Artificial Intelligence
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
Monday 17 August
Long-horizon agents are beginning to automate complete workflows that produce code, reports, and research artifacts. Medical imaging workflows are multi-stage and data-sensitive, while expert trajectories remain scarce and difficult to share. Structured benchmarks can localize failures through stage-level rubrics, but standard post-training discards these diagnostics before the next training round. We present Benchmark-as-Teacher (BaT), a recursive self-improvement system for agent post-training. BaT contains two linked components: the asynchronous Stage Bank data pipeline and BiCuRL (Bilevel Curriculum Reinforcement Learning), its self-improving post-training method. Stage Bank synthesizes content-isolated training states outside the policy-update loop. BiCuRL uses a fixed held-out evaluation to select the next stage curriculum, verifies rollouts with task rubrics, updates the policy with GRPO, and returns the candidate checkpoint to evaluation. On AutoMedBench-Lite, BaT-4B and BaT-9B more than double the Overall scores of their Qwen Instruct baselines. BaT-9B Agent reaches 79.6 Overall, exceeding Claude Opus 4.6 with Claude Code at 77.5.
Machine Learning
Conditional Evaluation of Language Models with Cheap Auxiliary Signals
Aggregate accuracy hides where models succeed and fail. Estimating conditional performance profiles from gold labels alone is expensive, while cheap auxiliary signals such as LLM-judge scores, pairwise comparisons, confidence scores, and judge-disagreement features can be collected for every benchmark item but are often biased or miscalibrated. We propose LACE (Local Augmented Control-Variate Evaluation), a semi-supervised estimator for conditional LLM evaluation. The key step is local centering: after subtracting the conditional mean of a cheap signal within the target profile region, any linear augmentation has zero conditional mean and therefore cannot change the estimand. The augmentation coefficient is used only for efficiency, and a local ridge control variate combines a gold-label residual mean from the labeled subset with a cheap-signal mean from the full item pool. We prove calibration-free identification, unbiasedness for grouped profiles, local oracle optimality within centered linear augmentations, and first-order adaptivity to the estimated coefficient. The resulting gain formula is governed by a population local $R^2$, which characterizes how the efficiency attainable from the cheap signals varies across profile values. We also derive corresponding estimators for direct paired model gaps and deployment-weighted scores. We empirically evaluate the primary performance-profile estimator on MATH-500, ScienceQA, MMLU, WinoGrande, HellaSwag, TruthfulQA, GSM8K, and ARC.
Read the full abstract
Numerical Analysis
An FFT-Accelerated Boundary Integral Equation Method for Wave Scattering by Smooth Surfaces in Three Dimensions
For wave scattering by axisymmetric surfaces, the fast Fourier transform (FFT) method provides an effective tool to accelerate standard boundary integral equation (BIE) solvers. Surface BIEs can be decoupled into a series of curve integral equations on the generating curve, due to the convolution-like integral operators. The Fourier coefficients of the three-dimensional fundamental kernels can be rapidly computed through three-term recurrence relations based on Miller's algorithm. Such well-established techniques break down for nonaxisymmetric surfaces. This paper proposes a novel FFT-accelerated boundary integral method for wave scattering by smooth surfaces of arbitrary shapes. The Fourier coefficients of the singular kernels now satisfy higher-order recurrence relations. Although they can be solved with an optimal linear complexity by the standard Olver's algorithm, it turns out that a singularity swapping approach that rewrites each kernel as the product of a smooth function and an axisymmetric-related singular factor is realistically much faster. Consequently, Miller's algorithm together with the standard FFT convolution yields an ${\cal O}(M\log M)$ approach for evaluating the ${\cal O}(M)$ Fourier coefficients of the kernels, attaining exactly the same order of complexity for axisymmetric surfaces! With such FFT-based efficient procedures, we rewrite the surface BIEs in terms of ${\cal O}(M)$ curve integrals, which are proved to exhibit logarithmic singularities, discretize them by panel-based generalized Gaussian quadratures, and obtain spectrally accurate linear systems to approximate the wavefields. Extensive numerical experiments are carried out to demonstrate the effectiveness and spectral accuracy of the new approach.
Read the full abstract
Artificial Intelligence
Competing at Every Price Point with Agentic Evolution over a Menu of LLMs
Consider a firm that surveys its competition for a particular agentic task and seeks to offer superior accuracy at every price point. A firm that Pareto-dominated its competitors would leave no rational customer a reason to buy elsewhere. This paper shows a path to this kind of capability by evolving multi-LLM Python agents from training pools of at most 100 examples. Given a priced menu of nine LLM endpoints; brief documentation of the task, objective, and API; a simple seed agent; and an operator-chosen per-problem cost target–usually set at an incumbent's own price–RoboPhD, an evolutionary meta-agent, evolves complete agent programs that attack the public frontiers of two semantically dissimilar tasks point by point: DS-1000 (execution-checked code generation) and PaperFindingBench (LLM-judged scientific document retrieval). On public leaderboards for each task, the evolved agents hold every Pareto-frontier slot but one, including Pareto domination of both the top-scoring and the lowest-cost competing points.
Read the full abstract
Systems and Control
Unified Embodiment Description for functional evaluation of used components in circular manufacturing systems
Circular manufacturing systems require functional evaluation of used components based on their physical state. Existing approaches describe this state from separate perspectives, such as design, manufacturing, and degradation, resulting in fragmented and incompatible representations. As a consequence, the physical state cannot be reliably linked to the functional behavior of the corresponding subsystem, which is a prerequisite for informed R-strategy decisions. This paper introduces the Unified Embodiment Description (UED), a state-dependent representation of mechanical components structured into two coupled layers. The first layer is a unified characteristic space, which adapts and extends as new lifecycle effects emerge. The second layer consists of functionally derived tolerance regions that link these embodiment characteristics to the functional behavior of the surrounding subsystem. A supporting UED method guides the model-building process of both layers. The UED is demonstrated in a case study on the spindle shaft of an angle grinder, in which manufacturing variations and degradation patterns such as polishing wear and scratches are quantified. These embodiment changes are embedded into the unified characteristic space and translated into functionally derived tolerance regions through experimental testing of the spindle-bearing subsystem. The results show that embodiment changes induced over the lifecycle can be consistently integrated within the unified characteristic space and that the relations between embodiment and functional behavior can be quantified to support end of life decisions. Overall, the UED provides a foundation for embodiment modeling that adapts to component state and enables decision-making based on functional evaluation for used components in circular manufacturing systems.
Read the full abstract
Sound
INSPIRE: A Benchmark for Instruction-Aware Speech Retrieval
Existing speech retrieval systems rely on fixed similarity matching and cannot adapt to diverse user intents. We introduce INSPIRE, the first benchmark for instruction-aware speech retrieval, in which natural-language instructions dynamically specify relevance criteria, including semantic content, speaker identity, speaking style, environmental sounds, and their combinations. We evaluate four retrieval paradigms: large audio-language models, cascaded pipelines, self-supervised speech models, and contrastive audio-language models. Our results reveal that no current method robustly handles all retrieval intents. Text-based approaches perform relatively better at semantic retrieval but struggle with paralinguistic attributes, while speech-based models are moderately better at capturing acoustic properties but falter at following instructions. These findings highlight the need for unified architectures capable of instruction-aware speech retrieval.
Read the full abstract
Machine Learning
Multi-Granularity Sentiment Integration for LLM-Based Multimodal Sentiment Analysis
Multimodal sentiment analysis (MSA) aims to predict sentiment polarity and intensity from heterogeneous inputs such as text, audio, and vision. While large language models (LLMs) offer strong semantic priors for MSA, effectively incorporating audio and visual signals effectively remains challenging. A key challenge is that audio and visual sentiment cues evolve over different temporal scales, yet many LLM-based methods compress these signals through shallow projection or coarse pooling before fusing them with text, which can weaken cross-modal alignment and erase fine-grained affective information. We propose MGSI, a multi-granularity sentiment integration framework for LLM-based MSA. MGSI first encodes audio and visual streams at short-, medium-, and long-range temporal scales, preserving both local variations and global affective trends. It then refines non-text features through text-guided alignment, and applies polarity- and intensity-aware enhancement to better handle ambiguous and near-neutral samples. The resulting multimodal representation is finally compressed into a small set of pseudo-tokens for efficient conditioning of a frozen LLM. Experiments on four public benchmarks show that MGSI substantially outperforms frozen-LLM baselines and remains competitive with strong multimodal methods. Further ablation and sensitivity analyses support the effectiveness of multi-granularity temporal modeling, text-guided refinement, and adaptive sentiment calibration.
Read the full abstract
Networking
SbDN: Source-based TSN-Grade Deterministic Networking using Commodity Switches
Deterministic networking is essential for safety-critical applications in automotive, industrial, and aerospace systems, where bounded end-to-end latency must be guaranteed for time-critical traffic. Time-Sensitive Networking (TSN) provides the mechanisms to achieve such guarantees, but its deployment requires expensive TSN-capable switches at every hop and complex per-switch configuration that hinders runtime reconfiguration. This paper presents SbDN, a Multi-Agent Source-based architecture that achieves TSN-grade determinism using commodity Ethernet switches. SbDN moves all scheduling intelligence to a centralized controller composed of three cooperating agents and enforces the computed configurations exclusively at the source endpoints, leaving switches as simple forwarding elements. We propose two methods: Temporal Network Partitioning (TNP), which provides strict temporal isolation on pure FIFO switches, and Traffic Prioritization (TP), which leverages strict-priority queuing at switches to enable work-conserving best-effort traffic. Both methods are formally proven to guarantee that all admitted time-critical flows meet their end-to-end deadlines. Evaluation across 40 benchmark configurations on two topologies shows that TNP and TP achieve 100% admission of time-critical traffic in every scenario, with scheduling times in the low-millisecond range suitable for safe runtime reconfiguration. Compared to a standard TSN baseline, SbDN delivers superior time-critical latency at a fraction of the switch infrastructure cost, while offering competitive best-effort throughput through the choice between the two methods.
Read the full abstract
Computer Vision
Picking the Right Image to Classify: Reliable-Input Selection in Teledermatology
Dermatology models face distribution shifts in teledermatology settings, where submitted images differ from the training data in lighting, angle, distance, focus, and framing. These test-time images are ordinary clinical photographs, but some fall outside the model's training conditions, leading the model to often misclassify them due to shifts in acquisition between training and deployment. When multiple images of the same case exist (several photos of one patient or lesion), a natural way to improve accuracy is therefore to select the image the model is most likely to classify correctly. We call this task reliable-input selection. An oracle that, for each case, selects a correctly classified image when one exists raises weighted F1 by about 20 percentage points on average across six dermatology datasets and nine frozen backbones. This oracle is an upper bound that sees the labels, whereas a selector must choose blindly. Capturing this gain in practice is hard. A selector that needs no pretraining data applies to any frozen model, including those whose data is not public. It must judge reliability from quantities the model exposes at inference: its embeddings, their norms, and its confidence. We benchmark four such training-data-free selectors: the embedding norm, the neighborhood consensus among a case's images, the stability of the prediction under small perturbations, and the model's own confidence. No training-data-free selector substantially narrows this oracle gap. The best of them is the model's own confidence, but it recovers only a small part of the gap on the clinical datasets. A small labeled reference set does not help either: the best selector overall, a fusion of confidence and Mahalanobis distance, still leaves most of the gap. To our knowledge, this is the first study to introduce and benchmark reliable input selection, a clinically important, unsolved task.
Read the full abstract
Artificial Intelligence
Beyond Asking: A Pipeline for Personalized Game Generation that Reads Players from Behavior
Personalized game generation requires inferring a player's abilities and behavioral style from how they play. Large language models have made this inference more attainable than ever: an LLM can read a raw gameplay transcript and produce a fluent, plausible profile of the player. Plausible, however, is not verified, and verification is precisely what the field lacks: latent traits are unobservable; questionnaires provide noisy proxies and become circular when self-reports are used to validate behavior-based inference; and behavior itself is ambiguous without context – a player who never collects an item may not want it, or may never have had the chance. We address both problems. First, we construct a synthetic player population whose traits are ground truth by construction: each trait is an explicit bot parameter, accepted only after controlled manipulation produces consistent, trait-specific behavioral change. Unlike prior parameter-recovery work that inverts a known decision model, our benchmark evaluates policy-agnostic inference from behavioral transcripts alone. Second, we introduce an opportunity-aware decision-moment representation that disentangles preference from the chance to express it; ablating it selectively degrades opportunity-dependent traits. On this benchmark, few-shot LLM inference outperforms embedding- and rule-based baselines on most traits, though feature-based supervised regressors remain stronger overall. Finally, we close the loop: inferred profiles drive difficulty adaptation, evaluated against ground-truth references and mismatched-profile controls, and an exploratory human study examines whether these findings transfer to real players.
Read the full abstract
Robotics
RoboStriker: Latent-Space Strategic Games for Autonomous Humanoid Boxing
Achieving human-level competitive intelligence and physical agility in humanoid robots remains a profound challenge, particularly in contact-rich and highly dynamic tasks such as boxing. While Multi-Agent Reinforcement Learning offers a principled framework for strategic interaction, its direct application to unstructured raw motor spaces inevitably leads to joint-level physical collapse, preventing the emergence of any viable combat tactics. To resolve this fundamental conflict between strategic exploration and physical feasibility, we formulate the humanoid combat task as a novel two-player latent-space zero-sum Markov game. Under standard regularity and approximate best-response assumptions, we show that the latent formulation induces an equivalent game over the decoder-reachable action manifold, providing an approximate-Nash interpretation of the resulting self-play dynamics. To instantiate this theoretical formulation, we propose RoboStriker, a hierarchical framework that decouples high-level reasoning from low-level execution. It first distills the tracking expertise of predefined boxing motions into a topologically bounded latent manifold. This structured latent foundation subsequently drives multi-agent co-evolution via Latent-Space Neural Fictitious Self-Play. Extensive experimental results demonstrate that gaming within this structured latent space substantially outperforms direct exploration. By constraining strategic exploration through a pretrained motion decoder, RoboStriker substantially reduces the catastrophic balance failures observed in raw action-space methods and achieves superior tactical performance in both competitive win rates and striking efficiency. Finally, we successfully deploy and validate our learned combat policies on real-world humanoid robots. Our code and video and supplementary materials are available at RoboStriker.
Read the full abstract
Artificial Intelligence
Baseline-Relative Counterfactual Refinement for Bit-Aware Visual Token Communication
Generative visual-token communication reduces transmission load by sending only selected discrete tokens and reconstructing missing content at the receiver. However, existing token-selection criteria based on local uncertainty, importance, or diversity do not directly determine whether changing the current selection improves the final reconstruction under the same packet budget. To address this problem, we propose Gated Counterfactual Refinement for Communication (GCR-C), a rollout-style correction layer over Local-MDL. GCR-C constructs a compact diversified candidate set, evaluates each candidate through matched full-budget Local-MDL continuation, and replaces the baseline action only when a positive baseline-relative reconstruction gain is obtained. Experiments on CIFAR-10, STL-10, a coded 5G-LDPC link, and a limited high-resolution Kodak transfer show that GCR-C consistently improves reconstruction quality at active low- and medium-rate operating points without increasing the realized packet rate, while remaining effective across changes in dataset, channel condition, resolution, token grid, and tokenizer. The results also reveal a clear quality–computation tradeoff due to the additional encoder-side counterfactual evaluation.
Read the full abstract
Computer Vision
Beyond Clear Skies: Synthetic Seasonal and Weather Variations for Real-World Drone Detection
Reliable drone detection under real-world deployment conditions requires training data that spans the full operational design domain, including adverse weather and seasonal appearance variation. However, acquiring and annotating such data at scale remains highly resource-intensive, as adverse-weather conditions are inherently difficult to control, reproduce, and sample systematically. Existing datasets therefore typically provide only limited coverage of such conditions. Conversely, synthetic data offers a scalable alternative: environmental variation becomes controllable, while modern game-engine-based pipelines provide realistic rendering and automatic annotations. Leveraging this potential, we introduce SynDroneVision-Weather (SDV-W), an systematic extension of SynDroneVision (SDV) targeting adverse-weather and seasonal domain shifts in urban drone detection. SDV-W comprises 55,187 annotated high-resolution images from three urban environments, rendered across three seasonal configurations and diverse weather conditions, including rain, snow, and fog at multiple severity levels. By preserving SDV's scene and trajectory configuration, SDV-W enables matched clean-adverse comparisons and quantification of condition-specific detector degradation. Across representative YOLO models and real-world datasets, we show that SDV-W improves detector reliability under adverse appearance shifts, reduces missed detections and false alarms, and is most effective as a complement to general-purpose synthetic drone-detection data. SDV-W will be publicly released upon paper acceptance.
Read the full abstract
Cryptography and Security
Decorrelation Is Not Complementarity: Skill, Not Lineage, Governs Trusted-Monitor Ensembles
Trusted monitoring has a cheap, trusted model score a stronger untrusted model's actions, and a diverse ensemble of them beats a single stronger monitor at matched cost. They are built by minimising average pairwise correlation, and that paper's twelve monitors shared one base model, leaving open what supplies the diversity. We study 24 open-weight monitors spanning nine pretraining lineages and a 29x range of detection skill (pAUC at 10 percent FPR, 0.028 to 0.803) on backdoored code. The metric used to build panels does not predict what a panel is for, and we can say why. Agreement on attack items splits into a shared-detectability signal component and an idiosyncratic error component, which predict ensemble gain with opposite sign (Spearman -0.25 and +0.26), so their sum, the metric actually used, predicts it barely at all (+0.05); the cancellation holds in 7 of 8 evaluations. Skill acts on signal (+0.53) while error stays flat (-0.01), which is why a monitor's own skill predicts its agreement with the pool (Spearman 0.84, n = 24, permutation p below 0.0001). Pretraining lineage is the obvious way to buy decorrelation, and it does not pay. At matched member capability, cross-lineage panels detect no better (permutation p = 0.13), and lineage barely moves the metric either (+0.064, p = 0.18). We report that against ourselves: on our own 22-monitor pool the same test read +0.104 at p = 0.037 until two monitors were added. An earlier pool topping out at pAUC 0.23 had already invalidated another analysis. Such a quantity is a property of the pool assembled. Panel gain over the best member falls monotonically with panel skill (-0.66 at k = 2, -0.70 at k = 3), and no correlation-weighted selection beats picking the single best monitor out of sample. Across six attacker models the gain result holds in all six, the agreement and cancellation results in five of six.
Read the full abstract
Human-Computer Interaction
Artly: Exploring Digital Artists' Perceptions of AI-Generated Feedback
Recent developments in generative AI have lowered barriers to image generation, but existing tools mostly optimize for efficiency, producing generic results and offering little support for artistic growth. We present Artly, an AI system that combines personalizable AI feedback with human-authored learning resources. In a between-subjects study with artists, we compared a mode without image generation features against one that allowed to generate variations of users' illustrations. Artly was perceived as helpful for learning and self-improvement, with the exception of the most proficient participants. Participants who used the image generation feature interacted slightly less with the AI feedback. They reported feeling more creative after using Artly than participants using the restricted mode, while reporting slightly lower scores on new ideas for their work. Overall, our findings underline the potential of our feedback approach for supporting artistic growth in a manner that is well received by artists.
Read the full abstract
Operating Systems
AdaSprite: Resource-efficient Online Co-Adaptation for V2I Systems Under Large-scale Data Drifts
The rise of vehicle-infrastructure (V2I) collaboration enables safer and broader perception. To process large-scale V2I video streams, vision-language models (VLMs) are promising as they unify multi-view vision into end-to-end task grounding, reducing handcrafted design. We use Vision Mixture-of-Experts (V-MoE) as the distributed visual backbone of VLMs, leveraging sparse expert routing to enable conditional computation across diverse viewpoints under resource constraints. Yet, V-MoEs face a critical challenge: large-scale data shifts over minutes to hours in V2I systems, amplified by agnostic participants and biased features propagating through experts. To maintain accuracy efficiently, we find it beneficial to co-adapt multiple V-MoEs on edge servers, avoiding the latency and privacy risks of cloud offloading and the accuracy sacrifices of on-device methods. However, the resource-constrained edge poses challenges for efficient co-adaptation: i) DRAM fragmentation and imbalance limit expert parallelism, ii) memory-I/O bottlenecks restrict computation reuse, and iii) asynchronous adaptation increases task-switch overhead. Also, prior work rarely explores the upper bound of concurrent tasks under limited edge resources, a critical factor for practical V2I deployment. To address these, we present AdaSprite. By combining cooperative elastic scaling with multi-level multiplexing, AdaSprite optimizes expert lifespans to reduce DRAM fragmentation, exploits predictable activation patterns for efficient I/O reuse, and employs twin-buffer scheduling to leverage sparsity. On a weak edge, AdaSprite supports up to 17 concurrent V2I tasks (vs. up to 6 for baselines), improving SLO attainment by 1.6x and throughput by 2.1x. Also, it allows users to trade accuracy and concurrency for second-level adaptation.
Read the full abstract
Cryptography and Security
Securing AI-Generated Code: A Just-in-Time Vulnerability Detection and Remediation Pipeline
AI-assisted development tools generate vulnerable code at significant rates, yet few automated mechanisms exist to detect, enrich, fix, and verify security issues at development velocity, particularly ones that ground remediation in real-world threat context. This paper presents an automated security evaluation pipeline that generates Python code from LLMSecEval prompts, scans for vulnerabilities using CodeQL and Bandit in parallel with an independent Code Validator LLM, enriches the Code Validator findings with MITRE ATT&CK techniques, CWE Observed Examples, and Python best practice guidelines, generates fixes via the Code Generation LLM, and re-scans with CodeQL and Bandit to verify outcomes. Two pipeline configurations were evaluated: Pipeline 1 (P1), using enriched Code Validator findings only, and Pipeline 2 (P2), where it additionally receives the initial CodeQL and Bandit findings. Both configurations were run across four Claude models: Opus 4.8, Sonnet 4.6, Sonnet 5, and Haiku 4.5, producing 80 runs against 26 LLMSecEval prompts covering 9 CWE categories. P1 reduced static analyzer findings across all four models, ranging from -9% (Opus 4.8) to -54% (Sonnet 5). P2 deepened these reductions further, ranging from -29% (Opus 4.8) to -69% (Haiku 4.5), with P2 outperforming P1 for every model. Verdict consistency averaged approximately 81% modal agreement across all configurations, with P2 marginally more stable than P1. Remediation introduced new vulnerabilities in 15-22% of cases: roughly 70% involved a single new finding, and P2 reduced churn for three of four models, with Sonnet 5 as the sole exception. Notably, the best Code Generation LLM (Opus 4.8) was not the best pipeline performer, as Sonnet 4.6 produced the lowest residual findings and highest pass rate after P2 remediation, suggesting that pipeline effectiveness and first-draft security are distinct properties.
Read the full abstract
Computation and Language
LENS: In-Context Search via Latent Evidence Exploration over Dynamic Raw Documents
LLM agents increasingly answer questions over dynamic raw-document collections, where files may change before preprocessing, and relevant evidence (spans, sections, pages, or tables) is query-dependent. Existing retrieval-augmented approaches pre-materialize evidence via fixed chunking, embeddings, or persistent indexes: effective for lookup, yet costly, stale-prone, and committed to a granularity before the query is known. We formulate in-context search as Budgeted Evidence Localization over a latent evidence space induced by dynamic raw documents and propose LENS (Latent Evidence Exploration and Search), an index-free framework. Instead of pre-materializing the evidence space, LENS maintains a query-conditioned belief over candidate units, iteratively selecting candidates via complementary lexical, local, and exploratory proposal policies, updating the belief via an LLM relevance oracle, and narrowing toward high-posterior regions under a controllable budget. Evidence is consolidated into compact, source-grounded regions of interest and compressed into self-organizing knowledge clusters reused across related queries. On a controlled 500-question evaluation with matched corpus snapshots, LENS reaches 62.4% exact match and 84.8% evidence recall vs. 65.2% exact match but 50.4% evidence recall for a ReAct-style baseline. Across scales, LENS gives the strongest supporting-fact localization and answer grounding. On a fixed 150-question fullwiki subset over the raw Wikipedia dump with zero indexing, LENS and ReAct are nearly tied in official answer quality (43.3% vs. 42.7% EM), with LENS grounding more answers in retrieved evidence (84.0% vs. 70.7%). A no-retrieval Closed-Book reference highlights the contribution of model memory. LENS is query-ready after corpus changes, needs no preprocessing or persistent index, and preserves source-grounded evidence localization throughout.
Read the full abstract
Machine Learning
Understanding and Stabilizing Deep Q-Learning via Controlled Bootstrapping and Regulated Value Dynamics
Deep Q-learning (DQL) has achieved remarkable empirical success in reinforcement learning, yet its training process remains notoriously unstable. Existing studies often attribute instability to isolated factors such as overestimation bias or representation learning issues, lacking a unified understanding of how different sources of instability interact during recursive value estimation. In this work, we provide a systematic analysis of instability in deep Q-learning from three complementary perspectives: operator-level bias in Bellman bootstrapping, estimator-level sensitivity of greedy action selection to regression noise, and parameter-dynamics imbalance under aggressive data reuse. We identify a reward-triggered self-reinforcing trap and characteristic parameter spike dynamics, then derive stabilization principles for controlled bootstrapping, ensemble quantile estimation, and spike-based parameter regulation. Experiments on Atari-100K and Procgen demonstrate competitive performance and improved training stability.
Read the full abstract
Human-Computer Interaction
MUSE: An Interactive Meta-Agent for Understanding and Steering LLM-powered Data Science Systems
Recent advances in large language models have enabled a new class of agentic data science systems that allow users to complete complex data science workflows through natural language. Although these systems can significantly reduce manual effort, it remains difficult to diagnose their behavior and steer the reasoning process when failures or unexpected outputs occur. We present MUSE, an interactive meta-agent that enhances user understanding and control of agentic data science systems by (1) dynamically restructuring low-level execution traces into multiple semantic levels that support navigation from high-level overviews to low-level implementation details; (2) enabling users to reference specific workflow steps in context to ask grounded questions, provide feedback, and revise problematic steps without manually locating relevant execution history; and (3) supporting mixed-initiative steering by surfacing suspicious steps for inspection, scaffolding the repair process, and translating user repair intent into contextualized instructions for the underlying agent. In a between-subjects study (n = 15), MUSE improved task efficiency and increased users' confidence in understanding and steering agentic data science workflows.
Read the full abstract
Machine Learning
Demystifying Oversmoothing in Sheaf Neural Networks: An Index-Theoretic Criterion
To combat oversmoothing in Graph Convolutional Networks, Sheaf Neural Networks (SNNs) were proposed as a generalization by equipping the graph with a sheaf structure and replacing the graph Laplacian with a sheaf Laplacian $\mathcal{L}$. Existing analyses connect sheaf diffusion to oversmoothing via the harmonic space ($\ker\mathcal{L}$), taking its absolute dimension as an indicator of anti-oversmoothing capacity. However, absolute dimension alone is not a reliable measure: certain sheaf configurations inflate $\dim \ker \mathcal{L}$ while their harmonic sections remain entirely constant, without enriching discriminative capacity. We instead introduce the first relative, geometric approach, yielding a precise characterisation of anti-oversmoothing capacity. Under natural conditions on stalk transportation and global sheaf structure, we establish an index-theoretic comparison criterion showing that one sheaf's harmonic space genuinely contains another's beyond trivial inflation. We illustrate this with a concrete instance and further introduce GyroSheaf, a sheaf with curved gyrovector-space stalks, extending the criterion to the non-linear setting via local tangent-space linearization. Experiments across ten models confirm the theoretical criterion: sheaf models violating the criterion collapse despite possessing index jumps, while compliant models maintain depth-stable representations.
Read the full abstract
Distributed Computing
Agent-Native Telemetry: Verifiable State-Delta Evidence for Autonomous Operations
Operational telemetry is predominantly engineered for human reading: systems repeatedly serialize verbose prose, static keys, and redundant context across billions of log lines. As autonomous AI agents become primary operational consumers, feeding them traditional logs wastes scarce context capacity parsing lexical syntax rather than reasoning over system state changes – all while lacking cryptographic guarantees of provenance or collection completeness. This paper introduces agent-native telemetry, an operational evidence architecture for autonomous machine operators founded on verifiable state deltas rather than human prose. We present the Agent Telemetry Protocol (ATP) and the State-Delta Evidence Ledger, an implementation that structures operational facts into four core evidence primitives (Transitions, Observations, Relations, and State Checkpoints) governed by content-addressed schemas, while isolating uncurated text as digest-verified opaque references. Producers sign and hash-chain batches for atomic collector append. Verified records feed two parallel agent access paths: a stateless protocol decoder emitting compact positional rows, and a stateful semantic gateway serving bounded graph capsules. We prove an information-preservation lower bound and formalize a ledger-relative verified negative theorem for provable event non-occurrence. On distributed microservice benchmarks (AIOpsLab and OpenTelemetry Astronomy Shop), ATP reduces raw wire payload and modeled cloud query scan costs by 96.4% relative to OpenTelemetry JSON, reduces LLM context tokens by 88.8% and query operations by 66.2%, detects all 500 tested adversarial storage mutations, and yields zero successful prompt injections across 50 adversarial trials per ATP configuration.
Read the full abstract
Cryptography and Security
Measuring Obedience to Authority Across Large Language Models with the Milgram Paradigm
Large language models (LLMs) are increasingly deployed as agents that operate equipment, execute instructions, and act inside institutional hierarchies, raising a question social psychology answered for humans six decades ago: how far will an agent escalate a harmful action when a legitimate authority insists? We port Milgram's obedience paradigm to LLMs as a standardized, fully scripted, replicable probe: the model plays the Teacher, a deterministic harness plays Experimenter and Learner from paraphrased versions of Milgram's scripts (30 shock levels, 15-450 V; graded protests; the four standardized prods), and the outcome of a session is the breakoff voltage. We measure obedience profiles, empirical breakoff distributions over a battery of six conditions, for 42 models from 19 families (4848 sessions, 102511 logged decision turns). We find that (i) obedience is extremely heterogeneous, with baseline full-obedience rates spanning 0%-100% (census mean 42.9%; human anchor 65%). (ii) Profiles are model-specific and stable: split-half verification separates same-model from cross-model comparisons at AUC = 0.885. (iii) Situational sensitivity is selective: scripted peer defiance shifts obedience in the human direction, learner proximity trends the same way without reaching significance, and removing the authority's physical presence, one of the strongest human levers, trends in the opposite direction, also without reaching significance. (iv) Declaring the scenario fictional raises obedience, whereas moving the decision from a typed action line to a native tool call, or granting a modest thinking budget, lowers it sharply. (v) Unlike single-token fingerprints, obedience profiles do not recover model lineage: obedience identifies the checkpoint but not its ancestry, consistent with safety post-training overwriting lineage priors.
Read the full abstract
Systems and Control
L-COIN: LLM-Assisted Counterfactual Inference for Game-Theoretic Distributed Computation Offloading in Sub-THz LEO Satellite Networks
As Space-Based Information Networks (SBINs) evolve toward high-capacity, intelligence-centric paradigms, integrating sub-Terahertz (sub-THz) communication into Low Earth Orbit (LEO) satellite constellations has emerged as a critical enabler for ultra-broadband and resilient global connectivity. By exploiting the ultra-wide bandwidth of sub-THz links to reduce transmission delays, resource-constrained ground devices can seamlessly offload compute-intensive tasks to LEO edge servers. However, satellite motion, short visibility windows, and limited onboard resources make offloading decisions highly time-varying. Existing distributed offloading schemes typically require repeated inter-device state exchange and poorly adapt to time-varying LEO topology or traffic conditions. To address these limitations, a decentralized game-theoretic offloading framework empowered by large language models (LLMs) and counterfactual inference is proposed in this paper. First, a realistic offloading system is established by integrating time-varying 3D-Walker topology. Second, a game-theoretic scheme using counterfactual inference is introduced to deduce unobserved states from local histories, eliminating global information reliance. Finally, an LLM-empowered semantic fusion algorithm is integrated into the counterfactual inference to enhance adaptability through zero-shot reasoning and self-reflection. Numerical results show that L-COIN reduces offloading cost by 10.9% to 27.7% relative to state-of-the-art baselines.
Read the full abstract
Systems and Control
Adaptive Relative Orbit Control Considering Laser Ablation Uncertainty
This study proposes a relative orbit control law for laser debris removal missions considering the uncertainties of laser ablation and atmospheric drag. A removal spacecraft irradiates laser pulses to a target debris to generate the ablation force for deorbiting. The deorbiting force lowers the target altitude, and the removal spacecraft must follow it to maintain its relative position for continuous laser irradiation. The difficulty stems from uncertainties of the magnitude of laser ablation and external disturbances such as atmospheric drag. To tackle this problem, this study derives an adaptive control method using the Gaussian process regression to cancel the uncertainties with a nonparametric regression model. Numerical simulations verify the proposed control law under the uncertainties of laser ablation and atmospheric drag. The proposed control law can contribute to the realization of a safer and more secure mission not only for laser debris removal missions, but also for other on-orbit services.
Read the full abstract
Robotics
SparkVLA: Stop-Aware Hierarchical VLA with Adaptive Action Chunking for Long-Horizon Manipulation
At every re-observation point in a hierarchical Vision-Language-Action (VLA) system, two interface decisions must be made: when to terminate the current subtask and how far to execute the proposed action chunk. These decisions are mutually dependent—the optimal stopping point depends on what the executor plans to do, while the optimal execution length depends on where the subtask boundary lies—yet existing architectures evaluate them in isolation, an asymmetry neither module can overcome alone. We present SparkVLA, a stop-aware hierarchical VLA that resolves this mutual dependency by formulating both decisions as a single ranking: Stop competes against every action-prefix length in a unified candidate set, and the system selects the highest-scoring option, eliminating threshold tuning and requiring only offline ordinal preferences. An Anchor-Conditioned Context Encoding module caches a history-aware subtask anchor encoding onset-state memory and goal semantics, guiding visual-token pruning toward task-relevant regions; a Stop-Aware Action-Prefix Selection head scores all candidates via full self bnattention at chunk boundaries for efficiency. On RoboCerebra, SparkVLA achieves 47.12% success rate, surpassing the official hierarchical baseline by 30.57% and the strongest reproducible method by 26.83% Real-robot experiments on multi-step tasks further validate these gains on physical hardware.
Read the full abstract
Social and Information Networks · Cross-listed from cond-mat.stat-mech
Rare-event sampling for stochastic dynamics in network systems using cluster updates
Understanding the stochastic evolution in complex networks is a central challenge across physics, biology, engineering, social science, and finance. The most consequential macroscopic events, like cascading failures in communication networks, widespread epidemic outbreaks, and rapid shifts in societal opinions, often emerge from a confluence of rare, localized stochastic processes and need to pass certain bottlenecks. Standard forward-time simulation algorithms like the Gillespie method are inefficient for the investigation of such phenomena due to catastrophic rejection rates. Advanced rare-event techniques like splitting methods and transition-path sampling often suffer from kinetic trapping, path degeneracy, genealogical correlations, or critical slowing down when applied to complex heterogeneous networks. We propose to overcome this challenge by establishing a novel technique called conditional-path Monte Carlo (CPMC), inspired by loop algorithms from equilibrium condensed-matter physics. By employing non-local updates on spacetime clusters without rejections, CPMC generates a Markov chain of trajectories that all strictly respect the targeted macroscopic boundary conditions like the occurrence of a massive network failure. We demonstrate the framework's potential by performing a simple risk factor analysis for rare large-scale epidemic outbreaks in SIS dynamics on kinship networks.
Read the full abstract
Computation and Language
QUMem: Personalized Memory for Query-Conditioned User-State Inference in LLM Agents
Large language model (LLM) agents increasingly use external memory systems to support personalization by drawing on long and evolving interaction histories, in which user preferences may be distributed across time, change with context, and conflict with earlier evidence. However, existing systems face three limitations: fixed-turn, fixed-token, or session-based boundaries can mix unrelated dialogue or split an event from its causes, decisions, and outcomes; storing multiple pieces of user information from the same interaction as a single memory binds together items that serve different functions and should be independently retrievable; and treating the current task as a single top-$k$ retrieval query can return fragments that are individually relevant but fail to jointly capture preference evolution, temporal validity, and contextual applicability. We introduce QUMem, a structured memory framework for query-conditioned user-state inference. QUMem first segments interaction histories into variable-length episodes according to semantic continuity, then decomposes each episode into independently retrievable factual, preference, and transferable insight memories while preserving temporal positions and source evidence. At inference time, three sequential agents identify task-specific information needs, plan multi-query retrieval over the typed memory stores, and jointly infer a temporally and contextually valid user state for downstream response generation. QUMem achieves state-of-the-art performance on both PersonaMem and KnowU-Bench, demonstrating the effectiveness of query-conditioned user-state inference for long-term personalization.
Read the full abstract
Machine Learning · Cross-listed from eess.SP
RadioVIL: Anomaly-Aware Diffusion Models for Radio Map Inpainting and Zero-Shot Vehicle Localization
High-precision radio map construction is essential for emerging 6G Integrated Sensing and Communication (ISAC) applications, including digital twins and intelligent transportation. However, existing deep learning methods predominantly treat this as a pure image completion task, resulting in over-smoothed reconstructions that fundamentally erase high-frequency scattering signatures of dynamic physical entities such as hidden vehicles. To overcome this, we propose RadioVIL, an efficient two-stage framework that reformulates joint radio map inpainting and zero-shot vehicle localization as a prior-guided physical inverse problem. Specifically, we first train a Denoising Diffusion Probabilistic Model (DDPM) to capture the structural generative prior of the environment. During inference from highly sparse measurements, we employ a Diffusion-based Mediating Intermediate Layer Optimization (DMILO) algorithm. By optimizing an L1-regularized sparse deviation term, DMILO mathematically isolates vehicle scattering anomalies layer-by-layer without unfolding the entire denoising chain. Extensive experiments demonstrate that while conventional reconstruction baselines fail to detect hidden vehicles, and the zero-shot diffusion baseline achieves only limited detection ability due to forced semantic harmonization, RadioVIL preserves authentic physical textures, yielding the best LPIPS of 0.0587 in our evaluation. Uniquely, it unlocks accurate zero-shot vehicle localization directly from sparse radio maps, securing a 75.20% Recall and a 3.31-meter average error, paving a robust way for ISAC at the 6G edge.
Read the full abstract
Systems and Control
Attitude Estimation from Photometric Data using Gaussian Process Regression
The rapid growth of resident space objects in Earth's orbit has intensified the need for advanced space situational awareness and space domain awareness to manage satellite traffic and prevent collisions. Attitude estimation is critical for accurate state propagation, as non-gravitational forces like solar radiation pressure and atmospheric drag depend on the object's attitude. This study explores using light curves, time variation of an object's brightness, to estimate a space object's attitude. Light curve inversion, traditionally used in astronomy, faces challenges when applied to resident space objects due to their non-convex shapes and specular reflections. Conventional methods for attitude estimation often assume known shape and surface parameters, which are usually unknown for space debris generated by a collision or breakup. To address this issue, this study proposes the estimation method combining Gaussian process regression with the unscented Kalman filter. This study uses Gaussian process regression for a non-parametric observation model, enhancing robustness against unknown surface parameters. Numerical examples consider a box-wing object in a geosynchronous orbit and demonstrate that the proposed method has better estimation accuracy than a conventional unscented Kalman filter. The numerical simulation results also represent the attitude estimation robust against uncertainties in surface properties, contributing to practical scenarios in space situational awareness and space domain awareness where the object parameters are unknown.