Papers

Retrieval-augmented generation (RAG) is hard to monitor in production: exhaustive relevance labels do not exist for non-stationary multi-million-passage corpora that re-index in real time. As a result, retrieval quality is generally understudied and often deprioritised in favour of generation-oriented metrics. In this work, we propose auditing retrieval coverage by probing for evidence of missing documents rather than enumerating every relevant one. Our method Re:CAP (REtrieval Coverage Audit by iterative Probing) is a reference-free audit loop applied to a deployed RAG pipeline's initial answer and retrieved context: it identifies the topics already covered, generates probing questions for plausibly missing topics, retrieves candidate documents, and applies an LLM-as-judge to retain only those that introduce previously-unretrieved information. On four public benchmarks, Re:CAP recovers 9-29% of gold labels that flat BM25 top-500 cannot reach, rising to 48% on TREC-COVID. On MuSiQue Re:CAP beats flat hybrid top-500 by +12.9 pp on recall at less than half the document budget. An ensemble BM25, dense, and hybrid baseline (top-500 each) still leaves out 21.2% of gold docs on TREC-COVID that Re:CAP recovers; human annotators judge that 78.9% of those structurally distinct documents add new information to the baseline answer (Fleiss $\kappa$ = 0.79, n = 123), and 73.9% on live production traffic (n = 180). End-to-end recall is reproducible to within $\pm$1% across three independent runs, making Re:CAP a stable instrument for periodic retrieval audits.

Orthogonal time frequency space (OTFS) modulation is well suited for high-mobility communications, but channel estimation and efficient multi-user transmission remain challenging. In this paper, we develop an OTFS-rate splitting multiple access (RSMA) framework that jointly considers pilot transmission, channel estimation, message detection, and resource allocation under imperfect channel state information (CSI). We consider guard-based and superimposed transmission for the common and private messages to account for the tradeoff between channel-estimation reliability and spectral efficiency (SE). We derive a linear minimum mean square error (LMMSE) channel estimator, error-aware message-passing detectors, and tractable surrogate rate expressions that account for channel-estimation errors. Based on these expressions, we formulate a joint resource-allocation problem and develop an SCA-based two-dimensional search algorithm together with a low-complexity alternative that exploits the analytical structure of the resource allocation. Numerical results demonstrate reliable channel-estimation and BER performance and show that configurations with guard-based common-message transmission achieve favorable sum-SE performance, while the preferred private-message structure depends on the operating regime. The low-complexity algorithm achieves near-SCA performance with substantially reduced computation time, and OTFS-RSMA remains competitive with OTFS-NOMA while requiring only a single SIC stage.

In this paper, we provide a single-letter characterization of the exact common information for a pair of correlated discrete sources, as well as a single-letter characterization of the admissible region for the shared randomness rate and the communication rate in exact channel synthesis.

Spectrum cartography is crucial for spectrum-aware resource management in low-altitude networks, where uncrewed aerial vehicles (UAVs) collect measurements to reconstruct power spectral density (PSD) radio maps. However, limited UAV energy budgets constrain both sampling density and onboard computation, while complex urban blockages hinder reliable measurement transmission to a remote station. To address these challenges, we propose an air-ground cooperative framework for active online spectrum cartography. Through adaptive bandwidth allocation, the UAV actively collects PSD measurements using uncertainty and simultaneously transmits them to a mobile uncrewed ground vehicle (UGV), while the UGV maintains a reliable air-ground link for map updates. Specifically, we first develop an online deep-unfolded tensor decomposition method to accelerate online map updates. We then derive a bit-depth-dependent interpolation-error model to guide the selection of quantization bit depths to balance overhead and accuracy. Finally, we use uncertainty information to coordinate the UAV and UGV for active sampling and link maintenance, respectively. Extensive experiments show that the proposed reconstruction method achieves a nearly 27-fold speedup over online tensor decomposition. In urban scenes, the proposed framework outperforms the representative baselines, reducing both normalized mean squared error (NMSE) and outage ratio by more than 23% each and increasing the packet-service completion ratio by over 4%.

Vision-Language-Action (VLA) policies achieve strong performance in robotic manipulation but remain brittle once execution deviates from nominal trajectories. We propose CARE (Corrective Atomic Robotic Execution), a framework that improves recovery by learning from failures encountered during execution. Instead of generating corrective data from manually designed or random perturbations, CARE collects failed rollouts, models stage-conditioned post-failure deviations, and uses the resulting empirical distributions to synthesize representative failure states and corrective demonstrations. At inference time, CARE combines stage-wise planning with physically grounded 3D monitoring to trigger atomic adjustments or re-operations while preserving task progress. We further introduce the Failure State Recovery Benchmark (FSR-Bench), which evaluates recovery from intermediate failure states under local deviations and structural anomalies. Experiments across multiple VLA backbones, simulation benchmarks, and real-world dual-arm tasks show consistent improvements, with average task-success gains of 14.5 points in simulation and 15.9 points in the real world. Code, models, and data are available at https://github.com/xiaojunlan/care

Identifying linear time-invariant (LTI) dynamical systems is challenging when trajectories are short, noisy, or high-dimensional. Traditional system identification typically treats each system independently and cannot exploit shared structure across related systems. We propose PBML-LTI, a PAC-Bayesian meta-learning framework for few-shot LTI system identification that learns a transferable prior over task-specific dynamics while preserving task heterogeneity. Each task corresponds to an unknown LTI system, and the meta-learner uses training trajectories to learn a data-dependent prior over transition matrices. For a new system with limited data, PBML-LTI performs Bayesian adaptation under this prior to obtain a task-specific posterior, providing accurate estimates and principled uncertainty quantification. A key challenge is temporal dependence, since LTI trajectories violate the i.i.d. assumptions underlying most PAC-Bayes meta-learning analyses. We address this with a martingale PAC-Bayes analysis for dependent trajectory losses and derive a support-query predictive-risk bound that motivates a fit-KL meta-training objective. The bound clarifies the roles of empirical fit, posterior complexity, and prior quality in few-shot adaptation under sequential dependence. We further derive corollaries for transition-matrix recovery and multi-step trajectory prediction, connecting uncertainty-aware meta-identification with finite-sample guarantees for dependent dynamical data.

Although deep learning has advanced Whole Slide Image (WSI) Analysis, tissue artifacts like bubbles and folds often cause silent failures by concealing essential morphology. Current pathology image restoration methods are mostly restricted to small patches, struggling to maintain global structural coherence at a megapixel scale. We introduce RestorePath, a framework for globally consistent megapixel scale inpainting that reconstructs diagnostic structures in histological image to prevent incorrect high-confidence predictions and lower error rates. Our model utilizes a Latent Diffusion Model (LDM) conditioned on Pathology Foundation Model (PFM) embeddings, integrating Large Kernel Attention (LKA) to manage long-range dependencies during random patch diffusion. Enhanced by Distance-Weighted Interpolation (DWI) and an Adaptive Guidance Scale (AGS), RestorePath ensures structural consistency and fidelity by modulating information from surrounding patches. Evaluations across TCGA-BRCA, BACH, and Camelyon16 datasets for images ranging from 512 to 4608 pixels demonstrate state-of-the-art performance in maintaining histological consistency. RestorePath significantly improves downstream Computational Pathology (CP) tasks, outperforming both raw artifact images and the conventional Detect-and-Discard (D&D) approach. The code is available at https://github.com/PathfinderLab/RestorePath

Tool-calling LLM agents are increasingly deployed in enterprise applications. However, effective evaluation and optimization require high-quality, diverse task datasets that are often difficult to obtain due to privacy and other constraints. Existing synthetic task generation methods often produce generic tasks that ignore an agent's underlying state or database and fail to reflect real-world usage diversity. We propose EdgeGen, a synthetic task generation framework that extracts compliance rules from an agent's specification and uses them to generate database-grounded edge-case tasks designed to violate these rules. When combined with existing synthetic data generation techniques, EdgeGen enables agent improvement through finetuning and harness optimization. The resulting pipeline forms a fully automated closed-loop system that requires no human annotation. Finetuning on data generated by EdgeGen yields a consistent mean progress improvement of 2 percent to 42 percent on tau2bench airline domain, while other baseline methods show degradation for some models. On the other hand, for harness optimization, our method shows a mean progress improvement of 10 percent and 30 percent over the human-curated and base harnesses, respectively, for the Gemma-4-e4b model.

We establish a new method for the study of imaginary non-CM fields of arbitrary \(2\)-power degree and class number one. In particular, we show that for degrees greater than \(16\), dihedral fields either lie in the ray class field of one of the three imaginary quadratic fields \(\mathbb{Q}(\sqrt{-2})\), \(\mathbb{Q}(\sqrt{-3})\),\(\mathbb{Q}(\sqrt{-19})\),\(\mathbb{Q}(\sqrt{-43})\), \(\mathbb{Q}(\sqrt{-67})\), whose conductor is supported only on primes dividing \(2\), with explicit upper bounds on the relevant exponents, or are Hilbert class fields of imaginary quadratic fields. We also give a complete classification of the \(161\) imaginary non-CM fields of degree \(16\) with class number one, providing explicit tables of defining polynomials, Galois groups, base fields, and conductors. These \(161\) non-CM fields, together with the six CM fields classified by Louboutin and Okazaki \cite{lou7}, complete the classification of all imaginary fields of degree \(16\) with class number one. For CM fields, the problem reduces to computing relative class numbers via analytic class number formulae, whereas for non-CM fields the absence of a totally real subfield of index \(2\) requires a different approach and leads to a richer variety of Galois groups and ramification patterns.

We give a unified study of local derivations on conformal Galilei Lie algebras in three natural settings: the algebra without central extension, the mass central extension, and the exotic central extension. We first determine the structural form of the derivation algebras and isolate the relevant outer derivations. For the mass extension, we complete the one-string rigidity at the two low parameters not covered by the previous one-dimensional treatment, using direct calculations, and combine this with new spatial compatibility arguments to obtain the result in arbitrary spatial dimension. The exotic extension is treated by a self-contained argument from its defining relations. For the algebra without central extension we resolve the remaining algebraic-reflexivity problem completely. If $d\ge2$, every local derivation is again a derivation. In spatial dimension $d=1$, the same conclusion holds when $2\ell$ is even and also for $\ell=\frac12$; however, for odd $2\ell\ge3$ there is an additional $(2\ell-1)$-dimensional space of pure local derivations. This gives a sharp contrast with the mass central extension, where the central Heisenberg pairing removes precisely this extra local freedom. We also determine the Lie algebra structure of the full local-derivation space: in the exceptional one-dimensional non-central case it is an explicit semidirect product with an additional abelian irreducible ideal.

Causal Bayesian Optimization (CBO) combines causal inference with Bayesian optimization to enable sample-efficient intervention selection in systems with causal structure. This survey provides a systematic review of CBO through a unified BO-loop perspective, showing how causal assumptions shape intervention search spaces, surrogate models, acquisition functions, and decision policies. We organize existing methods by graph and system-knowledge assumptions, environment, intervention representation, surrogate architecture, and decision rule, and connect CBO to causal bandits, Bayesian experimental design, safe optimization, policy search, and causal abstraction. We also introduce a reproducibility-oriented benchmark spanning hard- and soft-intervention settings, with standardized GAP and a new trajectory-aware Path-Aware GAP (PA-GAP), evaluating seven CBO methods and a non-causal BO baseline across thirteen datasets, three budgets, and two metrics. Results show that no method dominates uniformly: rankings depend on dataset, budget, metric, and how causal information is used, while strong non-causal baselines remain competitive in several settings. Controlled graph-misspecification and omitted-variable stress tests further show that rankings can change substantially when learner-side causal information is perturbed. We conclude by identifying key open challenges, including robustness to causal-assumption violations, scalable unknown-graph optimization, mixed intervention types, realistic cost models, stronger theoretical guarantees, and integration with modern representation learning and causal abstractions.

Test-time adaptation (TTA) addresses distribution shift using only unlabeled test data. Existing methods typically adapt pretrained models by updating their parameters, limiting both what is adapted and where adaptation can occur within the network. We instead keep the pretrained network frozen and steer its intermediate representations. We introduce SPeaR (Steering Primitive for Realigning Representations), which inserts lightweight learnable modules at stage boundaries and optimizes them directly from the test stream, requiring neither source data nor supervised warm-up. Each primitive is optimized using a gated objective that reduces uncertainty only when adaptation is beneficial, along with a diversity regularizer to prevent collapse, and a multi-depth anchor to stabilize adaptation. We show that steering early representations is the most effective strategy, and that the same primitive transfers across convolutional and Transformer architectures. Across CIFAR-10-C, CIFAR-100-C, and ImageNet-C, SPeaR consistently matches or outperforms methods that adapt orders of magnitude more parameters, remains robust across a wide range of batch sizes, and preserves source-domain performance during continual adaptation.

Analysts often need to identify images that jointly satisfy multiple visual conditions, such as a crossroads with traffic lights at dusk, for model diagnosis, dataset curation, and targeted training. Embedding-based retrieval can rank the large gallery efficiently, but a visually dominant condition can obscure weaker conditions, and a single similarity score does not enforce the required conjunction. Visual question answering (VQA) can explicitly verify conditions, yet exhaustively applying it to the full gallery is costly, especially when analysts refine their query. These limitations motivate keeping humans in the loop at the attribute level, where analysts can quickly build evidence for each condition and reuse it when the request changes. We therefore present ProbeScout, a visual analytics system that supports this loop. It first builds composable attribute probes from sparse VQA labels and fuses them into a conjunction-aware initial ranking. Coordinated views support rapid screening, near-miss diagnosis, and on-the-fly subset construction by filtering and combining these probe outputs. Analysts provide lightweight attribute- and query-level feedback, which drives staged refinement of fusion weights while keeping the probes fixed. These verified attributes can be reused for future queries. We evaluate ProbeScout on 17 retrieval tasks across three datasets, showing improved retrieval over embedding baselines. A separate 10-task comparison achieves higher task-macro AP and F1 than exhaustive VQA while labeling at most 2% of the gallery images. Two case studies further demonstrate how ProbeScout supports interactive analysis and refinement in realistic workflows.

Zero-shot brain-to-image retrieval requires robust alignment between noisy EEG responses and visual representations. Existing EEG-vision alignment methods often operate in sensor space and apply fixed visual supervision to all responses, ignoring both spatial mixing in scalp EEG and response-wise variability in alignment reliability. We propose an adaptive cortically constrained EEG-vision alignment method for zero-shot brain-to-image retrieval. The method reconstructs EEG responses into predefined ROI-level source-pattern representations and encodes them with a Neuro-ROI Attention Encoder. To handle response-wise variability, we introduce an evidence-based adaptive visual supervision strategy that weights detail-controlled visual targets using model-based alignment evidence. On THINGS-EEG, the proposed method achieves strong 200-way zero-shot retrieval performance, with ROI-level attribution providing post hoc interpretability of the learned source-pattern representations. These results show that cortically constrained representation learning and adaptive supervision can jointly support EEG-vision alignment for zero-shot brain-to-image retrieval.

Hyperdimensional computing (HDC) is a computing framework that represents information as high-dimensional pseudorandom vectors called hypervectors (HVs), enabling learning and inference through simple algebraic operations. The HV dimensionality provides computational capacity and error robustness, but the required high-dimensional randomness must be stored or regenerated, imposing a memory–computation tradeoff. Here, we address this tradeoff by physically embodying the randomness required for HV generation in the static disorder of a scattering medium. Specifically, a silicon photonic circuit combined with the scattering medium generates high-dimensional HVs from a small number of input values in a single optical shot. Detector-side encoding programs HV correlations for continuous and categorical representations. The resulting HVs reproduce key statistical and compositional properties of ideal i.i.d.\ random HVs, including pairwise similarity statistics, associative-memory capacity, and factorization capacity. We further demonstrate an optically assisted inference path using the generated HVs and optical similarity evaluation.

Wearable health questions often combine data retrieval, longitudinal analysis, and health advice over structured records. Prompting a single large language model with a complete record and a composite query obscures whether every request is executed and which evidence supports the answer. We propose a task-oriented multi-agent framework that represents a composite query as distinct intents and typed tasks with explicit intra-intent dependencies. Specialized agents execute retrieval, analysis, and advice tasks; isolated intent states preserve request boundaries and evidence relationships before aggregation. We evaluate the framework on a synthetic dataset of $10{,}000$ virtual users with one month of longitudinal wearable records, covering structured data retrieval, multi-intent recognition, and overall response quality. Across $1{,}500$ retrieval questions, the Query Agent achieves $98.3\%$ accuracy, compared with $97.9\%$ for the Direct LLM baseline, while reducing average query-stage token consumption from $6{,}869$ to $3{,}136$. On $180$ multi-intent questions, the Manager Agent achieves $100.0\%$ Multi-Intent Coverage and $94.4\%$ Multiset Jaccard Similarity. Under the current synthetic evaluation setting, our method receives higher mean Trustworthiness and Transparency scores on both question categories, whereas Actionability does not improve consistently. These results provide preliminary evidence that explicit task organization can support task-relevant data access and data-grounded longitudinal analysis, while leaving health advice generation and validation on real wearable data as open challenges.

Questions scraped from the web are used across academia and industry as a proxy for what people want to know. Across QA training data, retrieval benchmarks, and content strategy, questions on a page are assumed to reflect human intent. We test this assumption at scale by extracting 13.4B question occurrences across 110 FineWeb snapshots (2013-2025), and report three findings. First, you can tell who is asking: provenance (the host/page of questions) leaves a signal in question form, and a logistic model can separate genuine user questions from templated/manufactured ones at AUC 0.725 via length and surrounding context rather than question type, though only 0.554 against commerce FAQ writing. Second, question frequency does not measure demand: the most-frequent questions are boilerplate/templated (over 70% of the top thousand), so occurrence counts measure how often a string was published and not how often it was asked. Third, over twelve years the genuine share of occurrences fell by 79% (42-56% after controlling for crawl composition), with question length and context decreasing. We present the first diachronic, occurrence-level measurement of web question provenance, and find the crawlable web's questions have shifted from being asked by humans toward manufactured for machines to read.

Genetic Programming Guided Local Search (GPGLS) uses genetic programming to evolve utility functions for guided local search in large-scale vehicle routing problems (LSVRPs). Evaluating every GP individual on every training instance at every generation is expensive, so GPGLS is usually trained on small instance batches. Existing curriculum-based GPGLS orders these batches mainly by instance size. Adaptive Curriculum Learning GPGLS (ACL-GPGLS) improves training efficiency by adapting when the search moves between fixed curriculum stages, but the instance difficulty order remains predefined. We propose DCL-GPGLS, which estimates the difficulty of each training instance from the current population's solution quality and updates the estimates during evolution. Each generation then receives a batch near a scheduled difficulty level, with a correction that limits repeated selection of the same instances. Experiments on a fixed training-test split of the CVRPLIB X set show that DCL-GPGLS achieves the best observed average rank and mean test cost among six training policies. It obtains the lowest mean cost on 36 of 65 unseen test instances and is significantly better than the static feedback-derived curriculum, matched in total evaluator calls, on 6 instances, with no significant difference on the remaining 59.

Genetic Programming Guided Local Search (GPGLS) learns utility functions that guide local search for vehicle routing. Its evolving programs can have similar fitness while inducing different search behaviour, making fitness alone an incomplete basis for population diversity management. We propose GPGLS with Behaviour-based Niching (BN-GPGLS), which characterises programs through six operator-level descriptors collected during local search. A current-generation archive selects fitness-competitive, compact representatives from strata of a behaviour score. Fixed policies use archive parents continuously, whereas adaptive policies activate them using training-fitness and standardised behaviour-dispersion signals, optionally with a tree-size condition. We compare four behaviour-based variants with a no-archive GPGLS control and fitness-based niching over 30 seed-matched runs on generated 200-customer instances. BN-Adaptive achieves the best descriptive average rank on a separate 90-instance monitoring set; aggregate routing-cost differences are small. All five archive policies produce lower final-population median tree sizes than the GPGLS control, with paired Wilcoxon comparisons remaining significant after Holm adjustment. These results identify useful solution-quality and program-size trade-offs within the evaluated setting, without attributing the size reductions to behaviour representation alone.

In many real-world planning tasks, agents must tackle uncertainty about the environment's state and variability in the outcomes of any chosen policy. We address both forms of uncertainty as a first step toward safer algorithms in partially observable settings. Specifically, we extend Distributional Reinforcement Learning (DistRL)-which models the entire return distribution for fully observable domains-to Partially Observable Markov Decision Processes (POMDPs), allowing an agent to learn the distribution of returns for each conditional plan. Concretely, we introduce new distributional Bellman operators for partial observability and prove their convergence under the supremum p-Wasserstein metric. We also propose a finite representation of these return distributions via psi-vectors, generalizing the classical alpha-vectors in POMDP solvers. Building on this, we develop Distributional Point-Based Value Iteration (DPBVI), which integrates psi-vectors into a standard point-based backup procedure-bridging DistRL and POMDP planning. By tracking return distributions, DPBVI lays the foundation for future risk-sensitive control in domains where rare, high-impact events must be carefully managed. We provide source code to foster further research in robust decision-making under partial observability.

Sixth-generation (6G) wireless systems are envisioned as networks of networks, integrating diverse in-X subnetworks that provide localized, high-performance connectivity. Ensuring reliable communication in dense deployments, such as industrial robots and vehicles, is challenging due to dynamic interference and strict performance requirements. Traditional radio resource management (RRM) methods have limitations, prompting the need for AI-based solutions. In this paper, we address the challenges of dynamic resource allocation and scheduling in 6G in-X subnetworks supporting applications with heterogeneous characteristics by proposing a novel framework that combines Multi-Agent Reinforcement Learning (MARL), Federated Learning (FL), and Explainable AI (XAI). Our solution is designed to improve the reliability, robustness, and transparency of resource management and intra-subnetwork scheduling while ensuring data privacy and fairness across multiple co-existing subnetworks. Unlike existing works, our approach considers a realistic scenario with multiple devices per subnetwork, thereby offering a more comprehensive and scalable solution. The proposed explainable RL framework enables agents to collaboratively optimize channel allocation and scheduling without the need to share raw data, preserving the privacy of each participating subnetwork. Extensive simulations based on 3GPP scenarios demonstrate the effectiveness of our approach, showing significant improvements in performance, and transparency over existing solutions.

Automated biomedical evidence synthesis depends on retrieving published studies, but the biomedical literature is systematically skewed toward positive findings. Deeper retrieval can therefore make a system more likely to falsely infer benefit when the true effect is null. We formalise this phenomenon as evidence drift and prove that, under a standard publication-bias model, the false-positive probability on null-effect queries follows a strictly increasing large-sample envelope in retrieval depth, approaching one. Empirically, on a held-out test set of 140 Cochrane-derived queries, drift rises monotonically from 7.9% to 15.7% as the retrieval budget grows from 3 to 20 steps, and concentrates in the null-effect class. We present DACG-agent, a drift-aware causal-graph agent that incrementally builds a causal knowledge graph from PubMed abstracts and applies a two-layer stopping policy with complementary roles: a KL-divergence monitor that detects posterior convergence (the accuracy layer), and a Bradley–Terry process reward model (PRM) whose online decline detection halts retrieval once evidence quality peaks (the efficiency layer). Against full-budget retrieval, DACG-agent reduces evidence drift from 15.7% to 6.4% and improves null-effect accuracy by 21 percentage points (40.0%$\to$61.4%) while using 67% fewer retrieval steps; overall accuracy rises from 61.4% to 69.3% (95% CI 61–77). A simulation confirms the drift result transfers from the analysed vote-counting aggregator to the deployed noisy-OR one.

Relaxor ferroelectrics retain crystalline order while exhibiting glass-like dielectric and thermal responses. Using atomistic models trained on first-principles data, we investigate these responses in the prototypical relaxor Pb(Mg_{1/3}Nb_{2/3})O_3 (PMN). Compositional disorder couples elastic motion to electric polarization, producing a continuum of polar vibrational excitations, unlike the discrete polar optical modes of conventional ferroelectric PbTiO_3. The simulations qualitatively reproduce the temperature dependence of the low-frequency susceptibility, while the calculated far-infrared spectra agree closely with experiments. At low frequencies, disorder in PMN generates a boson peak comprising excess phonon-like modes and quasi-localized excitations. The latter exhibit a quartic frequency dependence of their density of states and pronounced anharmonicity, as in structural glasses. These excess modes account for PMN's anomalous behavior in the specific heat for temperature higher than 5 K. These results connect the dielectric, optical, and thermal anomalies of a crystalline relaxor through its disorder-induced polar excitation spectrum.

Body-cue recognition can support assistive robots, but benchmark accuracy does not guarantee reliable behavior under a robot-camera viewpoint. We present Nuni, a bedside robot prototype that treats a detected distress cue as a reason to ask rather than a reason to alert. We compare two X3D-UGT RGB appearance classifiers, which reach 97.7% and 94.8% six-way accuracy on NTU RGB+D, with a pose-centric hybrid pipeline on 28 single-actor scripted clips recorded from the robot camera. The hybrid path achieved 0.71 six-way macro recall, versus 0.25 and 0.29 for the fine-tuned and from-scratch RGB variants. More importantly for interaction, it produced a question-triggering distress cue in 12/16 distress clips and would have prompted unnecessarily in 2/8 normal clips; the RGB variants yielded a question-triggering cue in only 2/16 and 3/16 distress clips. We separately tested the question-first controller through event injection. All 13 state-transition trials passed: valid responses caused stand-down, two unanswered prompts produced one alert, and three boundary conditions were handled correctly. These results are a preliminary technical evaluation, not a user study or medical validation, but they show how interaction policy can limit the consequences of uncertain perception.

Visual Question Answering (VQA) with Multimodal Large Language Models (MLLMs) requires not only producing safe and effective responses, but also grounding safety decisions in the multimodal evidence that determines risk. Recent safety-alignment methods improve refusal behavior and contextual risk awareness, yet correct safety outcomes may still rely on superficial textual or visual correlations, particularly when risk emerges from interactions between individually benign image and question content. To address this issue, we propose A$^2$Safe, a counterfactual evidence-aligned adaptive agent collaboration framework for safe and effective VQA. A$^2$Safe organizes localized visual observations, textual intent, and cross-modal risk relations through a Grounded Safety Evidence Board, making the basis of safety decisions explicit. Counterfactual safety evidence alignment enforces invariance to safety-irrelevant changes while requiring appropriate safety-state and response-mode transitions when risk-critical evidence is minimally altered. The resulting evidence state further supports adaptive collaboration, enabling direct answering when grounded evidence is sufficient and invoking policy critique and response revision when evidence is risky, uncertain, or conflicting. Under complementary safety-critical and general VQA protocols, A$^2$Safe achieves a 95.72 SIUO safety score, reduces the benign refusal rate on MOSSBench to 14.67%, and maintains an average general VQA score of 78.34 with 27.8% token overhead. These results support counterfactual evidence-aligned adaptive collaboration for safe and effective multimodal question answering.

We prove that every triangle-free graph with minimum degree at least $\frac{n}{3}$ is $4$-colorable and thereby settle a problem of Brandt and Thomassé (2005) at the threshold $\frac{n}{3}$. The number four is best possible. For a positive integer-valued function $f(n)=o(n)$, we relate the chromatic number of $f(n)$-vertex subgraphs of the Kneser graph $KG(n,f(n))$ to that of triangle-free graphs with minimum degree at least $\frac{n}{3}-f(n)$. Consequently, for every $0<\delta<1$ and $\varepsilon>0$, and for all sufficiently large $n$, every $n$-vertex triangle-free graph with minimum degree at least $\frac{n}{3}-n^{1-\delta}$ has chromatic number at most $10^{391}+1+\left\lceil{(1+\varepsilon)(1-\delta)}/{\delta}\right\rceil$. We also show that every sufficiently large $n$-vertex maximal triangle-free graph with minimum degree at least $\frac{n}{3}-f(n)$ and chromatic number at least $10^{391}$ contains a bipartite subgraph with parts of orders $\frac{n}{3}-O(f(n))$ and $\frac{2n}{3}-O(f(n))$; the remaining induced subgraph admits a homomorphism to $KG(\frac{n}{3}-O(f(n)),O(f(n)))$. Finally, we connect maximal $K_r$-free graphs with minimum degree at least $\frac{2r-5}{2r-3}n-f(n)$ to $K_{r-1}$-free graphs and extend these results to $K_r$-free graphs. Our proofs employ the recent strong Brandt–Thomassé theorem of Łuczak, Polcyn, and Reiher.

Motivated by results in Kähler geometry, in this work, we are interested in understanding the relation between integral curvature bounds and volume growth, under non-negative curvature in dimension three. In case of non-negative sectional curvature, we show that for metric on Euclidean space, it is of Euclidean volume growth if and only if it has average quadratic curvature decay. This is based on showing that metrics on three-dimensional Euclidean space with non-negative sectional curvature is of Euclidean volume growth if its asymptotic scaling invariant integral of scalar curvature is smaller than the sharp constant $8\pi$. We also show a gap Theorem if the curvature decay fast enough in the average sense, under non-negative Ricci curvature.

While deep learning has enabled language decoding from intracranial brain recordings, extending this capability to non-invasive recordings remains an unresolved challenge. Decoding individual words from non-invasive brain recordings is particularly difficult, as word-level neural evidence is weak, temporally distributed, and entangled with acoustic, lexical, and semantic structure. Existing retrieval pipelines often collapse these factors into a single representation, potentially discarding information available at intermediate temporal scales. Here, we introduce Hierarchical Dynamic Neural Decoding (HDND), a hierarchical dynamic decoding framework that treats word decoding as structured refinement rather than flat label retrieval. HDND combines intermediate neural representations, contextual semantic predictions, and, for selected reading conditions, an auxiliary character-form objective. We evaluate HDND across seven electroencephalography (EEG) and magnetoencephalography (MEG) datasets spanning English, Dutch, Mandarin, and Cantonese listening, reading, and reading-aloud conditions. Across the nine-condition word-retrieval benchmark, the proposed HDND yields a higher participant-averaged balanced Top-10 point estimate than the matched contextual word-decoding baseline in every condition and achieves the highest mean among all compared methods in eight of nine conditions. Across the same nine matched conditions, HDND also yields higher token-micro and pooled word-macro Top-10 point estimates in every setting. Sentence retrieval favors HDND in eight of nine conditions, while auditory speech-segment retrieval is mixed across the six listening conditions. These results show that hierarchical residual refinement can improve multilingual word retrieval from heterogeneous non-invasive brain recordings.

Visual analytics (VA) enables sensemaking through interactive visualization, but effective analysis often requires experts to translate high-level intents into long sequences of interface operations and iteratively interpret visual feedback. We study whether modern vision-language models (VLMs) can take on this role as autonomous VA operators that observe the interface, plan multi-step exploration, execute interactions, and adapt based on intermediate visual feedback. To support systematic development and evaluation, we first introduce WidgetVA, a widget-centric agentic VA framework that standardizes interactive components as structured widgets with unified action (e.g., filter and zoom) and perception-query (e.g., selection summaries) APIs. This standardization supports two modes of system construction: wrapping an existing VA system to make it agent-operable without rebuilding it, and composing a new system from widgets as modular building blocks. To help agents coordinate across widgets rather than plan each interaction from scratch, each widget further packages reusable analytical workflows, giving agents more than a bare set of callable functions to plan over. Building on this framework, we present WidgetVABench, a benchmark of single- and multi-widget VA tasks that require agents to perform multi-step interactions to uncover evidence and produce verifiable results. Each task also provides fine-grained reference annotations so that WidgetVABench can score Answer, Reference Trace Similarity, and State separately rather than collapsing agent performance into one success score. Experiments across multiple VLMs show that our framework provides an effective scaffold for agentic VA, while the diagnostic measures expose persistent limitations for future work. The WidgetVA framework and WidgetVABench have been released in https://github.com/Hiverwin/widgetva.

Learning dexterous manipulation from demonstrations is bottlenecked by data: the contact forces that determine whether a grasp succeeds are absent from every scalable source of human demonstrations. This paper builds on two observations. First, what survives the change from a human hand to a robot hand is the contact structure of a demonstration - which finger regions touch which object locations, and in what order - rather than its joint motion. Second, physical consistency need not be engineered per task: a single residual reinforcement learning (RL) policy, trained once across diverse demonstrations, can repair kinematic recordings into physically consistent, contact-annotated trajectories, and the same residual formulation restores dynamic feasibility after retargeting. These observations yield a three-stage pipeline that converts human motion-capture recordings into dexterous robot policies with no real-robot training data: physics refinement with a simulated MANO hand recovers contacts and forces, contact-anchored retargeting transfers the demonstrated contact structure through an objective independent of hand morphology, and residual policy learning adapts the result to robot actuation. The pipeline reconstructs 25,454 single-hand trajectories (success 7.3% -> 59.3%) and 25 dual-hand tasks (16.0% -> 62.4%) with one shared policy per setting, transfers one human dataset to four morphologically distinct robot hands (+62.4 pp), and executes four contact-rich bimanual tasks on physical hardware with zero real-robot training data.