Lightweight auxiliary geometric graph encoders generate cutoff-specific representations, which are incorporated into each pre-trained layer through node-wise cutoff attention and gated residual connections. This design preserves pre-trained knowledge while incorporating permeability-relevant geometric information with a small trainable-parameter budget. Experiments on a curated BBBP dataset show that BBBP-GeoPEFT achieves competitive performance compared with full fine-tuning and representative PEFT baselines. Under both random and scaffold splitting, BBBP-GeoPEFT achieves competitive or improved ROC-AUC and accuracy in most experiments while updating only 10.1% of the model parameters.

Graph inference over relational data can expose sensitive edge information, and this risk becomes more severe in dynamic graphs, where repeated model updates cause privacy loss to accumulate. We formulate Edge-level Differentially Private Dynamic Graph Inference (EDG) and propose PriDyG, a private inference framework that combines GNN-based structural learning with LLM-based semantic reasoning. PriDyG introduces incremental private multi-hop aggregation, which buffers newly arrived edges and processes each edge exactly once. By parallel composition, the total privacy cost equals that of a single static release, independent of the number or schedule of model updates. Compared with geometrically decaying budget allocation, incremental aggregation avoids exponentially increasing noise while preserving exact one-hop signals and at least half of two-hop information transfers. PriDyG further complements privatized GNN outputs with LLM predictions derived solely from node text, incurring no additional edge-level privacy cost. Experiments on four benchmarks for node classification and link prediction show that PriDyG consistently outperforms geometrically decaying baselines under the same privacy budget and matches the utility of naive per-update retraining while reducing cumulative privacy cost by up to three orders of magnitude.

The dynamical Lie algebraic (DLA) theory of variational quantum algorithms (VQAs) predicts commonplace exponentially vanishing loss and gradient variances for sufficiently deep parametrized circuits. In this work, we show that these predictions fail dramatically in the shallow-circuit (and particularly constant-depth) regime for the Quantum Approximate Optimization Algorithm (QAOA) applied to the maximum independent set (MIS) problem. In a large-scale numerical study across $\sim$23,000 problem instances, we find that barren plateaus are rare, while landscapes whose variances polynomially increase with system size---which we term "cragged terrains"---are common across graph families. This aggregate polynomial growth persists both for generic, low-symmetry random graphs and for highly symmetric vertex-transitive graphs, indicating that DLA-based variance predictions do not describe landscape scaling in this regime. As a stopgap alternative to the theory, we train empirical hardness models to predict instance-wise hardness metrics for QAOA-MIS. While these models generalize poorly, they nonetheless recover the correct landscape scaling class (barren plateau vs. cragged terrain) with high fidelity. Taken together, our results identify shallow QAOA for MIS as a prototypical setting in which asymptotic, unitary-design-centric predictions may be fundamentally insufficient to describe shallow variational quantum algorithms more broadly, emphasizing the need for more empirically-informed models of VQA loss landscapes.

Multimodal large language models (MLLMs) make grounded predictions in real-world scenes by combining visual and textual cues, yet existing benchmarks rarely reveal how they arbitrate between these evidence sources when they conflict. We introduce SIGNPOST-Bench, a controlled counterfactual benchmark for evaluating text-vision conflict resolution. Each source image is transformed into a counterfactual quintuplet of Original, Blank, Similar, Random, and Adversarial variants. Synthetic, localized scene-text interventions are designed to preserve non-textual content, enabling paired measurements of changes in localization performance and directed shifts toward geographic targets introduced by conflicting text. SIGNPOST-Bench contains 5,111 counterfactual groups and 25,555 image variants from four datasets. We evaluate 20 MLLMs from seven providers. Compared with Original images, Adversarial variants raise median localization error from 282 km to 1,347 km, a 4.8-fold increase. Among geocodable adversarial samples, 6.5-20.1% of predictions lie less than 50 km from the injected target across models, and every evaluated model exhibits a positive mean paired reduction in target distance from Blank to Adversarial. Compatible, unrelated, and conflicting text replacements produce distinct effects on model predictions, while clean-input localization performance does not fully predict robustness to conflicting text. These results establish visual geolocation as a continuous diagnostic of scene-text arbitration and provide a controlled framework for evaluating how MLLMs resolve conflicting multimodal evidence.

Multi-head attention layers produce vector representations that support multiple downstream tasks. We establish bounds on the number of heads required in two simple and concrete multi-task scenarios. In the first scenario, a vector representation is sought so that linear predictors can compute both the smallest and largest numbers in a given list. In this case, it is known two attention heads with small embedding dimension and bit precision level suffice. We prove that a single attention head requires exponentially higher embedding dimension or precision level. In the second scenario, a vector representation is sought so that a polynomial threshold function can compute the XOR of a given string of $n$ bits. This scenario is analogous to the first one for $n=2$, since XOR is readily computed by a linear function using a vector representation that encodes both the AND and the OR of the two bits. We observe that $n$-bit XOR requires the product of the number of heads and the polynomial degree to be at least $n$, and we construct multi-head attention layers that match this lower bound. These results generalize to arbitrary (symmetric) Boolean functions, where the bound is given in terms of the threshold degree.

Superhuman game engines in domains like chess have made expert-level evaluations easily accessible, yet they communicate what is true without the natural-language explanations that make such expertise educationally useful to experts and non-experts alike. Large language models could, in principle, bridge this gap, but they frequently hallucinate due to limited domain-specific knowledge, and standard reference-based or LLM-as-a-judge frameworks cannot reliably detect these errors. In this work, we present ACT-Eval, an evaluation framework that decomposes chess commentary into atomic claims and routes them to engine-supported tools and expert-annotated gold references to assess factual correctness, conceptual coverage, and move-quality judgment. We release a benchmark of 325 position--move pairs spanning pedagogical, tournament, and critical positions, including 125 positions with expert-verified gold atoms and a five-class error taxonomy. Evaluating leading proprietary and open-weight models, we find that factual hallucinations remain pervasive in chess commentary: GPT-5.4 without tools produces incorrect sub-claims 22.0% of the time, while smaller open-weight models exceed 40%. Although tool augmentation substantially improves factual correctness and move-quality assessment, coverage of expert strategic and tactical ideas remains limited across all models. Human calibration shows that ACT-Eval's factual judgments fall within the observed range of inter-human agreement, while its coverage scores correlate strongly with human assessments of strategic completeness.

Spectral submanifold (SSM) reduction has emerged as a mathematically principled route to reliable nonlinear reduced-order models, capturing dynamics beyond the reach of linear techniques such as Dynamic Mode Decomposition (DMD). The computation of SSMs, however, remains computationally expensive, particularly for high-dimensional systems. In this work, we introduce equivariant spectral submanifold (eSSM) reduction, a novel extension of the SSM framework that explicitly incorporates symmetries of the full-order model into the reduction process. We establish the mathematical foundations of this approach by showing that SSMs are naturally equivariant submanifolds and that the associated charts and reduced dynamics inherit the appropriate induced group actions. Building on this framework, we develop a novel equivariant SSM reduction algorithm that exploits these symmetries to achieve substantially faster computations while also improving model robustness. We demonstrate the advantages of this approach on several benchmark problems including a test from the Common Task Framework for Science.

We study the problem of aligning data from multiple modalities into a shared representation space, focusing on settings where strong pretrained unimodal encoders are available but cross-modal paired data are scarce. We propose a structure-preserving alignment framework, joint kernel entropic Gromov--Wasserstein Optimal Transport (JK-EGW), which maps multiple modalities into a common latent space by minimizing a quadratic optimal transport objective. JK-EGW leverages fine-grained similarity relationships within and across modalities to construct a global affinity kernel instead of relying on raw feature-space distances. Our framework naturally provides explicit control over the geometry and distribution of the latent embedding. On the theory side, we establish parametric sample complexity rate of $n^{-1/2}$, matching the corresponding rates for standard, entropic and Gromov--Wasserstein optimal transport. On the algorithmic side, we derive a scalable alternating procedure to solve JK-EGW with entropic optimal transport (EOT) updates through a low-rank kernel approximation and a variational lifting. This lifting scheme effectively relieves the burden of a quadratic objective, and allowing us to take the advantage of existing EOT solvers. Empirically, we focus on post-hoc alignment of embeddings from pretrained encoders in data-scarce regimes, and show that our proposed method achieves improved multimodal retrieval performance compared to existing alignment baselines.

Biological systems must regulate competing needs under limited perceptual bandwidth, where sharpening one estimate costs the capacity to sharpen the others. Any fixed-budget system therefore has to decide where to allocate its perceptual precision. We study this in a foraging agent that must keep several bodily needs satisfied to survive, modelled with active inference. At each step it reads its own body-state beliefs, identifies the most-needed channel, and reallocates a fixed budget of interoceptive precision toward it, so that the same precision-shaped likelihood feeds both belief update and planning. In AffectWorld, a four-channel foraging gridworld, this selective allocation more than doubles learning-phase survival at matched budget against a uniform-precision agent ($0.414$ vs $0.199$ across 11 layouts, $n{=}32$ seeds each, paired cluster-bootstrap $p \leq 10^{-4}$). Two further results sharpen the mechanism. The benefit runs through planning as well as perception, since denying the shaped likelihood to the planner alone removes about half of it. It is also need-aligned, since aiming precision at the least-needed channel does worse than spreading it evenly. The attended channel additionally learns its own dynamics about twice as fast, and stays ahead even at matched observation count, a behavioural trace of the same precision routing, visible in learning speed, not survival.

Deep learning models for scientific spatio-temporal downscaling often minimize reconstruction error while failing to preserve physically meaningful multi-scale structure. For sea surface temperature prediction, this can yield outputs that are numerically plausible yet overly smooth, missing mesoscale variability critical to regional ocean dynamics. Existing methods often focus on pixel-wise objectives or single-context conditioning, which limits their ability to preserve spectral fidelity and generalize across regions. To address this, we propose EddyFlow, a representation learning framework for kilometer-scale sea surface temperature downscaling that balances predictive accuracy, scale-dependent structure, and regional generalization. EddyFlow is trained on the Gulf of St.~Lawrence and evaluated in zero-shot and few-shot settings on the Bay of Fundy and the Gulf of Mexico. EddyFlow demonstrates that physics-informed representation learning reduces zero-shot RMSE by 21%, achieves up to 85.6% skill relative to persistence on unseen domains, and maintains near-ideal spectral fidelity with a PSD ratio of $\approx 1.00$.

We propose a family of random feature maps for scalable kernel machines on low-dimensional subspaces, ie on the Grassmannian manifold. Such representations are useful when data classes or clusters are well described by the span of a few samples. Classical Grassmannian kernels, including the projection and Binet-Cauchy kernels, require full Gram matrices, which leads to prohibitive computational and memory costs for large high-dimensional subspace datasets. We address this limitation using random features based on rank-one projections of subspace projection matrices followed by bounded non-linear transforms, either periodic or binary, to control the resulting distributions. We show that inner products in the random feature space approximate well-defined rotation-invariant Grassmannian kernels that depend only on the principal angles between subspaces. When the number of features is sufficiently large relative to the intrinsic subspace dimension, the approximation holds uniformly over all fixed-dimensional subspaces with high probability. For periodic transforms, the approximated kernel has a closed-form expression with tunable behaviour between inverse Binet-Cauchy and Gaussian-type regimes. Binary transforms yield compact one-bit subspace features, although no closed-form kernel is known. Structured rank-one projections based on randomised fast Fourier transforms further reduce computation without sacrificing practical accuracy. Experiments on synthetic data and ETH-80 classification tasks show that these features accurately preserve Grassmannian geometry while reducing computation, memory, and storage. Rank-one embeddings therefore provide a practical and scalable alternative to classical Grassmannian kernels.

Turbulence is a central testbed for machine learning on physical dynamics because its governing laws are known exactly. However, most existing studies remain in 2D, while 3D turbulence has fundamentally different physics and is far more costly to simulate. Existing 3D resources also typically provide only one realization per configuration, making it difficult to distinguish learning the dynamics from fitting the statistics of a single flow. In this paper, we introduce TIDE (Turbulent Incompressible DNS Ensembles), a 256^3 DNS corpus and benchmark for 3D incompressible turbulence, with 15 configurations on eight controlled axes, independent ensembles, pressure fields, and equation-level verification. The benchmark includes five tasks, standardized learned baselines, controlled generalization splits, and physical-fidelity metrics alongside pointwise error. Across the main forecasting configurations, current learned models barely outperform persistence and still make about twice the error of a spectral solver given the true equations. Moreover, lower pointwise error can coincide with severely distorted small-scale dynamics, showing that accuracy alone does not ensure physical fidelity. Generalization results further show that most regime shifts reflect limited training coverage, whereas forced-to-decay transfer exposes a missing conditioning variable: operators trained under forcing continue to predict driven evolution when the external drive is removed. Closing these accuracy, fidelity, and conditioning gaps is the central open problem made measurable by TIDE.

Software vulnerabilities are persistent, but validating them remains difficult: a Proof-of-Vulnerability (PoV) requires a concrete input that triggers the vulnerable behavior, yet public triggering inputs are often unavailable for disclosed vulnerabilities. Existing techniques make different tradeoffs in effectiveness, scalability, cost, and controllability, leaving room for complementary designs. To complement them, we present POVGEN, a low-cost neuro-symbolic framework that makes PoV generation cost-effective via semantic focusing and LLM-guided constraint reasoning using open-weight models. POVGEN first localizes vulnerability-relevant regions (utilizing patch information if available), then performs path-sensitive reachability analysis, and finally generates PoVs by extracting and solving constraints with LLM-guided reasoning backed by an SMT solver. POVGEN successfully generates PoVs for 78.98% of vulnerabilities in a recent benchmark, outperforming fuzzing (up to 50.20%) and symbolic execution (2.45%). On 250 real-world CVEs without public PoVs, it generates valid PoVs for 74.80% of cases and reproduces 65.1% when without patch information. The fine-tuned open-weight models match frontier commercial LLMs on key sub-tasks (i.e., the core constraint-reasoning steps) while running locally at no per-sample API cost. Applying the generated PoVs revealed six flawed patches in disclosed CVEs (all subsequently fixed) and five previously unreported vulnerabilities (of which four have been confirmed and fixed by the developers).

The growing diversity of code clone types, from syntactic copies to cross-language semantic clones to AI-generated duplicates, has created a fragmentation crisis in clone detection. Current deep learning detectors are domain specialists that degrade significantly outside their training distribution, with F1 drops exceeding 70% across domains. Deploying multiple specialized models is impractical, yet training a single cross-domain detector requires simultaneous access to all training data. To address this, we investigate model merging, a family of post-hoc techniques that operate solely on trained checkpoints. We evaluate parameter merging with five task-vector methods, architecture merging via greedy layer stitching, and cross-tokenizer alignment across four code models, three benchmarks, and twelve configurations. Same-base TIES merging creates effective cross-domain detectors, validated across two model families and three random seeds, reaching 0.865 combined F1 on UniXcoder, 93% of multi-task performance without any training data at the merging step. WUDI achieves the highest in-distribution combined F1 at 0.899, but TIES generalizes better to unseen AI-generated clones, making it our recommended method. Cross-base merging yields only marginal and high-variance gains across all five methods, indicating that task vector compatibility through a shared pre-trained base is the binding factor for effective merging. Merged detectors also outperform zero-shot code LLMs on GPTCloneBench at lower inference cost and generalize up to 4x better than multi-task training to unseen AI-generated clones, suggesting a trade-off between in-domain performance and OOD robustness. This work provides one of the first systematic empirical studies of model merging for software engineering and a practical recipe for building cross-domain clone detectors.

Existing studies on self-supervised learning for white-box networks typically decouple the derivation of white-box networks via optimization algorithms from self-supervised learning paradigms. In this work, we instead revisit the two components from a joint perspective. The LeJEPA-based self-supervised framework assumes an isotropic Gaussian distribution as the optimal embedding distribution for downstream tasks, which is conceptually equivalent to the expansion term $R(Z)$ in the sparse rate reduction objective guiding white-box Transformer optimization. Building on this observation, we use the LeJEPA self-supervised paradigm to optimize $R(Z)$, and derive the remaining terms $R^{c}(Z\mid U_{[K]})+λ\lVert Z\rVert_{0}$ via the alternating direction method of multipliers (ADMM) into an attention-only Transformer that dispenses with the ISTA structure or MLP layers of the original design. Experimental results demonstrate that our attention-only white-box Transformer achieves classification accuracies of $88.88\%$ on CIFAR-10 and $63.54\%$ on CIFAR-100 at the Base scale under the LeJEPA self-supervised paradigm, while the original white-box Transformer CRATE achieves classification accuracies of $89.18\%$ on CIFAR-10 and $63.56\%$ on CIFAR-100. Our model achieves competitive performance while reducing the parameter count by roughly $31\%$. Beyond the white-box setting, we further investigate standard ViTs and find that replacing all MLP blocks with ReLU activations under knowledge distillation removes approximately 66\% of the parameters while preserving competitive accuracy, motivating further investigation into the potential redundancy of MLP modules in standard ViT architectures.

Physics-informed neural networks (PINNs) often require high-accuracy quasi-Newton refinement to obtain reliable partial differential equation solutions, but their residual objectives can exhibit indefinite, nearly singular, and poorly scaled local curvature. Regularized quasi-Newton methods provide established mechanisms for stabilizing secant models, while self-concordant methods provide local-metric rules for curvature-dependent step selection. Building on these two lines of work, we propose SCORE, a self-concordance-inspired quasi-Newton method with decrement-coupled shifted secant geometry for PINN training. Its distinguishing mechanism is that a single quasi-Newton decrement computed from the learned inverse metric jointly determines a strong-Wolfe-tested candidate step and an adaptive shift used to define the next secant geometry. The shifted displacement represents the action of an averaged shifted metric along the accepted step, while requiring neither Hessian construction nor Hessian-vector products. Under a local spectral-equivalence condition, we show that the quasi-Newton decrement and candidate step remain comparable to their counterparts in a positive shifted metric, and recover the normalized self-concordant rule in the matched-metric case. Strong Wolfe acceptance, fallback line search, and standard curvature safeguards provide globalization without modifying the underlying PINN objective. Experiments on the viscous Burgers, Kuramoto--Sivashinsky, Korteweg--de Vries, and complex Ginzburg--Landau equations show that SCORE attains lower final errors than the tested BFGS and self-scaled Broyden baselines. The Burgers ablation further indicates that shifted curvature stabilization and decrement-based step selection make complementary contributions to high-accuracy refinement.

Human evaluation of AI systems and digital products is costly, slow, and difficult to scale. Offline evaluations are more scalable but often abstract away human diversity and interactive behavior. We therefore introduce MatrAIx, a population-scale simulated-user evaluation infrastructure for testing AI systems and digital products with heterogeneous users. MatrAIx has three core components: First, Persona 8B contains 8.3 billion persona records represented by 1,290 categorical dimensions. Records are either sampled from a dependency graph that preserves correlated attributes or derived from human-authored profiles. We release a quality-filtered coreset of approximately 1 million personas, comprising 599,847 human-grounded and 400,000 synthetic records. Second, the MatrAIx Playground provides four environments in which diverse users evaluate and interact with digital products: Survey, AI Chatbot, Web, and App. Third, MatrAIx provides 1,010 application tasks spanning more than 25 domains, including Commerce, Software, Finance, and Healthcare. We conducted 18,189 evaluation trials across eight representative tasks. Persona agents were powered by three LLMs: Claude Opus 4.8, GPT 5.5, and Claude Haiku 4.5. The resulting feedback captures how decisions and preferences vary across persona backgrounds, including hesitation after a price increase, willingness to continue after an AI assistant fails, and latency tolerance. We conducted two main validation studies: First, a 400-trial controlled study evaluated persona adherence across ten behavioral attributes and all four environments. The declared behavior was expressed or correctly suppressed in 366 trials (91.5%). Second, human and LLM judges evaluated the extraction quality of human-grounded personas. Overall, MatrAIx provides an end-to-end infrastructure for evaluating AI systems and digital products with diverse simulated human users.

State estimation for nonlinear dynamical systems is commonly performed with the Unscented Kalman filter (UKF), which propagates the state moments through deterministic sigma points and reports a posterior covariance at every step. In practice, however, unknown and time-varying noise statistics and model mismatch degrade both estimation accuracy and covariance calibration. Existing learned filters improve accuracy but are largely built on the extended Kalman filter and either forgo an explicit covariance or learn uncertainty without correcting mismatch-induced gain bias. This paper introduces the Unscented KalmanNet (UKN), a hybrid recursive estimator that augments the UKF with two structurally distinct learned components while preserving its explicit sigma-point covariance recursion. NoiseNet predicts time-varying process and measurement covariances as bounded multiplicative corrections to fixed baselines, guaranteeing positive definiteness, while GainNet applies a bounded residual correction to the analytical gain. A calibration-aware training objective combines state error with covariance- and innovation-consistency terms through adaptive weights, jointly optimizing accuracy and calibration. UKN is benchmarked against UKF, KalmanNet, and Bayesian KalmanNet on three synthetic systems and UZH-FPV real-flight data. It achieves the lowest aggregate state-estimation error in all four examples and reduces RMSE by $26.4$-$49.7\%$ compared to UKF in the synthetic cases. Leave-one-sequence-out cross-validation over 11 flights shows $22.4\%$ and $34.3\%$ reductions in mean position and velocity RMSE, respectively. UKN also yields the lowest fold-to-fold variability, with normalized NEES and empirical coverage closest to nominal values among the other filters.

Financial sentiment classifiers are commonly evaluated against human labels, but strong linguistic performance does not necessarily imply economically useful return predictability. This study separates these questions through two experiments. First, we construct a unified three-class benchmark from five financial text datasets and compare TF--IDF Naive Bayes, off-the-shelf FinBERT and Financial-RoBERTa encoders, zero-shot Qwen2.5-7B, and QLoRA-adapted Qwen2.5-7B, LLaMA3-8B, and Mistral-7B models. Mistral-7B achieves the best test accuracy (0.8840) and macro-F1 (0.8771), while QLoRA raises Qwen2.5's macro-F1 from 0.7274 to 0.8615. An inverse-frequency class-weighted loss does not improve Qwen2.5. Second, we evaluate economic validity on a temporally separate 2019 Benzinga sample containing 10,637 unique headlines and 13,115 headline--stock observations for a fixed S\&P~100 universe. Model probabilities are converted into continuous sentiment scores, aggregated by stock and signal date, and aligned with next-session returns over one-, two-, three-, and five-day horizons. All seven downstream models produce positive but small mean rank information coefficients at the one-day horizon; the largest is 0.0143 for FinBERT. None of the 28 model--horizon tests remains significant after Newey--West inference and false-discovery-rate correction. Portfolio results likewise fail to establish a robust advantage for the best-performing classifiers. The findings show that QLoRA is effective for financial sentiment adaptation, while also documenting a clear gap between classification accuracy and tradable cross-sectional signals.

Recent years have witnessed an explosive trend of scaling ego-centric human videos for robot manipulation, yet it remains unclear which data actually benefits dexterous manipulation. We present SiMDex, a similarity-based data mining framework that casts human data selection for VLA post-training in dexterous manipulation as a recommendation problem. For each robot demonstration, SiMDex employs a three-layer recall-ranking-re-ranking pipeline to extract task-relevant subsets from a pool of ~32M egocentric human samples, operating in a morphology-agnostic action space that requires no changes to VLA architecture or training. Against a strong baseline trained with an equal amount of randomly sampled human data, SiMDex uses only ~1.49M mined samples (<5% of the pool) yet improves the overall success rate from 47.7% to 61.1%, showing that selective curation outperforms indiscriminate data mixing.

Source code summaries have traditionally been evaluated from a human developer's perspective, with quality determined by how closely they resemble developer-written references and how well they align with human preferences. But this overlooks a growing reality: LLM-based tools and agents increasingly consume code summaries as inputs for software engineering (SE) tasks, and what makes a summary useful for a consuming agent on a task remains largely unexplored. To bridge this gap, we propose SONAR, a reference-free framework that evaluates source code summaries along four dimensions: Correctness, Abstraction, Conciseness, and Fluency. Rather than optimizing for a pre-written "gold standard", SONAR introduces a novel code regeneration-based approach that uses a summary to regenerate code and leverages that reconstruction as a quality signal of the summary. This provides an empirical grounding that requires neither a reference summary nor the subjective judgment of humans or LLMs. We evaluate SONAR's dimensions on their ability to influence LLM performance across four downstream SE tasks. We find that Correctness, followed by Abstraction, significantly correlates with LLM performance, with correlations up to 14X higher than the best baseline. Conciseness and Fluency, though widely valued by human developers, remain mostly insignificant to an LLM consumer, suggesting that what makes a summary useful is task- and consumer-dependent. Through a large-scale evaluation of 11 popular LLMs using SONAR, we further identify the strengths and weaknesses of different models across each quality dimension, while offering insights to facilitate future research on task-aware summarization.

Language models (LMs) offer strong textual representations for electronic health records (EHRs), but they encode patient sequences in isolation and provide limited explainability. Graph neural networks (GNNs) complement LMs by incorporating inter-patient relationships and enabling reference-patient attribution, yet they rely on high-quality patient representations. We propose Patients-like-me (PLM), a unified LM--GNN framework that integrates local patient semantics with global cohort structure. To train PLM efficiently, we introduce a Variational Expectation-Maximization algorithm that alternates LM and GNN updates under a supervised variational objective. Extensive experiments on MIMIC-III and MIMIC-IV show that PLM consistently outperforms state-of-the-art methods, with improvements generalizing across encoder-only and decoder-only LM backbones. These gains are achieved with only modest additional computational overhead. PLM also provides reference-patient explanations by retrieving influential similar patients, while edge-masking experiments confirm that the highest-ranked references have the greatest impact on model predictions.

Closed source agent skills may encode proprietary instructions, scripts, constants, and data. Providers may offer their capabilities as services while keeping the underlying packages hidden. Prior work focuses on prompt injection attacks that directly disclose these artifacts, and existing defenses accordingly aim to prevent such leakage. However, preventing file disclosure does not prevent users from recovering the functionality those files implement. This raises a fundamental question: can a user reconstruct a skill's functionality through ordinary use while its files remain hidden? We study behavioral skill reconstruction (BSR), in which an attacker uses valid task requests and observed responses to build a functional clone of a hidden skill. We introduce SkillClone, a black-box attack that clones a target skill by forming an interface hypothesis from its public advertisement, issuing structured benign probes, synthesizing an executable replica, and iteratively repairing it through differential validation against the victim skill. Across 30 skills spanning rules, tables, procedures, and algorithms, SkillClone achieves exact or partial recovery on held-out inputs for several targets. Iterative requerying closes gaps missed by single-round reconstruction. Because SkillClone uses only legitimate interactions, disclosure-focused defenses provide limited coverage, and less detailed skill descriptions offer limited protection. These results show that file secrecy alone does not ensure functional secrecy. Defenses must also limit cumulative information leakage from ordinary use.

Deploying pre-trained perception models in novel environments degrades their accuracy under distributional shift, and assembling them alone does not recover it: combiners such as majority voting trade recall for precision and are brittle to coordinated failures. Prior metacognitive methods learn logical rules that flag a model's errors, but rely on hand-authored domain-knowledge cues (object-size priors, segmentation masks) that do not transfer to genuinely novel scenes. We show that this metacognitive layer can be learned without any domain knowledge by exploiting vector-space geometry: per-model Label Vector Pools (LVP), built from each model's own training embeddings, yield error-detection rules from the geometry of detections relative to training-determined prototypes, reaching parity with domain-knowledge rules to within $0.002$ every F1 on test set. Because the approach remains neurosymbolic, these geometric rules share a single logical framework and can still be complemented by domain knowledge when available. We frame the fusion of multiple imperfect ViT-based detectors as a consistency-based abduction problem solved at test time by an exact Integer Program (IP) and a polynomial-time heuristic. On an aerial-imagery benchmark of 15 weather-shifted test sets and six ViT detectors, our domain-knowledge-free layer matches the strongest majority-vote variant on clean data (within $0.005$ F1) and, unlike every majority-vote baseline, retains its performance under a coordinated label-flipping attack: at a $90\%$ flip rate it averages $0.42$ F1 versus $0.35$ for MV-Plurality (a $22\%$ relative gain) and attains the highest F1 on \emph{every} test set once the flip rate exceeds $0.4$

This paper presents a conceptual framework for developing an electronic explanatory dictionary of the Tajik language using large language models (LLMs). The relevance of the work stems from the absence of a comprehensive digital lexicographic resource for Tajik that is comparable in functionality to dictionaries for high-resource languages, and from the limited adaptation of modern natural language processing technologies to low-resource language systems. Based on a systematic survey of existing linguistic, statistical, and corpus resources, we propose a dictionary architecture that integrates modules for morphological analysis, lemmatization, semantic clustering, and dictionary entry generation using LLMs. The choice of subword tokenization is justified by the agglutinative nature of Tajik morphology and its high morphological variability, along with a parameter-efficient fine-tuning (PEFT) strategy suitable for limited annotated data. The novelty of the work lies in proposing the first holistic conceptual architecture of an explanatory dictionary for Tajik that unifies classical lexicographic methods, language statistics, and generative capabilities of LLMs into a single system. The practical significance of the study is the formation of a methodological foundation for developing a full-featured electronic dictionary that can serve both as a lexicographic tool and as a core resource for machine translation, automatic summarization, sentiment analysis, and other applied NLP tasks. The paper is intended for specialists in computational linguistics, lexicography, and developers of natural language processing systems working with low-resource languages.

When a language model follows an in-context conditional rule such as "if P(x) then A else B," does it assemble a runtime circuit with one module that tests the predicate and another that routes the answer? We probe this with activation patching under a four-donor design whose two swapped-rule donors make the condition and the answer word disagree, so each layer reveals which of the two it carries. Across three open models from two families and six languages sharing one fixed item bank, a mid-stack residual band carries the predicate's truth value: patching it reroutes the answer with predicate-outcome flip near 1.0 and mapping flip near 0.0, meeting a strict pre-specified isolation criterion in 17 of 18 cells, and the same localization holds across five predicate families. The router shows the opposite profile. A learned subspace flips A and B near-perfectly within the trained pair yet transfers to a new pair at approximately 0 in every model, while in Gemma-3-4B (the only model probed cross-lingually) it transfers at approximately 0.98 to the same pair in other languages. Under every probe we ran, the router direction is token-bound and non-transferable (largely answer-readout in Gemma, pair-specific in Qwen) rather than an abstract routing module. Test is modular; under these probes, route is not.

Fine-tuned code models often behave as domain specialists and can degrade sharply under distribution shift: in our clone-detection setting, a model trained on same-language clones drops 71\% F1 on cross-language clones, while multi-task training falls to 0.151 F1 on unseen AI-generated clones. Our companion study shows that post-hoc model merging can address this fragmentation, achieving 93\% of multi-task performance without training data while generalizing 4$\times$ better to unseen clone types. However, no practical tool exists that lets SE researchers diagnose checkpoint compatibility, merge specialists, validate results on SE benchmarks, and export models for deployment. We present \textbf{MergeSE}, an open-source CLI and web tool for training-free model merging of HuggingFace encoder checkpoints. While motivated by OOD generalization in clone detection, MergeSE supports SE classification workflows more broadly through a built-in registry of nine task types, including vulnerability detection, defect prediction, and code-smell detection. MergeSE provides five operations: \textit{tasks}, \textit{inspect}, \textit{merge}, \textit{evaluate}, and \textit{export}. It supports five merging algorithms, including TIES, DARE-TIES, Wudi, PCB, and averaging; detects cross-task classification-head mismatches; produces seedable deterministic outputs; and includes bundled benchmark samples for smoke-test reproduction. A full merge of two 124M-parameter checkpoints completes in under 5 seconds on CPU. End-to-end validation confirms that MergeSE-produced checkpoints match reference implementations and recover cross-domain performance from domain-specific specialists. The tool is available online at https://mergese.usask.ca, and the development repository is at https://github.com/srlabUsask/MergeSE.

Feature selection is a critical step in electronic health record (EHR)-based predictive modeling, where input variables are often high-dimensional, sparse, noisy, and redundant. Large feature sets not only increase computational burden and overfitting risk, but also make model interpretation difficult, leading to limited usefulness in clinical settings. In this study, we focus on diagnosis-related features and compare five feature selection paradigms for opioid use disorder (OUD) prediction: recurrence enrichment, NTK-motivated early gradient sensitivity, LightGBM-SHAP, Elastic Net, and large language model (LLM)-guided semantic selection. We use a unified preprocessing and evaluation framework and assess each method by downstream predictive performance, resampling stability, and representation of infrequent diagnosis codes. Our results demonstrate that performance improves with larger feature budgets with diminishing returns beyond a moderate size. NTK sensitivity provides the best overall balance of accuracy and stability, and LLM-guided selection contributes complementary clinically meaningful signals despite lower standalone performance.

Time series data are ubiquitous in practical applications, where classification (TSC) and extrinsic regression (TSER) have emerged as essential tasks for obtaining value from temporal sequences. While the literature has seen significant progress through feature-based and deep learning models, existing methods often focus either on the quality of feature extraction or on the intrinsic predictive power of complex architectures applied to raw data. This division creates a gap between the control offered by feature engineering and the automated performance of end-to-end models. This paper proposes TS2TabPFN, a framework that bridges this gap by integrating explicit feature extraction with TabPFN 2.5, a cutting-edge foundation model for tabular data, to leverage its predictive capabilities. Our extensive experimental evaluation demonstrates that TS2TabPFN significantly outperforms state-of-the-art models in TSER tasks with statistical significance, providing a robust and efficient alternative for TSC and surpassing most of the currently best-performing algorithms. These results suggest that combining foundation models with structured features overcomes single-paradigm limitations, establishing a new time series state-of-the-art.

Deep neural networks (DNNs) deployed on resource-constrained neuromorphic hardware face three concurrent challenges: the need for model compression through pruning, vulnerability to adversarial input perturbations, and susceptibility to hardware-induced weight faults such as stuck-at-zero errors. While each of these factors has been studied in isolation, their combined effects on model reliability have received little attention. This paper presents an empirical investigation of how pruning, adversarial training, and hardware fault injection interact to affect the robustness of convolutional neural networks. Using a compact three-layer CNN trained on MNIST, we conduct three experiments: (1) comparing the fault tolerance of naturally and adversarially trained models under simultaneous hardware faults and adversarial attacks, (2) evaluating how pruning affects adversarial robustness, and (3) characterizing the joint accuracy surface across fault rates, adversarial perturbation magnitudes, and pruning levels. Our results show that adversarial training improves robustness against input perturbations but increases sensitivity to stuck-at-zero weight faults. Contrary to intuition, pruning did not significantly increase fault sensitivity, and varying the pruning level had little effect across fault rates and attack strengths. These results highlight the need to jointly consider adversarial robustness and hardware reliability.

AI co-scientists can generate fluent materials-science hypotheses, but fluency does not show that an answer preserves a scientifically meaningful mechanism. We present a graph-to-answer mechanism-tracing case study for Graph-PRefLexOR-8B, a Qwen3-8B model adapted to expose distinct stages for brainstorming, graph construction, pattern extraction, and synthesis. We organize semantic backtracking, graph corruption, activation-based recovery measurements, and layer-by-token-region grids into a visual diagnostic workflow for inspecting this pathway. Across 100 open-ended materials-science questions, final answers remain closest to the model's own structured stages, especially synthesis. Under graph corruption, a full sweep over 37 residual-stream checkpoints, the embedding output and 36 transformer blocks, shows little mechanism recovery in the earlier transition region at layers 7--10, recovery instead concentrates in late synthesis and answer-start regions around layers 30 and 36. The workflow is intended to help scientists and model developers identify where a generated hypothesis loses or regains mechanism support before it is passed to downstream experimental planning.

Multilingual evaluations report accuracy at a single output-token cap, but languages need different numbers of tokens to express the same content, so the cap is a hidden experimental variable. We test whether the native-vs-translate gap on MGSM (German, Thai, Swahili) is a token-budget artifact for Qwen3-8B and Llama-3.1-8B-Instruct under four prompting strategies. The measured gap swings by up to 57 points across budgets, length normalization moves it by up to 38.9 points where the cap binds, and at tight caps normalization can reverse which strategy scores higher. We prospectively froze the sweep's three Qwen peaks and its near-zero value at 1024 and evaluated them on 540,000 independently hard-capped decodes: a second frozen family of six Holm-corrected tests rejects every null. The frozen test at $B^*=1024$ still fails to reject because native accuracy has already saturated there; above saturation, the residual difference is a strategy-performance gap, not an identified reasoning deficit. The same truncation channel prices a cost-ordered adaptation ladder: a cross-fitted Thai vocabulary extension closes 0.0 points of the gap at the frozen budget and 4.9 points where 19% of traces still truncate. A third frozen family varies only the announced budget at a fixed enforced cap; announcing 128 rather than 2048 tokens moves Thai native accuracy by 5.1 points, so accuracy is not a function of the enforced cap alone. A correct-emission timing identity computed from one long-cap run matches the three pre-specified MGSM peaks to 0.65 points and, in an exploratory Qwen-only analysis of three further benchmarks, tracks held-out items to 0.92 points, locating the peak exactly in five of seven cells. Treat the output cap as an independent variable and report accuracy across the budget regime, not at a single budget.

Recurrence plots are a time series data mining primitive applied to a variety of domains (e.g. star light curves, sound waveforms, CCT telemetry). This work proposes tensorized self-similarity matrices as a primitive for univariate time series datasets ($N\times n$) of $N$ time series of length $n$ with a subsequence window of length $m$, and whose tensor-based nature is naturally extensible to multivariate datasets. The proposed method to compute this primitive computes dot plots of size $N \times (n-m+1) \times (n-m+ 1)$ from these datasets, where the subsequent tensor is mined using tensor decomposition methods to mine for co-clustered patterns. We demonstrate our results in mass rapid transit, electricity demand, wind turbine, and car traffic data, finding the MINT pipeline effectively co-clusters cross-sensor patterns in highly regular datasets containing motifs at regular intervals.

Electroencephalography (EEG) analysis extends beyond assigning predefined labels to recordings; it requires workflows connecting natural-language instructions, signal processing, quantitative evidence, and scientific interpretation. We term this capability \emph{comprehensive EEG understanding}. Existing evaluations, however, primarily target isolated decoding tasks or system-specific demonstrations, leaving the competence of large language models (LLMs) insufficiently quantified. We introduce \benchmarkname{}, a unified benchmark for comprehensive, instruction-conditioned EEG understanding. It comprises four subsets---Foundational Analysis, Sleep Assessment, Neurocognitive Assessment, and Physiological Integration---covering 17 datasets, \numcases{} tasks, and over \numinstances{} real-data instances. Given an instruction and EEG recordings with optional physiological signals, a system must perform the analysis and produce a scientifically grounded report and, when required, artifacts. Outputs are assessed through numerical, categorical, set, sequence, semantic, and artifact validation. We evaluate \nummodels{} representative LLMs across more than 100K executions under two paradigms: autonomous code execution with CodeAct and structured agentic analysis with BrainAgent. Results vary substantially across models, subsets, difficulty levels, and execution paradigms, showing that EEG competence depends on the model and its operationalization. \benchmarkname{} provides a reproducible testbed for advancing LLM-based EEG understanding. The code and benchmark will be released soon, with evaluation results continuously updated.

Mapping paddy rice from very-high-resolution imagery in mountainous and hilly regions is difficult because terrain alters optical appearance and increases confusion with visually similar vegetation. We present TRNet for 0.5-m GaoJing-1 red--green--blue (RGB) imagery, a 5-m TanDEM-X digital elevation model (DEM), and derived slope. Separate visual and terrain encoders preserve modality-specific features. At an early encoder stage, Topographic Energy-Spectral Rectification applies terrain-conditioned low-frequency modulation and asymmetric high-frequency regulation to suppress steep-slope clutter and conditionally enhance compatible low-slope rice cues. The Topography-guided Paddy Structure Decoder combines semantic, rice--background boundary, and interior cues, using coarse terrain as context. Experiments used an Area A internal test set and held-out Area B, which had steeper terrain and lower rice prevalence. TRNet achieved rice intersection-over-union (IoU) values of 85.10\% and 80.68\%, exceeding the original Dual-Encoder U-Net by 9.15 and 18.83 percentage points, respectively. Ablation and slope-stratified results linked these gains to frequency rectification, structure learning, and fewer steep-terrain false positives. The results support coarse topography as a contextual prior for very-high-resolution paddy rice mapping.

Swap regret governs the rate at which uncoupled learning dynamics converge to correlated equilibria in multiplayer general-sum games. Under full-information feedback, the best previous guarantee when every player follows the same dynamics grows logarithmically in the horizon $T$. We construct uncoupled dynamics under which every player incurs only $O(nm^2\sqrt{\log m\log T})$ swap regret, where $n$ is the number of players and $m$ bounds the number of actions per player. To our knowledge, this is the first sublogarithmic individual guarantee in this setting, and it implies that the time-averaged product distribution of play is an $O(nm^2\sqrt{\log m\log T}/T)$-approximate correlated equilibrium. The key algorithmic choice is to combine the Blum--Mansour reduction with optimistic follow-the-regularized-leader using a hybrid regularizer that separately weights negative Shannon entropy and the log-barrier: the entropy controls the optimistic prediction error, whereas the log-barrier controls the transition-matrix movement through its Bregman divergence. A new sensitivity theorem for stationary distributions of Markov chains, which involves neither mixing parameters nor the smallest transition probability, transfers this control to the played strategies and yields a simpler analysis without local-norm or self-concordance arguments. The guarantee is preserved by an adversarially robust variant that additionally ensures $O(nm^2\sqrt{\log m\log T}+\sqrt{mT\log m})$ swap regret against arbitrary utility sequences, and by a horizon-free variant that requires no prior knowledge of $T$.

Agentic AI is increasingly used to coordinate planning, implementation, review, and testing in software development, yet it often offers limited transparency into its decisions and interactions. Many such systems also assume that users can effectively guide the AI's decisions and validate its outputs. This assumption poses a particular challenge for novices, who must simultaneously learn how agentic AI works, how to collaborate with it effectively, and how to evaluate its outputs critically. To address this challenge, we present \textit{AgentForge}, an immersive learning system in which novices take on one of four software-engineering roles: Task Planner, Patch Author, Code Reviewer, or Test Runner, within a multi-agent code-repair workflow. In each practice session, the novices perform their chosen role while AI agents perform the remaining three. Through role-based scaffolding and metacognitive support, AgentForge clarifies role-specific responsibilities, makes agent coordination and intermediate artifacts visible, and encourages novices to monitor and evaluate their decisions. In a study with 37 novice developers, participants achieved high task-completion rates with AI-agent support. However, interaction demands differed significantly across practices: the Code Reviewer practice required more interaction turns, reroutes, and completion time ($p_{\mathrm{adj}} = .004$) and was perceived as the most challenging. Participants nevertheless reported significant gains in their understanding of software repair and agent collaboration ($p_{\mathrm{adj}} < .001$). These findings suggest that AgentForge can help novices develop practical software-engineering skills while learning to collaborate with agentic AI more critically and effectively.

Label noise is common in medical imaging datasets due to factors such as inter-rater variability, annotation errors, and ambiguous cases. This can severely undermine the reliability and clinical effectiveness of machine learning models trained using those datasets. To address this challenge, we introduce Lightweight Noise Correction (LiNC), which adds a single trainable trust parameter per training sample and learns when to use the observed label and when to defer to the model during a standard training loop. The key idea is to train using a convex combination of the observed label and the model's own predictive distribution, controlled by a per-sample trust parameter. We show that the gradient of this objective drives trust values in opposite directions for clean versus noisy samples in the early training phase, yielding separable trust distributions. We use a 3-component Gaussian Mixture Model over the trust values to separate them into clean, ambiguous, and noisy cases and then execute a short soft-correction phase on the noisy cases and a final hard correction phase. Experiments on ten 2D datasets from MedMNISTv2 under label noise of up to 50% show consistent gains in accuracy and strong mislabel detection. LiNC adds negligible asymptotic overhead: the training-time complexity remains dominated by the base network, with additional memory growing linearly with the size of the training set.

Biomedical entity linking grounds mentions in clinical and scientific text to entities in a curated knowledge base (KB) with ontological structure, which supports downstream applications such as literature-scale information extraction and patient-record normalization. The task has several challenges at once: the KB contains large numbers of entities, mentions are often ambiguous, and gold labels follow annotation conventions specific to each corpus. To address these challenges, we propose PILOT, a three-stage framework made up of neighborhood-aware retrieval, dual reranking, and score fusion. The retriever injects ontological structure from both the query and KB side, by reformulating mentions and pooling entity embeddings. The retrieved pool is then scored from two complementary views, one over surface forms and one over context, and fused together. PILOT achieves the state of the art on average across five widely-used benchmarks and remains efficient at inference.

Existing reference-free methods for evaluating music perceptual quality alleviate the need for paired noisy-clean data, but they still rely on a background set, which is used to compute aggregated statistics of clean audio samples. In this work, we propose a novel approach that eliminates this requirement, achieving background-set-free and reference-free quality estimation using only a pre-trained Flow Matching backbone. We demonstrate that unconditional Flow Matching inversion via simple Euler integration is sufficient to detect various artificial distortions and accurately rank music generation models against human perceptual judgments. We introduce InvFlowFD, which performs flow inversion and compares a group of inverted samples to the prior distribution. We evaluate our method against prior work, quantitatively and with a thorough human study. Results suggest that InvFlowFD is highly correlated with human perception of sound distortions, as well as generative models' quality, while being more flexible and less restrictive than existing metrics.

Reusing embedded software functions across products is economically valuable but technically difficult: the same functionality implemented for two different microcontroller platforms can be entirely incompatible at the hardware level, even when the functions score above 0.90 cosine similarity and both pass SonarQube quality checks. Static analysis tools were designed to measure code quality, not hardware-domain compatibility, and have no model of peripheral interfaces, hardware abstraction layer (HAL) dependencies, or register-map constraints. This paper presents a domain-aware retrieval-augmented generation (RAG) pipeline for embedded C software reuse detection that addresses the hardware-compatibility gap directly. The pipeline enriches each function by extracting its existing inline comments, call-graph context, and a project README before embedding it with eight backbone models (MiniLM, MPNet, BGE, E5, GraphCodeBERT, OpenAI text-embedding-3-small, LLaMA 3 8B, StarCoder2 3B) acting as feature extractors. Four hardware-compatibility validators---covering peripheral token overlap, parameter count parity, call-graph dependency overlap, and structural branching pattern---filter candidates directly in the retrieval stack. Evaluated on six public embedded C software projects (184 functions, 4,815 above-plateau pairs), the pipeline reveals that SonarQube produces a 93.6% false-positive rate as a reuse filter, with 83.5% of failures caused by hardware-environment mismatches that static analysis cannot detect. Manual verification of 40 rejected pairs confirms 97.5% validator accuracy, and a diagnostic rule-injection variant identifies the dominant failure categories (McNemar chi-squared~=~294.0, p~$<$~0.001).

Video question answering requires models to ground language queries in visual evidence and, when necessary, reason over that evidence across time. Existing methods typically rely on long textual chain-of-thought rationales, even though many questions can be answered as soon as the relevant object, action, or frame is localized. We propose Dynamic Latent Reasoning (DyLaR), which first grounds a question in a short block of perception latents (continuous hidden states that encode query-relevant visual evidence), and then adaptively decides whether to append reasoning latents (continuous thoughts that reason over this evidence in latent space) before answering. DyLaR learns this behavior by grounding perception latents in verified visual evidence and distilling verified rationales into reasoning latents, followed by reinforcement learning that further refines when to reason. Across nine video benchmarks and four multimodal language model backbones, DyLaR improves average accuracy over same-backbone baselines while generating fewer than 20 tokens per query. On Qwen3-VL-4B, for example, DyLaR improves average accuracy over Qwen3-VL-4B-Thinking from 54.0 to 58.2 while reducing response length from 1,220.7 to 18.5 tokens per query. Ablations further show that grounded perception latents, rationale-supervised reasoning latents, and adaptive routing each improve accuracy.

Vision-based guidance of unmanned aerial vehicles (UAVs) toward unmanned ground vehicles (UGVs) supports cooperative aerial--ground robotics, but reliable continuous yaw estimation from onboard vision remains challenging because of sensing uncertainty, limited computation, and the need for interpretable control. Existing deep-learning and geometric-reconstruction approaches often require large datasets, external localization, or complex modeling assumptions, reducing transparency and deployment suitability on resource-constrained platforms. We present an interpretable fuzzy-inference framework that generates continuous yaw commands from low-dimensional features extracted from YOLO boxes: target centroid location, area, and aspect ratio. No explicit geometric modeling is required. A Mamdani fuzzy system serves as an interpretable baseline using a shoulder--triangle--shoulder input partition. It is followed by a first-order Takagi--Sugeno model with three antecedent membership terms per input, whose parameters are derived from training-set quantiles, yielding a compact 27-rule structure. Evaluation uses 6{,}169 labeled samples from a VICON motion-capture environment. Across five randomized train--test splits, the Takagi--Sugeno model achieves a test-set mean absolute error of $0.140^\circ \pm 0.003^\circ$, a root mean squared error of $0.200^\circ \pm 0.008^\circ$, and a maximum absolute error of $1.254^\circ \pm 0.121^\circ$. Within-threshold accuracies are $99.676% \pm 0.270%$ for $\pm1^\circ$ and $100.000% \pm 0.000%$ for both $\pm3^\circ$ and $\pm5^\circ$. Directional consistency between image-plane horizontal displacement and predicted yaw sign reaches $90.254% \pm 0.612%$. These results show that the framework is transparent, data-efficient, computationally lightweight, and suitable for real-time vision-based UAV guidance toward mobile ground targets.

Black-box optimization is a ubiquitous problem in science and engineering, often dealing with expensive objective functions with cheaper lower-fidelity proxies available. Multi-fidelity Bayesian optimization (MF-BO) is a principled approach to this problem, leveraging correlations across different fidelities when querying the objective. However, for many important MF-BO tasks, the true highest-fidelity function is prohibitively expensive to be part of the optimization loop. Nevertheless, practitioners often have gold standard data (observations of the highest-fidelity function) obtained from previous experiments that might provide information for the current task. For instance, in molecular optimization, chemists often pick the top-$k$ candidate molecules using various computer simulations, and later reveal their true objective function values. In this work, we demonstrate the suboptimality of standard MF-BO algorithms in the real-world scenarios above, even under ideal assumptions. Next, we mitigate this problem by incorporating historical high-fidelity data accompanied by task descriptors---which can be explicitly given or extracted from unstructured metadata. We demonstrate the effectiveness of our methods on synthetic functions, as well as real-world problems in chemistry and hyperparameter optimization.

Can a model look at a river delta and a lightning bolt and see that they share a structure? We introduce GEB-Bench, a benchmark whose unit is an abstract structural motif--self-reference, a strange loop, a Mobius twist--in the spirit of Godel, Escher, Bach. Each motif is told in several voices: a natural scene whose composition is the structure, a folk story whose telling enacts it through a mechanically checkable form device, a mathematical theorem, and a programmatic skeleton; surface parameters are declared nuisance variables and never scored. Motifs, voices, and the structural changes between them form a small cross-modal category, and GEB-Bench's tasks are its questions. Evaluating twelve open and proprietary models, we find that abstraction failure is lawful. The central finding is a gap between recognition and cross-voice mapping: models identify a structure within one voice far better than they carry it across voices; every model pays this tax, and mapping strong enough to narrow it appears only at the frontier tier. Two patterns support it. Errors align more strongly with the designed formal geometry than with measured perceptual geometries, and frontier models from different vendors converge on the same wrong answers; and surface complexity taxes every model that reads structure, with capacity buying headroom rather than immunity. GEB-Bench is fully generative and released with its pipeline.

Large language model (LLM) agents are increasingly used as personalized assistants in high-stakes domains such as financial advising, yet it remains unclear whether they can maintain and update an individualized user model over long horizons. Existing personalized-memory benchmarks primarily test factual retention or rely on weakly constrained model-generated trajectories, leaving event-driven preference adaptation underexplored. We introduce FinPerMA, an event-grounded benchmark that evaluates personalized memory against frozen longitudinal investor trajectories. Its generation pipeline combines deterministic, theory-informed impact rules, controlled LLM narration, and automated quality screening; a Post-Shock checkpoint isolates whether an agent has integrated a material event into its persistent user model. On 2,994 questions from 276 personas, seven frontier LLMs and up to seven memory configurations remain far from saturated: no full-context configuration exceeds approximately 0.47 overall accuracy or approximately 39% on multiple-choice questions. Attribution analysis shows that summary-based memory often preserves factual details while losing the preference signals needed for personalization; simple retrieval can therefore outperform purpose-built memory systems, with the gap widening after shocks.

Mixture-of-experts (MoE) networks pursue specialization through learned routers, gates, and load-balancing losses, yet at matched total-parameter budgets learned routers can underperform equal-weight No-Routing baselines. Is the bottleneck the routing algorithm, or the alignment between training-signal granularity and the target categories? We probe the question with SpecDrop, a fixed parameter-free routing scheme: each of $K$ branches receives weight $p_a$ for its assigned category and a small leakage $p_i > 0$ otherwise, merged through a category-independent fixed denominator, with no learned routing parameters and no auxiliary losses; the category label is required at inference. On vision tasks where each image has one superclass label (CIFAR-100 on ResNet-110; ImageNet-1K on ViT-S/16), SpecDrop reaches 79.23% on CIFAR-100 and 79.89% on ImageNet-1K, exceeding parameter-matched baselines that do not use the label (+4.75 over dense on CIFAR-100; +6.53 over the No-Routing+SE control on ImageNet-1K). These gains quantify what category supervision buys when deployed through routing -- not an advantage over label-aware deployments of the baselines: given the same label, masking a dense model's outputs is stronger for accuracy alone (85.2 / 83.7). SpecDrop's contribution is converting the label into trained-in modular structure: 58%/100% branch-category alignment, and masking gains of 0.00 (CIFAR) / +1.06 (ImageNet) -- the output-space restriction is largely internalized during training. On fuzzy partitions, where training units span multiple categories (SlimPajama-6B language modeling with a 30M Transformer; SuperNI instruction tuning over Llama-3.2-1B with LoRA), the routing mechanism reduces to the matched No-Routing controls within seed noise, the null our thesis predicts. Granularity alignment, not algorithm choice, localizes when routing helps. Code: https://github.com/Beryex/SpecDrop

Evaluating financial AI agents requires criteria aligned with real professional work. Existing rubric methods typically derive criteria from task prompts or model outputs, overlooking tacit standards visible only in practitioner deliverables. We introduce FinProBench, a benchmark for professional financial tasks, and Role-Grounded Rubric Construction (RGRC), a reusable pipeline that derives rubrics from deliverables produced by practitioners in the same role. RGRC comprises four stages: Deliverable Collection, Competency Extraction, Rubric Synthesis, and Validation. Its rubrics capture tacit standards, distinguish quality levels, and transfer across tasks within a role. Before analysis, we classified 57 occupations by deliverable genre into 30 prior-rich conventional roles and 27 prior-sparse role-specialized roles. Across all roles, Prompt-only nearly matches RGRC for conventional roles (89.2% vs. 90.7%), but RGRC substantially outperforms it for role-specialized roles (99.1% vs. 78.0%). This split indicates that prompt engineering can approximate rubrics when conventions are well represented in model priors, while professional grounding is essential for standards beyond those priors. FinProBench is built from 1,723 curated deliverables spanning 57 occupations, 8 financial sub-industries, and 161 deliverable types, and releases an initial evaluation set of 20 complete tasks covering 20 roles in 7 sub-industries. With heterogeneous LLM judges and role-level rubrics, human deliverables rank first on average (73.7 vs. 70.3, 70.2, and 69.6 out of 100), while all four systems show overlapping 95% confidence intervals and complementary strengths. Reusing rubrics at the role level reduces estimated per-task construction effort by 6.7 times relative to authoring each rubric from scratch.

LLM agents are increasingly applied to open-ended everyday requests that span work, study, and life. These tasks are long-horizon, cross-environment, and multimodal, forcing the agent to preserve goals and constraints across many steps while navigating heterogeneous tools and attachments. While prior work has addressed individual failure modes such as goals drift, states loss, and context overflow, whether a single harness can manage them jointly and remain effective across backends has received less study. We present OneDayAgent, a long-horizon harness for autonomous agents. OneDayAgent turns an open-ended request into a managed execution process that decomposes tasks into bounded subtasks, maintains execution memory under context pressure, and verifies and repairs the final deliverable. We evaluate OneDayAgent on AgentIF-OneDay across 104 tasks. With the GLM-5.2 backend, OneDayAgent sets a new state of the art with an overall score of 0.821. The same harness runs across five backend LLMs from three model families, indicating the harness generalizes across backends without tuning, even as different models induce distinct execution styles under the same workflow.

Aligning diffusion models with human preferences usually relies on a sparse terminal reward evaluated on the final generated samples, presenting a severe temporal credit-assignment challenge across the multi-step denoising process. We propose Latent Reward Registers, a mechanism that estimates terminal preference directly from intermediate noisy latents by prepending learnable, position-free register tokens to the input sequence of a frozen Diffusion Transformer (DiT). This independent readout mechanism extracts latent reward evidence without altering the generator's hidden states or velocity field. The resulting dense, differentiable reward signal throughout the full denoising process facilitates two alignment strategies. For training, Reward-Gradient On-Policy Distillation (RG-OPD) distills reward-guided updates along on-policy trajectories, bypassing the computationally expensive rollouts of standard policy gradients. For inference, Reward-Guided Sampling (RGS) steers trajectories via magnitude-matched reward gradients without parameter updates. Empirically, at high noise levels (u = 0.8), the registers reach the highest pairwise accuracy among the evaluated latent reward models. Furthermore, RG-OPD outperforms online reinforcement learning baselines while reducing GPU hours by up to 33x, and RGS establishes a new state-of-the-art among training-free methods, strictly enhancing both alignment and perceptual metrics. Code and weights are available at https://github.com/Guanys-dar/latent-reward-register

Tensor cross-concentrated sampling (t-CCS) bridges entrywise sampling and t-CUR slice-wise sampling by observing entries only within selected horizontal and lateral slices. Existing t-CCS completion methods, however, assume that the observations are free of gross corruption. In this work, we study robust recovery of a third-order low-tubal-rank tensor from partial t-CCS observations contaminated by sparse, arbitrarily large outliers. We propose Robust Iterative t-CUR (R-ItCUR), a tensor-native algorithm that partitions the sampled tensor cross into two exterior blocks and an intersection block, applies adaptive blockwise Welsch correction for outlier suppression, and updates the low-rank component through projected blockwise gradient descent. By operating directly on the sampled cross, R-ItCUR avoids reconstructing the full tensor throughout the iterations, resulting in substantial memory and computational savings. Experiments on synthetic tensors, cardiac MRI data, and three-dimensional seismic data demonstrate accurate recovery and strong robustness to sparse gross corruptions. The results further highlight the importance of explicitly exploiting the cross-concentrated sampling structure in robust tensor completion.

Accurate traffic forecasting is essential for proactive resource management in edge computing, where service demand evolves dynamically across both space and time. In practical cellular edge systems, traffic exhibits strong spatial correlations among neighboring service regions and long-range temporal dependencies driven by user mobility and application behavior. Existing recurrent forecasting approaches can capture short-term dynamics but often struggle to model long-horizon traffic evolution under non-stationary conditions. To address this challenge, we propose a spatiotemporal graph Transformer framework that jointly models spatial interactions and temporal dependencies for traffic forecasting in edge computing. The framework employs graph neural networks to capture spatial correlations among service regions and leverages Transformer-based self-attention to learn long-range temporal patterns from historical traffic observations. By decoupling spatial representation learning from temporal reasoning, the proposed approach provides an effective mechanism for large-scale spatiotemporal traffic modeling. Extensive experiments on a real-world cellular network dataset demonstrate that the proposed graph Transformer consistently outperforms recurrent graph-based baselines, including GCN-RNN, GCN-LSTM, and GCN-GRU models, across multiple forecasting horizons. The resulting forecasts enable more effective proactive resource provisioning and reduce overload risk compared with reactive management strategies. These results highlight the potential of graph-enhanced attention mechanisms for building intelligent and adaptive edge computing systems.

Long-context LLM decoding reads the key-value (KV) cache at every step. Loading it takes longer than computing attention over it, so throughput is bandwidth-bound. Hence, reducing the cache size can raise both decoding speed and serving capacity. The challenge is to reduce cache size while preserving the attention products, keeping reconstruction cheap, and using a fixed per-token bit count. At two bits per element, the most competitive methods rely on orthogonal transforms. However, existing techniques are either data-oblivious or use the query statistics without deriving the transform from a distortion criterion. Moreover, they rely on transforms built on top of random or Hadamard rotations, which equalize variances across entries rather than compacting energy, and fixed-width scalar quantizers, which are suboptimal at low rates. In this paper, we formulate KV cache quantization as a transform coding problem in which distortion is the error in the attention products. We derive closed-form optimal transforms for keys and values from calibration statistics, under a high-resolution model. We show that the optimal key transform is not orthogonal and satisfies a generalized Parseval relation: the attention-aware distortion becomes mean-squared error (MSE) in the transform domain. Thus, we can use MSE-optimal vector quantizers applied directly to the transformed key coefficients. To meet the fixed-width layout requirement, we show that grouping coefficients into equal-volume partitions makes equal-size codebooks attain the variable-rate optimum under the same high-resolution model. At two bits per element, our method, termed NOVA-KV, recovers most of the long-context retrieval accuracy lost by scalar quantization methods at comparable throughput.

Personalized Federated Learning (PFL) has emerged as a promising solution for intrusion detection in heterogeneous IoT environments, as it can improve local adaptation under highly Non-Independent and Identically Distributed (non-IID) data distributions. However, existing PFL methods often rely on client-side self-adjustment, which may lead to over-personalization and substantial degradation in out-of-distribution (OOD) attack detection. In this paper, we propose Federated Bandit Intrusion Detection (FBID), a novel adaptive PFL framework to address this limitation through server-side personalization control. In particular, FBID employs a contextual multi-armed bandit at the server to dynamically regulate each client's local training intensity according to its observed behavior and update quality. Moreover, FBID introduces a trust-based blending mechanism to derive client-specific interpolation coefficients between the global and local models, thereby preserving global attack-detection knowledge while still allowing beneficial local specialization. Through extensive experiments on the CICIoT2023 dataset under heterogeneous client distributions and OOD stress-test settings, we show that FBID improves individual client OOD Detection Rate (DR) by up to 7.66% and F1-Score (F1) by up to 5.08% (relative) over the strongest stable baseline, while also improving robustness to previously unseen attack classes.

When decoder language models are used as classifiers, predicted class probabilities depend on implementation choices, including the prompt template, verbalizer (label-to-token mapping), and scoring rule, that are rarely treated as experimental variables. We present a controlled evaluation of three Mistral-7B variants (Base, BioMistral, and Instruct) on PubMed RCT sentence classification (n=2000) under FP16, INT8, and INT4 precision using four answer-text prompt templates. Our primary finding is that the probability extraction protocol dominates apparent calibration. Switching from summed to mean token log-likelihood scoring reverses the calibration ranking between models: BioMistral average expected calibration error increases from 0.097 to 0.289, whereas Instruct decreases from 0.237 to 0.096, while accuracy changes by less than 1 percentage point for the specialized models but 4-6 percentage points for the base model. Prompt template choice produces accuracy differences of 7-24 percentage points, comparable to or larger than model-level effects. On one template, BioMistral outperforms Instruct although the overall mean favors Instruct by only 1.3 percentage points. For BioMistral and Instruct, INT8 quantization changes accuracy and F1 by only 1-2 percentage points relative to FP16, whereas the base model shows larger INT8 effects on some templates (up to +4.2 percentage points). INT4 produces heterogeneous but non-catastrophic effects. Temperature scaling reduces expected calibration error under summed scoring for both models but only for that scoring rule. A fine-tuned PubMedBERT reference achieves 82.7% accuracy but uses about 176000 labeled training examples, precluding direct comparison. These results demonstrate that prompt template design and scoring normalization are first-order experimental decisions when evaluating decoder language model calibration.

Automatically generating professional multimodal reports comprising both textual analysis and visual charts from structured tabular data is a critical challenge in data intelligence. Existing methods suffer from fixed linear pipelines and isolated subtask processing, which hinder joint optimization of factual accuracy, visual quality, and narrative coherence. To address these issues, this paper proposes MCTS-Report, a Monte Carlo Tree Search (MCTS)-driven framework that formulates multimodal table-to-report generation as a progressive construction process over a structured search space. The core idea is to decompose report generation into atomic actions, including chapter planning, visualization task identification, chart generation, insight organization, and narrative refinement, each executed by an LLM based on dynamic reasoning conditioned on the current report state. We use an LLM to generate step-by-step reasoning and actions during MCTS, storing the reasoning trajectory in each node for context-aware, coherent report construction. To guide the search, we design a multi-dimensional reward function that jointly evaluates numerical fact consistency (via SQL), chart quality, chart-text alignment, and structural completeness, while incorporating a diversity penalty to suppress repeated charts and a precondition check to prune invalid actions. We also construct MMRBench, a comprehensive benchmark comprising real-world tables from six domains, paired with expert-refined reference report structures and verifiable key insights. Experiments on MMRBench demonstrate that MCTS-Report significantly outperforms strong baselines across structural completeness, numerical accuracy, chart-text alignment, and insight novelty, achieving a 77.9 overall score.

Context. Microservice architectures are widely adopted for building scalable cloud-native systems, enabling independent deployment, fine-grained service composition, and operational elasticity. Problem. Despite growing interest in sustainable software, research on energy efficiency in microservices spans operational, infrastructural, and architectural perspectives, but these are typically addressed in isolation. Existing studies focus on optimisation techniques or measurement approaches, with limited synthesis of how energy efficiency is considered, measured, and addressed at the architectural level. Goal. This study synthesises research on energy-efficient microservices by examining where energy efficiency is considered, how it is measured, and which architectural solutions have been proposed. Method. We conduct a systematic literature review following Kitchenham's guidelines, screening publications from four major digital libraries through a six-stage process with backward and forward snowballing, resulting in 40 primary studies. Results. Energy efficiency is predominantly addressed at runtime through monitoring, scheduling, and resource management, while design-time integration remains limited. Measurement practices are largely infrastructure-oriented and rely on model-based estimation and coarse-grained monitoring. Conclusion. Energy efficiency in microservices is primarily treated as an operational optimisation problem rather than a lifecycle-spanning architectural concern, highlighting the need for earlier architectural integration and improved measurement practices.

How do you verify a long-horizon agent when its own state and self-reports are exactly what you cannot trust? We present an agent instrument built so that verification is structural rather than post-hoc. A deterministic Executive owns all belief; a language model may only file typed proposals, and a claim is admitted only when a prediction pre-registered before acting is matched against observation by code. Two properties make the instrument a verifier of its own science, not just of the agent: every run invalidates itself when per-organ write-error, render-size, or salted-canary-echo floors are breached (four of the first eight architecture runs were invalidated, each localizing a real defect); and a render-invisible shadow reference compiles the plan the full system would have committed in every ablation cell, so drift metrics are defined even where the mechanism under test has been removed. Using this instrument we report a clean, single-variable result on a failure every long-horizon agent suffers: ablating the commitment mechanism flips goal-abandonment from 0.00 to 1.00 while binding error stays flat at 0.00 (three seeds per cell, up to 394 reference beats per run, every run gated valid). The binding channel, by contrast, does not reappear as per-beat drift when its repair is ablated -- because binding is code-owned, the failure class is structurally absorbed, its only residue appearing one layer upstream as a collapse in hypothesis formation. We report these under full disclosure that task efficacy is null (zero level completions across 52 gated runs on ARC-AGI-3), pre-registered as a structural defeater. The contribution is a verification methodology for agent development and the drift decomposition it makes measurable.

Vehicle-to-everything (V2X) systems increasingly incorporate large language models (LLMs) for semantic tasks such as message summarization, operator assistance, and decision support at roadside units and edge nodes. Although these components are not part of safety-critical control loops, they introduce prompt-level attack surfaces that are not addressed by traditional V2X security mechanisms focused on authentication and message integrity. This paper presents Guarded-V2X, an inline semantic guardrail architecture for securing LLM-enabled V2X services under real-time constraints. The proposed system integrates rule-based ingress filtering, a lightweight safety classifier, policy-constrained structured generation, trusted-only retrieval, and post-decision adjudication to enforce machine-checkable safety boundaries prior to downstream execution. Guarded-V2X is evaluated using a four-stage experimental pipeline encompassing intrusion vulnerability analysis, calibration and latency benchmarking, guardrail validation, and robustness under adversarial stress. Experiments are conducted on a V2X-aligned simulated dataset derived from RSU advisories, operator messages, and annotated V2X message summaries. Results show that unguarded and prompt-only baselines retain residual vulnerability under multi-turn adversarial trials, while Guarded-V2X consistently reduces intrusion acceptance success rates and eliminates observed unsafe completions in two-turn settings, without exceeding latency budgets for V2X semantic advisory paths.

Automated Knowledge Base Construction (AKBC) is a core NLP task, and recent work proposes generating knowledge bases directly from large language models (LLMs), treating the model itself as the knowledge source. However, LLMs natively possess no representation of entities, leading to duplicate entries as well as conflations. We propose GPTKB 2.0, a methodology for constructing disambiguated KBs directly from LLMs. GPTKB 2.0 incorporates on-the-fly disambiguation of entities, relations and classes, and is meticulously designed to satisfy both scalability and disambiguation accuracy. We analyze the central design decisions and characterize the trade-offs between accuracy, scale, and cost. We execute GPTKB 2.0 at scale, obtaining a materialized KB containing over 1M disambiguated entities and 38.4M triples. This represents the first million-scale LLM-native KB with explicit internal canonicalization of entities, relations, and classes, a significant departure from prior Wikimedia-centric works. GPTKB 2.0 is available at https://gptkb.org/.

Collective intelligence research treats disagreement as evidence of epistemic diversity: if agents express different views, the group should retain capacity to revise. In LLM collectives this proxy can break: agents can produce diverse-looking arguments while preserving the same conclusion. We operationalize dispersion-revision coupling: the degree to which an intervention that verifiably increases the dispersion of a collective's outputs in embedding space is accompanied by genuine revision of its epistemic stance rather than premise-preserving reformulation. The diagnostic is black-box: it operates on generated text alone and makes no claims about the internal representations of the generating models. Two channels are measured independently: an output channel, the Coherence Index (CI), verifies that the intervention changed output dispersion; an epistemic channel, per-turn stance annotation, measures whether the collective revised. We propose CI with the Meta-Predictive Clarity System (MPCS), which inserts a Re-Differentiation Protocol (RDP) when outputs over-converge, as a reusable method for estimating this coupling regime. We evaluate five-agent collectives from two configurations (gpt-4o-mini and gemini-2.5-flash; 310 paired episodes per condition). On gpt-4o-mini, conditional dissent improves false-premise recovery by +17.7 points (p<1e-6) while static persona diversity harms recovery (-8.1, p=.007). On gemini-2.5-flash, the same intervention at a comparable budget yields no gain (26.1% vs 27.1%, p=.84) despite a verified dispersion drop; the two treatment effects differ from each other (z=3.79, p<.001). Mechanism tagging shows Gemini preserves the false premise via intra-framework dissent: 94% of tagged post-RDP responses reformulate rather than concede (vs 24% on GPT). We recommend reporting per-intervention stance shift and premise-preservation rate alongside accuracy.

Physical AI policies require inference throughout their lifecycle, including model evaluation, cloud reinforcement learning rollout, edge GPU serving, and onboard deployment. Although these settings share the same checkpoint and action semantics, they often rely on separate inference programs. To unify them, we build PhyAI, a Physical AI inference engine with a single runtime that keeps architecture-specific conditioning, solver, cache, and output logic in model adapters while sharing graph execution, kernels, memory management, and parallel services. The same codebase runs vision-language-action (VLA) models and world-action models (WAMs) on single or multiple GPUs across onboard, edge, and cloud deployments. We used the adapter interface to add MiniCPM-Robot on the day of its release. PhyAI achieves 1.40x-4.65x speedups over the official implementations of pi0, pi0.5, GR00T N1.7, and MiniCPM-Robot. On Cosmos3-Nano-Policy-DROID it reduces latency from 2.46 to 1.18 s on eight H20 GPUs (CFG=2, TP=4), a 2.08x speedup. Specialized runtimes remain faster in several configurations, so our goal is one runtime with competitive latency rather than the fastest result in every case. Detailed profiles reveal why different models need different execution policies: on a Hopper-series GPU at batch size one, the pi0.5 action expert accounts for 8.8% of FLOPs but 57.2% of latency; at batch size 32 its share drops to 13.5% and throughput reaches about 100 samples/s. Cosmos3 remains generation-dominated and gains only 14.3% throughput as batch size increases from 1 to 16. We further introduce the control-time Roofline, which distinguishes inference-bound from environment-bound control; the measured pi0.5 points on four LIBERO suites are environment-bound while Cosmos3 stays inference-bound. Code and benchmarks: https://github.com/mingti-org/phyai.

Joint-embedding predictive architectures learn abstract states by predicting target embeddings from context embeddings, but their transition models are typically opaque neural maps. We introduce SJEPA, a reconstruction-free JEPA framework that learns predictive representations whose induced dynamics admit compact symbolic descriptions. Its hybrid transition combines a symbolic law with a regularised neural correction for dynamics outside the selected grammar. The central principle is to learn the simplest adequate dynamics: representation constraints preserve informative, non-collapsed predictive coordinates, while operator compression favours low-complexity symbolic-neural transitions that remain predictively adequate. We formalise this principle through induced-dynamics complexity, analyse predictive-coordinate non-identifiability, and show that unconstrained operator compression creates a direct shortcut to representation collapse. The framework supports both alternating representation-equation learning and symbolic dynamics fitted to fixed representations. In controlled pendulum experiments, joint learning discovers substantially simpler symbolic dynamics with lower long-horizon rollout error and divergence than post-hoc fitting, while an unconstrained one-step diagnostic realises the predicted collapse shortcut. Under grammar misspecification, correction regularisation preserves the representable symbolic mechanism and directs the neural component towards residual dynamics. The results expose a controllable trade-off among predictive fidelity, representation quality, symbolic parsimony, and symbolic-neural allocation.

Test-time reinforcement learning (TTRL) improves the reasoning capabilities of large language models without labeled data by updating the policy with pseudo-labels constructed through majority voting. While effective, the reward signal assigned from majority voting is highly sensitive to consensus strength, defined as the frequency of the most common answer within a rollout group. In TTRL, consensus strength plays a dual role: it reflects both the reliability of the pseudo-label and the distribution of advantages. Low consensus can amplify updates from unreliable pseudo-labels through disproportionately large advantages, whereas high consensus reduces reward contrast and ultimately yields vanishing gradients. In this paper, we introduce Hi-TTRL, a test-time reinforcement learning framework that utilizes hints during sampling to regulate rollout consensus strength. Hi-TTRL first estimates consensus strength from a partial rollout group. When the consensus strength falls outside a target interval, it invokes a Markov chain Monte Carlo (MCMC) hint sampler. The sampler targets the power-transformed prefix distribution and uses finite-step approximate sampling to generate rollout prefixes as hints. By tuning the power exponent, Hi-TTRL generates hints with a sharpened or flattened power target, steering rollout consensus strength toward the target interval. Experiments on multiple datasets and backbones show that Hi-TTRL consistently improves over standard TTRL, with ablations and consensus-steering analyses validating the effectiveness of adaptive hint-guided consensus regulation.

Top-$k$ selection determines which components of a sparse model remain active. Hard selection blocks gradients, while continuous relaxations often couple mask hardness to the selected mass. We introduce LaPrune, a mathematically exact-budget differentiable layer that controls the normalized second moment while preserving the selected mass. A LapSum barrier preserves the selection mass, and a normalized second-moment constraint moves the mask from a dense equal-mass allocation toward hard top-$k$ at each budget. We derive a population prediction of the saturated fraction, a near-binary limiting law, and a tight worst-case guarantee on the near-zero fraction. The normalized hardness parameter is invariant to score scale, while a fixed LapSum temperature is not.

Biomedical literature is increasingly mined for knowledge beyond the questions it was written to answer. Because the target concepts are not known in advance, annotators prefer open-ended labels, whose agreement is hard to quantify. We study soft inter-rater reliability for annotators providing unstructured texts for biomedical annotation tasks. Synthetic experiments show that soft reliability can be quantified using a variety of semantic equivalence measures, and that the choice of measure affects failure modes of the estimation. Embeddings are scalable, but limited when differentiating similar but distinct concepts. Large language models are promising, but limited by scalability for estimating agreement by chance. Finally, we suggest measures based on natural language inference as a sensible compromise.

Constructed languages (conlangs) are intentionally created human languages with a rich tradition of linguistic creativity. Despite their potential for studying language learning in large language models (LLMs), existing conlangs remain largely underexplored in LLM research. We present ConlangBench, the first large-scale benchmark for evaluating and training LLMs on 21 existing conlangs. We collect over 21M conlang-English parallel sentence pairs (including 430K pairs across the 20 non-Esperanto conlangs) and 321K vocabulary entries. In bidirectional translation experiments, we find that models perform better on a posteriori conlangs, whose vocabularies are derived from natural languages, reflecting the design characteristics of conlangs. Training on ConlangBench also shows that models can learn all eight conlangs for which sufficient parallel corpora are available, while their learning curves vary depending on how the conlangs were created. Our findings suggest that conlangs provide a unique testbed for investigating how LLMs acquire low-resource languages.

When people label text for sexism, they often disagree, and not because some of them are wrong: they genuinely perceive sexism differently. Most NLP systems discard this disagreement by collapsing it into a majority vote. We propose the Multi-Agent Perspectivist Preference Optimization (MAP-PO) framework to keep these different perspectives. On the EXIST 2024 dataset of labeled English and Spanish tweets, we first cluster annotators by their labeling behavior rather than their demographic attributes. We then fine-tune one Large Language Model agent per cluster to reproduce that cluster's annotation behavior, and coordinate the agents with preference optimization that combines individual and team-level rewards. We evaluate MAP-PO in four settings defined by two languages and two backbone language models, asking whether each agent reproduces the annotations of its own cluster and whether the agents together reproduce the majority label. Two findings hold in all four settings. First, without fine-tuning the agents behave almost identically, so cluster-specific training is necessary. Second, we show that training each agent only on the labels of its own cluster pushes the agents far beyond the clusters they should represent, while adding a shared team-level training signal consistently keeps each agent calibrated to its cluster.

Retrieval-augmented generation (RAG), which augments large language model (LLM) generation with information retrieved from databases, has become a widely used approach for knowledge-intensive applications. Modern RAG systems, however, expose many configuration choices, such as retrieval indexes, model selections, and how models invoke retrieval. Each configuration yields a different trade-off between answer quality and serving performance, making it challenging to choose the optimal setting for a specific application deployment. We present RAG-Stack, a framework for efficiently discovering quality-performance Pareto frontiers across diverse RAG applications and serving systems. RAG-Stack consists of RAG-PE, an iterative design-space exploration algorithm that selects the next RAG configuration to evaluate; RAG-IR, a workload abstraction for diverse RAG algorithms; and RAG-CM, a performance model that predicts the optimal deployment and serving performance on the given hardware. Together, these components allow RAG-Stack to search the joint algorithm-system configuration space without deploying every candidate and to transfer an existing Pareto frontier to a new serving system. Given the same number of optimization iterations across diverse datasets, the Pareto frontiers found by RAG-Stack cover 52.5% to 153.2% more of the normalized quality-performance space than those found by state-of-the-art configuration-search methods evaluated over the same RAG design space.

The minimum change principle provides an information-theoretic characterization of the Bayes reversal channel in classical probability theory and has recently been proposed as a framework for extending Bayes' rule to quantum information theory. Using quantum relative entropy, we investigate a minimum change principle for the setting of quantum statistical inference. Specifically, we consider a forward process based on a classical-to-quantum preparation channel and a reverse process based on a quantum-to-classical measurement channel. We establish a closed-form characterization of measurements that are optimal for this principle, and this optimal measurement can be found via a dual formulation involving a single unconstrained Hermitian variable. This perspective allows us to recover some notable measurements within the same framework, including pretty good measurements and Fermi-Dirac thermal measurements, and we use it to discover a novel family that we call softmin thermal measurements. We further show that softmin thermal measurements arise as optimal solutions to entropy-regularized semidefinite optimization problems, demonstrating that they play a role for measurements analogous to that of thermal states in statistical mechanics. Finally, we prove an additivity property for the relative-entropy minimum change principle and investigate the performance of Fermi-Dirac thermal measurements for quantum hypothesis testing.

Reinforcement learning with verifiable rewards (RLVR) com- monly optimizes each correct completion as an independent learning signal. In GRPO, this completion-level uniformity creates structure-level skew: recurring correct solution forms accumulate positive coefficient mass in proportion to how often they are sampled, while rare forms receive limited credit. We formalize this behavior as multiplicity-induced structure-level credit concentration and introduce a partition- conditioned rule that redistributes positive advantages accord- ing to cluster rarity. Cue-GRPO instantiates this rule with- out auxiliary-model inference by using deterministic Strategy Cues to construct rollout-local partitions of verified-correct traces. Across Qwen2.5-Math-7B and Llama-3.1-8B-Instruct, Cue-GRPO improves AIME repeated-sampling performance, with the largest gains at high sampling budgets. Credit Re- distribution (CR) under Judge Partitions (JP) further indi- cates that the proposed redistribution mechanism can oper- ate with judge-derived partitions. Cue-GRPO adds only 6% wall-clock training overhead over GRPO. These results sup- port structure-level credit redistribution as a practical design axis for RLVR, with Strategy Cues providing a low-overhead implementation for competition mathematics. Code is avail- able at https://github.com/CzZ12/When-Correct-Solutions- Repeat-Rarity-Aware-Credit-Redistribution-for-GRPO.

Multimodal intent recognition requires understanding not only what textual, acoustic, and visual signals share, but also how they disagree. Such disagreement is frequently class-informative; for example, lexical positivity accompanied by incongruent vocal or facial behavior may indicate sarcasm or taunting, yet most fusion methods either encourage modality alignment or treat inconsistency as uncertainty to be suppressed. We propose MACH (Modality Agreement- and Conflict-aware prototype Hypergraph), a hierarchical prototype-hypergraph framework that represents multimodal agreement and conflict as distinct, recurring relational structures. MACH progressively composes unimodal representations into bimodal and trimodal abstractions. At each applicable level, modality-composition anchors activate sparse agreement prototype hypergraphs that capture reusable consensus patterns, while a separate conflict pathway maps cross-modal discrepancies to dedicated conflict prototype hypergraphs. The two pathways are combined through a feature-wise, sample-adaptive arbitration mechanism, enabling the model to preserve informative disagreement while suppressing incidental modality noise. A progressive optimization strategy stabilizes the interdependent hierarchy before joint agreement-conflict learning. Experiments on benchmark datasets demonstrate the effectiveness of the proposed formulation, while component and robustness analyses validate the distinct roles of hierarchical composition, prototype-mediated semantic refinement, and agreement-conflict arbitration.

Prompt injection remains a critical threat to LLM agents, yet existing defenses treat each task as a self-contained problem, independent of previous encounters. In practice, user requests are often underspecified: they describe the desired outcome without fully specifying acceptable behavior. An injection can exploit this ambiguity, causing the agent to complete the task in a way the user would reject. As the user's expectations become clearer through concrete cases, a defense should learn from each encounter and apply what it learns to the next. Inspired by adaptive immunity, we propose AgentAntibody, which equips LLM agents with a self-evolving immune system against prompt injection. AgentAntibody represents its evolving understanding of the user's security boundary as a persistent library of antibodies. At runtime, the library recognizes threats to this boundary and mounts corresponding immune responses. Across encounters, it evolves to strengthen the agent's immunity to future attacks. Extensive experiments across three benchmarks and four backbone LLMs show that, by learning the user's boundary through experience, AgentAntibody outperforms existing defenses in preventing harmful actions while preserving legitimate task completion, even when the harmful and legitimate actions are both compatible with the stated task.