Papers

Modern AI model training imposes unprecedented computational demands, making it a key contributor to datacenter energy consumption. Yet a significant fraction of the energy consumed during training does not translate to useful computation due to bottlenecks throughout the training pipeline. We present PAFT, a phase-aware, dynamically adaptable GPU frequency tuning system that reduces energy consumption of training workloads with minimal performance overhead. The key insight behind PAFT is that bottlenecks represent an energy optimization opportunity, rather than purely a performance problem: when GPUs are bound to stall, PAFT opportunistically reduces their clock frequencies to match the pace of bottlenecked devices, saving energy without impacting execution time. PAFT achieves this by continuously monitoring pipeline behavior and applying fine-grained frequency adjustments, adapting to workload and system changes. Experiments conducted on twelve widely used models show that PAFT consistently outperforms all baselines, achieving energy savings of up to 46% with an average overhead of 4%.

Visual segmentation systems encounter objects outside their training distribution during real-world deployment, hindering reliable autonomous systems that depend on scene parsing in the perception stage. Many recent methods address this by using self-supervised foundation models to train density estimators that yield low likelihood in anomalous image regions. Although promising, these methods suffer from poor feature semantics or they lack spatial consistency, both of which undermine critical downstream decisions. We address this problem with~\method, a generative method based on class-conditional density estimation over self-supervised representations. SAFe trains lightweight normalizing flows that produce class-conditional normalized likelihood estimates over frozen DINOv3 features. We combine density estimates from transformer features with density scores over multi-scale convolutional features to capture both global semantics and local detail. We introduce a method-agnostic post-processing step based on SAM3 that connects per-location likelihoods into spatially coherent segments while suppressing false positives, and enables instance-level anomaly detection without retraining. The post processing further distinguishes novel categories among anomalous objects by a similarity-based agglomerative clustering scheme. SAFe sets a new state of the art on the PANIC, OoDIS, SMIYC ObstacleTrack with strong performance on the ISSU benchmark.

Sparse attention reduces the quadratic cost of global self-attention while retaining strong empirical performance, but how its restricted interactions shape the evolution of token representations remains theoretically underexplored. Modeling tokens as particles on the unit sphere, we introduce opinion leader dynamics, a framework that identifies two mechanisms through which token groups converge internally while maintaining distinct limiting directions. In the explicit model, fixed representatives induce a potential that attracts tokens toward distinct local maxima. In the implicit model, disconnected interaction groups evolve toward separate consensus directions. We formulate both models as reverse Wasserstein gradient flows and establish exponential convergence under suitable conditions. We further connect these theoretical predictions to token evolution in frontier sparse-attention LLMs that motivate our framework. Across four benchmarks, Kimi-K3, MiniMax-M3, and DeepSeek-V4-Flash consistently exhibit clearer cluster separation and higher clustering scores than the dense-attention model GLM-4.7-Flash in projected token representations. These observations support the relevance of the predicted multiple-group structure to trained frontier LLMs, while finite-particle simulations illustrate the theoretical convergence behavior. Together, our results connect restricted token interactions to distinct group-level attractors, providing a dynamical account of how sparse attention can support alignment within groups while preserving separation between them.

In this work, we propose a novel design of high-rate Calderbank–Shor–Steane (CSS) quantum low-density parity-check (QLDPC) codes based on quasi-dyadic matrices. The Tanner graphs associated with the two component classical codes used to construct the proposed CSS codes are characterized by a girth of at least $6$. We also derive an exact count of length-$4$ cycles in the quaternary representation of the parity-check matrix. Monte Carlo simulations under code-capacity and phenomenological noise models show that the proposed codes have competitive finite-length logical error rate performance compared with state-of-the-art codes in both these settings.

AI is increasingly used to automate computer security testing, and the tools must decide for themselves whether an attack succeeded. A finding that a deterministic rule confirms by observation is reported as fact, whereas one that an LLM judges exploitable is treated as an opinion. We ask whether the system under test can forge that confirmation. In offline security testing of a four-stage AI-assisted pipeline, nine of its fifteen confirmation mechanisms are forgeable, and forgeability is predicted entirely by whether the decision reads attacker-controlled data. We formalise this as an auditable attack surface and test it prospectively: on sixteen held-out mechanisms, predictions fixed before any attack separated forgeable from unforgeable mechanisms exactly, and across 12,203 mechanisms in public scanner templates the prediction was 99.9% accurate. Deterministic rules proved cheaper to forge than eight open-weight LLM judges, failing at 2% of attacker-controlled response content against a median of 50%. No implementation of one check was both robust and precise, and routing between a rule and an AI judge raised forgery to 99%. Moving the decisive evidence to a channel the attacker cannot write cuts attack success from 97% to 0%, and an escalate verdict recovers the sensitivity this costs. The protection fails when the scanned host is itself the adversary. The results bear on AI security agents and on benchmarks that score success by string matching.

Evaluation benchmarks for Indian language automatic speech recognition (ASR) suffer from two systematic biases: optimistic scores from clean, controlled audio conditions, and pessimistic scores from overly rigid transcription standards that penalize valid linguistic variations. We introduce Vimarsha, a 100-hour benchmark spanning all 22 scheduled Indian languages, designed to address both distortions. Vimarsha combines demographically diverse on-field recordings with carefully mined in-the-wild audio selected for acoustic difficulty, alongside a lattice of variations framework that encodes multiple valid transcriptions per utterance. Evaluations of 10 state-of-the-art ASR models reveal substantial shifts in model rankings under realistic conditions, geographic and demographic performance disparities, and systematic failure modes across speaking rates and acoustic environments.

This paper describes the SKstars submission to SHROOM-Visions 2026, a shared task on fine-grained hallucination detection in large vision-language model outputs. The task requires systems to identify hallucinated character spans, assign hallucination categories, and provide confidence estimates for their predictions. Our approach combines zero-shot predictions from Qwen2.5-VL-72B-Instruct with those of a LoRA-adapted Qwen2.5-VL-7B-Instruct model. The outputs of the two models are integrated through a lightweight ensemble procedure, followed by span refinement and confidence adjustment. We evaluate the main system components on a small internal development subset and report the performance of the submitted system on the official English test set. SKstars achieved a Cor+Lbl score of 0.2902, ranking 15th among 29 teams, and obtained Cor and IoU scores of 0.3642 and 0.3151, respectively, ranking 18th on both metrics. The results show that combining a large zero-shot model with a smaller adapted model provides a practical framework for multilingual and fine-grained hallucination localization, while also highlighting the difficulty of transferring development-set improvements to hidden test data. Code and predictions: https://github.com/aliathar1401/SK-Stars-shroom-visions-2026

Speculative decoding losslessly accelerates large language model inference by having a lightweight draft model predict future tokens for verification by the target model. Recent block diffusion drafters further reduce drafting latency by predicting multiple tokens in parallel. However, existing block drafters project target hidden states at every input position into a separate drafter-side KV cache, incurring per-request memory and KV-write overhead that grow with concurrency; directly reusing target KVs in place removes this cache but fails to sustain draft quality throughout the block. We propose a hybrid target-context injection method that complements direct target KV reuse with target hidden states only at the last input position, requiring no separate drafter-side KV cache. Building on this design, we propose H-Spec, a hybrid Mamba-attention parallel drafter that consumes the two target-context sources through complementary modules. Mamba modules are initialized with projected last-token target hidden states, while attention modules reuse target KVs in place. Despite its recurrent formulation, Mamba's parallel scan allows H-Spec to preserve block-parallel drafting. Across three target models and diverse tasks, H-Spec improves over the best baseline by 5.0–13.3% in mean accepted length and 5.3–12.6% in batch-size-1 inter-token latency speedup. Under concurrent serving, H-Spec consistently achieves higher throughput while maintaining lower KV cache utilization than baselines across evaluated concurrency levels.

Looped Language Models (LoopLMs) perform "latent reasoning" by recursively refining internal latent representations with shared weights, offering a more effective alternative to explicit verbal reasoning. Despite their effectiveness, we find that LoopLMs remain prone to loop instability: unstable refinement across iterations can produce localized uncertain "hard" tokens associated with reasoning errors. To address this, we propose LoopCD, loop-wise contrastive decoding that enhances the reasoning performance of LoopLMs by intervening on these tokens at inference time. Specifically, we exploit the internal dynamics of LoopLMs and contrast the logits from earlier iterations with logits from the last refined iteration to form the final sampling distribution. We find that this strategy is highly efficient, introducing only negligible inference overhead and requiring no additional training, while effectively improving reasoning performance by naturally refining reasoning-critical hard tokens. Extensive experiments show that our method improves the performance of recent representative LoopLMs across various reasoning tasks.

Reliable perception is essential for underwater vehicles operating in complex environments, where light attenuation and scattering often degrade visibility and compromise optical sensing. Forward-looking sonar (FLS) offers an alternative by providing high-frame-rate acoustic imaging under poor optical conditions. However, real-time FLS mapping remains challenging due to unresolved target elevation, spatially non-uniform noise, and fragmented target boundaries, which hinder feature extraction and introduce geometric ambiguity during projection. To address these challenges, we propose a cascaded feature reconstruction pipeline combining fast Fourier transform (FFT)-based denoising, fast multiscale constant false alarm rate (MCFAR) detection, and gradient-adaptive boundary connection to extract geometric features from degraded sonar images with low latency. We integrate attitude-aware geometric projection with incremental occupancy accumulation to construct a depth-referenced 2.5D map for local mapping in confined underwater environments. The sonar's vertical position is referenced to an external sensor, while target elevation is assigned under an explicit geometric assumption rather than measured directly by FLS. Experiments in a 3 m X 5 m pool demonstrate centimeter-scale planar mapping accuracy, with a root-mean-square error (RMSE) below 3 cm across three sequences and an average processing time of 42.4 ms per frame.

Evaluation scores used around LLM systems – including reward models, rerankers, and LLM judges – can track surface form instead of the quality they claim to measure. When presented with a terse correct solution and a commented buggy solution for the same MBPP problem, a public preference reward model selects the correct one no better than a coin flip (0.507). Subtracting the predictable surface component from such scores is increasingly common, but removal alone does not yield a more valid measurement: the removed component may carry construct-relevant signal, and residualization cannot tell which is which. Under designed interventions – unit-test labels with comment-only edits – residualization attenuates the reward model's format effects by about 0.12 on both correct and buggy code, while the correct-versus-buggy margins move by less than 0.01. In observational NLI and QA settings, we freeze a held-out replication before scoring and re-evaluate it using labels from disjoint annotators; this supports only a narrower conclusion: better agreement with the construct labels on a pre-declared slice where a surface-only predictor errs, not a repaired score. Full-population agreement falls in every observational setting with a reported positive slice gain, and within-question ranking falls in every such QA setting. When construct and surface features are entangled, residualization can decorrelate a score while degrading construct alignment, and, in a controlled model, configurations just as damaging to construct alignment pass every pre-adjustment check, so no committed gate is a guarantee. We assemble these distinctions into a reporting protocol whose outcomes, refusal included, state what an adjusted score may be claimed to show: an audit-time diagnostic reported beside the construct-alignment cost it incurs, never a replacement for the raw score.

Pedestrian head orientation recognition plays an important role in autonomous driving by providing valuable cues for understanding pedestrian attention and anticipating potential crossing behavior. However, reliable recognition in real-world traffic scenes remains challenging because pedestrian head regions are often captured at low resolution. To address this challenge, we propose a lightweight Low-Resolution Head Orientation Convolutional Neural Network (LRHO-CNN) for pedestrian head orientation recognition. We construct a new dataset by extracting pedestrian head images from multiple public datasets and manually annotating them into eight orientation categories. The collected images are systematically preprocessed and augmented to increase data diversity and better represent variations in illumination and image quality. The experimental analysis compares LRHO-CNN with three fine-tuned baseline models, namely ResNet-18, ResNet-34, and VGG-16. The results demonstrate that LRHO-CNN achieves the highest classification accuracy among the evaluated models. LRHO-CNN is further evaluated on the JAAD and PIE datasets, demonstrating its effectiveness in recognizing pedestrian head orientation in real-world traffic scenes and providing informative head-orientation cues that can support downstream pedestrian behavior and intention prediction.

Many spatial (3D) vehicles, including AUVs and fixed-wing aircraft, are effectively nonholonomic and subject to limited actuation, such that continuous time-invariant stabilization is fundamentally obstructed by Brockett's necessary condition. To overcome this obstruction, we exploit the geometric singularity of spherical coordinates to design a backstepping continuous, time-invariant feedback law that exponentially stabilizes a 3D nonholonomic vehicle actuated solely by forward surge velocity, pitch rate, and yaw rate. The resulting closed-loop region of attraction excludes only the coordinate singularity, codimension-two, measure-zero set of initial conditions in which the vehicle lies on the line through the target orthogonal to the target plane, thereby covering the largest possible domain. We further construct a strict control Lyapunov function to prove global exponential stability of the origin on this domain with a user-specified decay rate, while simultaneously preventing the system from approaching the singular set. Finally, we show that the closed-loop system is exponentially attractive to the origin in Cartesian coordinates. Numerical simulation examples in both spherical and Cartesian coordinates illustrate the effectiveness of the control law.

Artificial intelligence (AI) has advanced at a rapid pace in recent years. Initially, breakthroughs in large language models caught widespread attention. However, recent generations of frontier AI models have adopted multimodal capabilities as a first class citizen, with vision capabilities being central to that. In this paper, we evaluate three recently released models on the task of diagnosing dermatological conditions from patient-submitted images. The models chosen are at the low to mid tier in terms of pricing and thus represent a floor on current AI capabilities, not a ceiling. We evaluate AI performance relative to a panel of three certified dermatologists, who grade each image, and we present four interesting findings. Firstly, depending on the metric, the tested AI models are either on par or slightly trail humans in terms of inter-clinician agreement. Secondly, we find that asking AI models for a confidence rating produces poorly calibrated answers, meaning use of confidence thresholds should not be relied upon in a clinical setting. Thirdly, the effect of providing additional patient metadata is strongly model-specific, with one of the three models degrading on every metric considered. Finally, model cost is not predictive of performance. The best-performing model we tested costs on average $0.0045 per case.

Vision-language navigation requires embodied agents to navigate environments using natural language instructions and visual observations. Existing approaches typically decompose navigation into sequential language-guided decisions or rely on online exploration without prior environmental knowledge. Scene graph representations offer compact semantic memory but remain decoupled from downstream navigation, which still depends on dense metric maps. To close this gap, we propose an object–path graph that unifies open-vocabulary semantic reasoning with topological navigation. The proposed representation jointly supports semantic grounding, graph-based localization, and navigation within a single lightweight topological framework. Building on this graph, we introduce a navigation strategy that combines global path planning with local inter-node execution through lightweight node localization and semantic visual servoing, enabling navigation directly over the graph without dense metric reconstruction. Experiments on HM3D and Replica demonstrate competitive performance in open-vocabulary object grounding through the proposed hierarchical graph structure, while achieving effective navigation performance. Real-world robot experiments further validate the practicality of the proposed framework.

We formalize Hastad's PARITY lower bound in Lean using the switching lemma. For every fixed d >= 2, formulas and DAG circuits of computation depth at most d computing PARITY on n inputs require size exp(Omega_d(n^(1/(d-1)))) for all sufficiently large n. This matches the classical upper bound up to constants in the exponent and implies that PARITY is not in nonuniform AC0. We also construct a polynomial-size, logarithmic-depth bounded-fan-in formula family for PARITY, providing a witness to NC1 is not a subset of AC0 for the formalized models. The Lean source code is available at https://github.com/formalcs/circuit-complexity and is checked with Lean 4.33.1 and mathlib 4.33.1.

Autonomous endoscopic navigation requires the policy model to predict actions from texture-poor monocular observations, make safe control decisions, and retain evidence of lesions after they leave the field of view. Existing vision-language-action (VLA) models primarily rely on visual appearance and short-term context, limiting geometric grounding and episode-level reporting. We introduce StenoVLA-3D, a 3D-aware VLA framework for navigating through stenotic regions. We integrate point-maps into the Cosmos-Reason 2 backbone through learned geometry-gated fusion, and also propose a temporal state branch to model traversal progress. Our reasoning-and-action backbone predicts grounded reasoning with actions, while dedicated heads estimate stenosis shape and generate the final lesion report. We further introduce EndoCausal, an episode-level dataset with lesion annotations, actions, and temporally grounded reasoning. On 40 held-out recorded test episodes, StenoVLA-3D reaches 95.2% semantic accuracy and 83.4% action accuracy. On the physical 3-DoF endoscope, it attains 88.9% and 77.8% task success in esophageal and colonic phantoms (36 trials each), substantially outperforming the evaluated baselines.

Large language models have achieved remarkable progress on Text-to-SQL through reasoning-enhanced fine-tuning, yet existing approaches predominantly rely on massive instruction corpora under the assumption that scale drives performance. We challenge this paradigm by investigating a fundamental question: what is the minimal data requirement for effective Text-to-SQL instruction tuning? We propose LIMIT(Less Is More for Instruction Tuning in Text-to-SQL), a data-centric framework that demonstrates strong database reasoning can emerge from an extremely compact training set when examples are strategically selected. LIMIT operates through four stages: difficulty-aware filtering that identifies samples within the model's learning frontier, chain-of-thought synthesis with consistency-based selection, multi-dimensional quality scoring via LLM-as-judge, and genetic algorithm optimization that jointly maximizes schema coverage and sample quality. On the BIRD and Spider benchmark, LIMIT selects only 796 and 863 samples while achieving 100% table coverage, enabling Qwen3-8B to reach 69.1% and 88.9% execution accuracy.This result surpasses methods trained on 20 times more data and establishes a new state-of-the-art among open-source approaches. Our findings suggest that careful data curation, rather than scale, is the key to efficient Text-to-SQL learning.

We prove the Courtade-Kumar conjecture: among all Boolean functions $f\colon \{-1,1\}^n\to\{-1,1\}$, a dictator retains the most information about a uniformly random input observed through independent binary noise.

Visual grasp proposal generation has advanced rapidly, yet converting a selected proposal into a stable physical grasp remains a central execution-stage challenge. This paper introduces GraspTune, a tactile-driven execution-stage refinement framework that starts from a nominal proposal and applies bounded residual TCP motions during approach, contact formation, and final grasp execution. GraspTune learns control-facing contact semantics from local depth, tactile signals, state, and history using state-conditioned expert contact queries and multi-task supervision for contact change, contact risk, and post-close readiness. The representation conditions a diffusion-pretrained residual policy and is aligned with PPO for closed-loop execution. Across more than 60,000 simulated executions over 20 object categories, GraspTune establishes an execution-layer benefit across four proposal generators, raising stable grasp success by +19.22, +9.55, +12.45, and +20.70 percentage points for GraspNet, Contact-GraspNet, AnyGrasp, and VGN. A four-fold held-out category study raises unseen-object execution from 54.58% to 70.33%, showing category-disjoint generalization of contact correction. Across more than 1,000 real-robot trials on a UR5e setup with Xense fingertip sensors, GraspTune raises GraspNet execution from 71.0% to 84.3%, validating direct transfer without realworld policy fine-tuning. Together, these results turn visually plausible proposals into stable physical grasps for downstream contact-rich manipulation. A supplementary video is available at https://youtu.be/kcq7fSLNtzU.

In this paper, we study remote state estimation under unreliable forward and feedback communication between a smart sensor and a remote estimator. When the sensor cannot perfectly reconstruct the filtering state maintained by the remote estimator, an information asymmetry arises between the two sides of the network. To analyze this asymmetry, we first characterize the internal state maintained by the remote estimator and its evolution under the packet reception process. The sensor's uncertainty about this state is then captured through a belief conditioned on its information set, and we derive a recursive update under noisy acknowledgment feedback. Finally, we show that the resulting belief admits a finite Gaussian mixture representation whose number of components grows at most linearly over time, ensuring computational tractability. Simulation results illustrate the impact of information asymmetry on the estimation performance and demonstrate the effectiveness of the proposed framework.

Legal information in Bangladesh is inaccessible to most citizens. Statutory text is English-only, trained lawyers are concentrated in urban centres, and cloud-dependent AI fails where mobile connectivity is unreliable, a setting in which hallucinated legal text causes direct harm. The system addresses statutory interpretation only; queries that require judicial precedent or case-law reasoning fall outside its scope. We target the statutory access gap by compressing a 9-billion-parameter Gemma-2 teacher into a 2-billion-parameter student through two-phase progressive knowledge distillation. Phase 1 performs supervised fine-tuning on 9,429 quality-gated legal question-answer pairs (65% acceptance from 14,514 generated queries); Phase 2 minimises sparse Kullback-Leibler divergence against the teacher's top-50 per-token logits at temperature tau = 4.0, implemented via QLoRA (4-bit NF4, rank-32 LoRA adapters). Prior legal language models target general legal English; this system specialises in Bangladeshi statutory law. Every response is grounded through hybrid retrieval combining dense semantic search (60%) and BM25 (40%) across 36,029 statutory passages from the Bangladesh Constitution and national legislation. On a 50-query English benchmark, the distilled model reaches ROUGE-L 0.4715 and BERTScore F1 0.5679, a 103% ROUGE-L and 143% BERTScore gain over the retrieval-augmented undistilled baseline (ROUGE-L 0.2323, BERTScore 0.2340). The adapter quantises to 1.6 GB (GGUF Q4_K_M) and runs at 4-8 tokens per second on a Pixel 6 with no network access. Cross-lingual evaluation on 50 Bangla queries yields ROUGE-L 0.4083 and BERTScore 0.8133, showing effective retrieval from Bangla input against an English-only corpus. In a single-evaluator pilot, a practising lawyer rated 50 responses at a weighted mean of 4.16/5 (90% rated 4 or 5), supporting utility beyond text-overlap metrics.

Whether a public blockchain is "decentralized" is routinely settled by citing a concentration statistic. On two ERC-20 ledgers, Chainlink (LINK) and Uniswap (UNI) over a 90-day window, we show that verdict depends on measurement specification rather than the ledger itself. Four discretionary choices (holder population, address type, temporal aggregation, and entity resolution) move the balance HHI for UNI from 109 to 2,336 (a factor of 21), with every specification defensible. Over the same range, the Gini coefficient moves by less than 0.003 and does not change under entity resolution, demonstrating that Gini and HHI answer different questions and cannot substitute for one another. We further document an implementation choice - summing versus overwriting repeated transfers - that discards roughly 85% of volume and overturns a finding on wealth and structural position. Substantively, both ledgers are extraordinarily unequal in ownership (balance Gini = 0.990 and 0.998) yet unconcentrated in routing (weekly flow HHI = 421 and 386), with the two dimensions close to statistically independent across addresses. A parameterized criterion for hidden brokers identifies 30 and 18 zero-balance intermediaries, 12 shared across ledgers; a matched control confirms that degree thresholding, rather than learned embeddings, drives the discovery. Finally, on the governance ledger, proposal-eligible addresses and routing intermediaries are almost disjoint, so routing contestability is held at the pleasure of a rule layer with a Nakamoto coefficient of two. We conclude that on-chain concentration should be reported as a specified range rather than a point estimate.

While Large Language Models (LLMs) excel at understanding unstructured financial contexts, their direct use in portfolio optimization is limited by a mismatch between next-token prediction and the listwise ranking objectives required for asset allocation. They also struggle with precise numerical forecasting, leading to instability and arithmetic hallucinations. To bridge this gap, we propose FinRankGRPO, a framework that shifts LLM based portfolio construction from direct numerical prediction to listwise ranking of financial assets. We introduce a two-stage training process, supervised finetuning on Chain-of-Thought reasoning data, followed by our Financial Asset Ranking via Group Relative Policy Optimization with a Spearman rank correlation reward that aligns generated asset rankings with ground truth market orderings. The second stage uses a novel Spearman rank correlation reward to explicitly align the model's generative preferences with ground truth market orderings. Experimental results show that FinRankGRPO outperforms traditional quantitative and state-of-the-art commercial models, achieving a Sharpe ratio of 0.636 and a Spearman correlation of 0.023.

Long-horizon language agents receive sparse terminal feedback, while intermediate rubrics provide structured but potentially misspecified assessments of progress. In resettable training environments, counterfactual continuation rollouts can measure local credit, but exhaustive replay is costly. We propose Credo, a framework that couples evolving semantic rubrics with selective, execution-based credit correction. A frozen judge maps visible transitions to rubric features, and a credit head predicts the change in expected terminal reward associated with the realized transition. Independently sampled two-sided replays correct prediction residuals using their recorded inclusion probabilities. We derive conditional unbiasedness and a variance decomposition that connects two design choices: which rubric features to retain, and where to allocate a fixed expected replay budget. The resulting criterion weights prediction errors by policy-score sensitivity and missing replay coverage; its allocation rule additionally accounts for continuation cost. We also describe a practical mixture with terminal leave-one-out advantages and distinguish its clipped, token-normalized PPO implementation from the ideal policy-gradient estimator. This preliminary report provides the method, proofs, an exact finite-model audit, and a controlled evaluation protocol. It makes no claim of empirical superiority on language-agent benchmarks.

The growing capability of Wi-Fi devices to identify human activities using channel state information (CSI) raises privacy concerns. To counter this threat, we propose GRAW, an adversary system, acting as a privacy defender, that degrades the human activity recognition (HAR) system at the user device by perturbing the router's signals that the device uses to estimate CSI. GRAW employs generative adversarial imitation learning (GAIL) to construct perturbation signals, and thereby eliminates the need for any information on the target HAR systems and their inputs (i.e., zero-knowledge operation). We evaluate GRAW against seven representative HAR models, using datasets collected in five environments, including our own dataset. We observe that GRAW is the only remote attack scheme that degrades every tested HAR model to a random-selection level. At the same perturbation level, GRAW achieves an attack success ratio up to 76.7% higher than comparison methods, while maintaining over 99% packet success rate on regular Wi-Fi communication. We demonstrate the feasibility of GRAW through real-time, over-the-air experiments with software-defined radios.

Legends are fundamental to chart understanding, as reliable interpretation requires correctly binding legend entries to corresponding visual marks. While vision-language models (VLMs) are increasingly applied to chart understanding, their legend understanding is poorly diagnosed by aggregate accuracy, which can be satisfied by superficial shortcuts and confound legend-specific errors with other reasoning failures. To enable fine-grained diagnosis and controlled testing, we introduce LegendBench, a parametric benchmark and generation pipeline that produces targeted legend-centric test cases. LegendBench contributes (1) a capability-task taxonomy spanning legend parsing, legend grounding, legend-conditioned reasoning, and legend-aware abstention to localize failures, and (2) counterfactual group generation, where each base chart yields multiple variants under controlled legend interventions to probe model invariance and sensitivity. Using LegendBench, we evaluate both general-purpose VLMs and specialized chart models and generate their capability profiles, revealing persistent bottlenecks in reliable legend-to-mark binding and counterfactual consistency. We then use these capability profiles to guide targeted fine-tuning, demonstrating that bottleneck-specific interventions can effectively close the localized capability gaps and generalize to unseen data. We further leverage our counterfactual design to conduct fine-grained diagnostic experiments, analyzing encoding-channel effects, legend-order shortcuts, and abstention under varying visibility.

Embodied AI systems are often organized into System 1 and System 2. System 1 is typically a pretrained policy that generates actions at high frequency, whereas System 2 is often instantiated as a vision-enabled language model for high-level planning. We ask whether a large language model (LLM) can act as the policy for robot manipulation without task-specific finetuning. We call this setting LLM as policy. We evaluate three LLMs on all 42 RoboDojo tasks and compare their scores with 40 public policies. Astra and GPT-5.5 use the official 50-episode-per-task protocol; DeepSeek-Flash uses 10 episodes per task. GPT-6 Astra achieves 22.48% average success rate and 28.97 Score over 2,100 trials, ranking above every public entry. Yet GPT-5.5 and DeepSeek-Flash reach only 0.88% and 1.92% average success rate with the same post-processing. We find that Astra exhibits a sharply polarized capability profile. It generalizes well to tasks that require semantic understanding but not high-precision control. In contrast, it performs poorly on tasks that require precision, dynamic control, or complex bimanual coordination. In-context experiments show no aggregate benefit from one-shot demonstrations, while selected interaction traces show within-episode corrections under perturbations. Overall, the evaluated LLMs vary substantially in manipulation performance. Astra stands out and provides initial evidence for the potential of a general-purpose manipulation model, although reliable precision and dynamic control remain limitations in the evaluated setting.

LLM agents can make expert scientific workflows accessible through natural language, but a plausible response does not establish that the requested computation ran. This matters at synchrotron facilities, where calibration and integration are multi-step operations over terabyte-scale datasets. We present APEXA, a deployed framework coordinating 61 tools for synchrotron data reduction. A deterministic execution-integrity guard prevents it from returning results for tool calls that did not execute, and the same tool layer enforces motor safety: across 50 adversarial scenarios and four models, it produced 0/200 violations against a simulated EPICS controller, versus 15/200 for a prompt-only baseline. We also introduce APEXA-Bench, 58 facility tasks organized by scientific correctness and physical consequence. On real APS beamline data, one prompt recovered detector geometry and integrated a full attenuation and exposure sweep. Trustworthy scientific agents need deterministic validation of execution, not model-side assurances. We release the framework, benchmark, and traces.

Dependency scanning tools surface hundreds of vulnerabilities but provide no exploitation-aware ordering, leaving practitioners to decide which upgrades to perform first with no principled guidance. The dominant practice, ordering by CVSS severity, is structurally misaligned with active exploitation: in our npm corpus, 186 non-KEV vulnerabilities carry CVSS scores greater than 8, all outranking three CISA Known Exploited Vulnerability (KEV)-listed packages and causing CVSS-first tools to defer the first actively exploited fix by 17 upgrade actions. KEVGraph is an eight-stage pipeline that frames remediation as a KEV-aware set-cover problem: it constructs per-repository dependency graphs from lockfiles, joins them against OSV and the CISA KEV catalogue, and produces a minimum-cardinality upgrade plan ordered to eliminate actively exploited vulnerabilities as early as possible via exact Integer Linear Programming (ILP) or a KEV-aware greedy algorithm. Evaluated on 924 real-world npm repositories (1,046 vulnerabilities, 5 KEV-listed), the ILP planner achieves AUCCKEV = 0.997 versus a random-baseline mean of 0.663 (95 percent CI [0.519, 0.831], n = 30), resolves the first KEV vulnerability at plan step 1, and requires only 417 upgrade actions, 15.9 percent fewer than the random mean of 495.4. CVSS-first and Dependabot-style ordering are strictly dominated: they defer the first KEV fix to step 18 while requiring more actions (419 and 421, respectively). The framework generalises: Maven (1,200 repos) achieves AUCCKEV = 0.988 versus random mean 0.486; PyPI (300 repos) achieves AUCCKEV = 1.000. Each plan is accompanied by a machine-verifiable certificate enabling automated compliance verification under CISA BOD 22-01.