arXiv:2607.17012v1 Announce Type: new
Abstract: Large language models (LLMs) can plan tool-mediated scientific work, but scientific instruments remain difficult to connect to such agents: vendor APIs may load only inside acquisition host processes, facilities may prohibit cloud-hosted agents, and natural-language interfaces can emit physically unreasonable arguments. We present a method for connecting local LLMs to scientific instruments through the Model Context Protocol (MCP). It combines: (1) a schema-bound tool surface that validates requests against physical bounds before adapter dispatch; (2) a vendor-neutral, host-process adapter pattern separating language-side reasoning from instrument-side execution; (3) a persistent lifecycle for long-running live-processing jobs; and (4) MCP-prompt-registered skills that compose typed tools into reusable multi-step protocols. Our open-source reference server exposes 30 typed tools, 5 live-job types, and 6 skills through a physics-plausible simulator implementing the same protocol surface. Validation is software-only: all 120 hardware-independent tests pass deterministically, while 15 local-LLM integration tests pass 12-15 of 15 across runs because of model nondeterminism. A preliminary single-run probe across five open tool-calling LLMs indicates that the schema-bound interface can be driven locally by small open-weight models without cloud dependency; it is not a benchmark and has no confidence intervals. The method provides a bounded, testable, and reproducible foundation for closed-loop agentic instrumentation research.
Science Journals
arXiv:2607.17657v1 Announce Type: new
Abstract: Multimodal large language models (MLLMs) still struggle with spatial reasoning that requires perspective transformation. In particular, they often rely on camera-centric cues rather than reasoning from the reference object's viewpoint, leading to systematic errors in non-camera reference settings. In this paper, we first analyze this failure mode and show that object orientation is a key factor underlying such camera-centric shortcut behavior. To address this issue, we propose OrientSAM, an orientation-aware spatial alignment framework for multimodal models. OrientSAM injects explicit orientation information into multimodal representations through orientation-aware tokens and Fourier-based angle encoding, and further adopts a curriculum learning strategy to progressively improve perspective-aware reasoning. In addition, we build a spatial data construction pipeline to generate orientation-aware spatial supervision from large-scale images. Experiments on Spatial-MM, ViewSpatial, and 3DSRBench show that OrientSAM consistently outperforms strong baselines, especially on non-camera-view, person-centric, and orientation-sensitive tasks. The results further demonstrate that explicit orientation modeling is important for mitigating camera-centric shortcut behavior and enabling more robust allocentric spatial reasoning in multimodal models.
arXiv:2607.16298v1 Announce Type: new
Abstract: Understanding whether visual recognition models rely on shape, texture, or color is central to interpreting their behavior. Prior cue-conflict studies have strongly influenced the view that CNNs are texture-biased, yet such tests measure cue preference under artificial conflicts rather than feature reliance during natural recognition. We revisit this question through controlled feature suppression and show that performance drops are difficult to interpret unless different suppression operations impose comparable category-level damage. We introduce a semantically matched evaluation framework that compares shape and texture suppression at matched levels of category separability loss. Under this framework, ImageNet-trained CNNs show stronger degradation under texture suppression than under shape suppression, revealing greater texture reliance than suggested by unmatched suppression analyses. Extending the comparison across architectures, we find that Vision Transformers retain higher accuracy than CNNs under both shape and texture suppression. Brain encoding further shows that ViT representations exhibit smaller suppression-induced decreases in neural prediction performance under the tested suppression settings. These findings indicate that semantic comparability is essential for interpreting feature reliance from suppression experiments, and suggest that the robustness advantage of ViTs may be related to representations more compatible with human visual cortex.
arXiv:2607.16943v1 Announce Type: new
Abstract: Safety validation at signalized intersections remains a critical bottleneck for the deployment of autonomous driving systems (ADS), as these scenarios involve dense heterogeneous traffic, contested right of way, and long-tail safety-critical interactions, posing significant challenges to the Safety of the Intended Functionality (SOTIF). Existing naturalistic driving datasets often suffer from geographical homogeneity, sparsity of safety-critical events, and lack of semantic risk annotations, which limit the evaluation of algorithmic generalizability and targeted SOTIF verification. To address these gaps, this paper introduces SinD 2.0, a large-scale drone-based intersection dataset dedicated to cross-domain ADS safety analysis. The main contributions of SinD 2.0 are: (1) Cross-domain diversity: It covers six signalized intersections across four Chinese cities, capturing distinct intersection topologies and regional driving behavior characteristics; (2) High-density risk interactions: A total of 32,682 safety-critical events are extracted via surrogate safety measures, significantly enriching the density of boundary test scenarios; (3) Hierarchical semantic annotations: Besides integration with high-definition (HD) maps and Signal Phase and Timing (SPaT) data, it provides multi-dimensional semantic labels including traffic violations, high-risk interactions, visual shielding, and narrow feasible areas; (4) Full-stack testing toolchain: It supports automated scenario extraction, prediction-only evaluation, open-loop replay, reactive closed-loop testing, and photorealistic rendering. Benchmark experiments demonstrate that SinD 2.0 exhibits significant domain shifts across cities, and the semantic risk subsets can effectively expose the performance limitations of ADS algorithms. The dataset, annotations, and testing toolchain are available at https://github.com/SOTIF-AVLab/SinD/tree/main.
arXiv:2607.17269v1 Announce Type: new
Abstract: Large language models encode world models implicitly in neural weights, which exposes four structural risks in high-precision domains such as medicine and finance: hallucination, frozen knowledge, poor explainability, and poor modifiability. This paper proposes data-first ontology: LLMs are treated as reasoning and language engines, while deterministic knowledge is moved into an explicit multimodal database, DaoQL. We formalize an explicit world model and show that, under rule independence, deterministic evaluation, and fixed conflict resolution, explicit models provide a sufficient condition for composable counterfactual decomposability; implicit models lack atomic read/delta semantics and therefore provide no comparable architectural guarantee. The implemented system focuses on DaoQL's verified storage layer and explicit Eval path, integrating graph, column, vector, and full-text engines within one process. KVCache graph nodes, expert hot updates, and the DaoQL-Agent runtime remain future work. On an embedded same-machine setup, DaoQL reports graph BFS at 1.20 ms, HNSW at 83.1 us, and a Fluent hybrid query at 105.8 us; these results indicate engineering potential but must be interpreted with deployment-shape differences from client-server systems. Exploratory measurements on LDBC SNB SF1 and ANN-Benchmarks further show 34/34 query coverage with interactive-class queries mostly in the sub-millisecond to millisecond range, but only 1.8 QPS overall due to long-tail BI/IC queries; ANN-Benchmarks reaches Recall@10 >= 99% at thousand-level QPS after a bridge-edge protection fix. In a five-domain counterfactual experiment (n = 1250), DaoQL+GPT-4o achieves 94% composable counterfactual decomposability, 49 percentage points above GPT-4o alone. The paper explicitly separates provable structure, preliminary empirical evidence, and architectural roadmap claims.
arXiv:2607.16299v1 Announce Type: new
Abstract: In competitive learning activities, inaccurate robot decisions may reduce students' perceptions of fairness and competence, ultimately affecting their motivation. This paper investigates whether the accuracy of sound detection algorithms influences student motivation during a robot-mediated quiz game. A Pepper humanoid robot hosted an interactive buzzer-based quiz in which two sound detection approaches, a Convolutional Neural Network (CNN) and a Cross-Correlation algorithm, were evaluated using a controlled between-subjects experiment involving 40 university students. Participants were equally assigned to a CNN group (n = 20) and a Cross-Correlation group (n = 20). Both groups completed the same quiz under identical conditions, differing only in the sound detection algorithm used for first-responder identification. Student motivation was assessed using the Intrinsic Motivation Inventory (IMI), while algorithm performance was evaluated through real-time detection accuracy. The results indicate that the Cross-Correlation approach achieved more reliable sound detection under classroom conditions and produced significantly higher scores across all IMI subscales, demonstrating greater student interest, perceived competence, effort, perceived choice, and lower perceived pressure (after reverse coding). These findings provide empirical support for the proposed Algorithmic Precision-Motivation Relationship (APMR) model, demonstrating that algorithmic accuracy is not merely an engineering performance metric but an important factor influencing learner motivation in robot-assisted educational environments.
arXiv:2607.17438v1 Announce Type: new
Abstract: Laser powder bed fusion (LPBF) is a metal additive manufacturing process where temperature stabilization is of vital importance to avoid defects such as distortion and cracking. Existing control methods require manual tuning, increasing the risk of part failure when printing complex geometries. This paper introduces a dual-loop, data-driven control strategy to stabilize the surface temperature, ensuring robustness and near-optimal performance in the presence of disturbances. The proposed method integrates (i) an in-layer linear output feedback control with gains optimized through policy gradient, and (ii) a layer-to-layer feedforward control combining temperature trajectory optimization and iterative learning control. Simulation results show that the multi-scale controller effectively stabilizes the temperature even under significant model mismatch and measurement noise. Experimental results demonstrate that a simplified, hardware-constrained version of this method matches the state-of-the-art performance of in-situ data-driven methods, reducing mean tracking error by 3.4% and mean input-constraint violation by 47.5% relative to a Bayesian Optimization-tuned baseline. For this physical LPBF validation, the controller is tuned entirely offline using uncontrolled print data from a single calibration layer. Our experiments also demonstrate a new class of high-frequency excitation dynamics that result in reduced vector head swelling, opening up new avenues of research in the additive manufacturing community. This work marks one of the first successful applications of sim-to-real policy optimization in LPBF processes.
arXiv:2607.17279v1 Announce Type: new
Abstract: Recently, text-to-video (T2V) models have been widely deployed, sparking growing concerns over their robustness against jailbreak attacks. Existing jailbreak methods, mostly adapted from text-to-image attacks, suffer notable drawbacks when applied to T2V systems. They fail to fully leverage temporal consistency, an inherent characteristic of video generation. Besides, these methods demand heavy video query optimization, which is infeasible in practical black-box scenarios. Their adversarial prompt search is also driven by heuristic local signals, lacking principled structured exploration strategies. To tackle these limitations, we propose BSB, a structured, query-efficient jailbreak framework for T2V models. BSB harnesses temporal consistency by encoding harmful intent as the transition between two individually harmless boundary states. Under this paradigm, the attack targets boundary-state pairs whose interpolation tends to produce unsafe intermediate frames during video generation. Directly evaluating all candidate pairs within the video space incurs prohibitive computation cost. Instead, BSB conducts Monte Carlo Tree Search (MCTS) in a cheaper textual proxy space and regularly calibrates search outcomes with sparse video-level evaluations. We conduct comprehensive experiments on mainstream commercial T2V models including Veo 3.1, Sora 2, Seedance and Kling v1. Results show BSB surpasses all existing jailbreak baselines, delivering an average 18.6% relative gain in attack success rate over the strongest competitor across evaluated models. Our findings identify temporal consistency as an understudied yet vital attack surface for T2V models and verify that structured search facilitates effective vulnerability discovery under constrained query budgets.
arXiv:2607.16302v1 Announce Type: new
Abstract: In recent years, motorcycle popularity has been declining, particularly among the younger generations. To rekindle interest in motorcycles among this demographic, we developed a rideable two-wheeled robot equipped with four limbs as a future partner mobility concept and intended for use in public events. The character reproduced by this robot carries the protagonist on its back and exhibits a dynamic quadrupedal gait. To maximize the riding experience, we aimed to match the robot's weight and size to the character's specifications while ensuring rider safety and enabling expressive movements of limbs, wrists, ankles, and facial features. However, achieving locomotion solely through limb movement would require excessive motor output and increased limb strength, resulting in higher weight and extremely slow gait, thereby reducing character fidelity. To overcome these challenges, the robot performs its primary locomotion using a self-balancing two-wheeled base, while the limbs provide auxiliary support during mounting and dismounting and move in coordination with the wheeled locomotion speed. This approach enables an animatronic robot capable of intuitive, weight-shift-based control for free and natural movement. In this paper, we focus on a robust selfbalancing control method that maintains stability even during rapid limb movements, as well as a motion control strategy that generates natural quadruped-like behavior.
arXiv:2607.16305v1 Announce Type: new
Abstract: Vision-Language Models (VLMs) have achieved strong progress in multimodal understanding. However, scaling dense or sparse Mixture-of-Experts (MoE) models to improve performance limits deployment in resource-constrained environments due to the trade-off between high memory usage from full loading and increased latency from on-demand loading. Recently, the Per-Layer Embedding (PLE) architecture addresses this by scaling models with large external embedding tables stored in ROM and performing lightweight lookup to retrieve relevant embeddings to enhance token representations. Nevertheless, existing PLE-style methods are primarily designed for text embeddings due to the convenience of ID-based retrieval, limiting their effectiveness in VLMs where multimodal embeddings contain richer information for visual tasks. In this paper, we propose LookME, the first framework that enables lookup-based enhancement for multimodal embeddings in VLMs while supporting partitioned storage and on-demand loading. To efficiently lookup arbitrary continuous multimodal embeddings from large-scale embedding tables, we propose a hierarchical two-level lookup method employing a coarse-to-fine strategy that performs lookups from the scene-level to the intra-scene primitive-level. Furthermore, we integrate the lookup method with a sparse injection strategy, which adaptively prioritizes critical embeddings over voluminous multimodal embeddings within layers, and facilitates embedding table reuse across neighboring layers, improving the trade-off among efficiency, model size, and performance. Experiments on multiple visual benchmarks show that LookME outperforms text-only PLE-style methods, validating the effectiveness of lookup-based multimodal embedding enhancement.
arXiv:2607.17660v1 Announce Type: new
Abstract: Camera-only 3D semantic occupancy prediction aims to infer voxel-wise scene semantics from multi-view images, yet remains fundamentally challenging due to depth ambiguity and occlusion. Along a single camera ray, multiple spatially separated surfaces may coexist, making occupancy inherently a multi-label existence problem rather than a single-depth estimation task. However, most existing approaches favor a single dominant depth hypothesis per ray, limiting their ability to model volumetric scenes under complex occlusion. To address this limitation, we introduce RayOcc, an occlusion-aware ray occupancy framework that reformulates ray modeling as multi-label existence prediction. Instead of predicting a categorical depth distribution, RayOcc estimates a non-normalized Gaussian mixture intensity along each ray and converts it into interval-wise occupancy probabilities via a Poisson event formulation, allowing multiple occupied hypotheses to coexist without enforcing mutual competition across depth. The predicted mixture components are interpreted as occupancy hypotheses to initialize sparse 3D Gaussian primitives, which are refined and rasterized for semantic occupancy prediction. Experiments on the nuScenes benchmark show that RayOcc achieves state-of-the-art overall IoU and mIoU among the compared Gaussian-based occupancy methods.
arXiv:2607.17665v1 Announce Type: new
Abstract: Mixed Reality (MR) is increasingly being used in exhibition settings to bring digital artworks into relation with the physical environment. However, existing MR exhibition systems are often confined to prototypes or case-specific deployments, offering limited guidance for large-scale practical implementation. To address this gap, this paper presents a practical pipeline for designing and deploying large-scale MR art exhibitions, treating spatial alignment not only as a technical mechanism but also as an experiential design decision. We first conducted a pilot study comparing marker-based and Simultaneous Localization and Mapping (SLAM)-based alignment methods in an MR exhibition setting. Based on the results, we developed a SLAM-based pipeline for MR exhibitions that integrates technical deployment with exhibition curation. We then evaluated the pipeline through both system overhead measures and users' experiential feedback. The results show that spatial alignment influences not only technical stability, but also overall exhibition coherence, visitors' sense of continuity and immersion, and artwork interpretation. These findings provide an empirically grounded reference for future large-scale MR art exhibition deployment.
arXiv:2607.16777v1 Announce Type: new
Abstract: We present JOR-Bench, a collection of five Japanese-language benchmarks for evaluating the ability of large language models (LLMs) to formulate and solve operations research (OR) problems. Each benchmark is a Japanese translation of an existing English benchmark: IndustryOR, MAMO Complex LP, NL4OPT, OptiBench, and OptMATH, covering 1,319 problems spanning linear programming, mixed-integer programming, non-linear programming, and combinatorial optimization. JOR-Bench is a solver-independent benchmark that can be used with any solver or programming language, and consists of pairs of Japanese problem statements and expected numerical answers. We evaluate seven LLMs, including multilingual general-purpose models and Japanese-specialized models, on both the original English and the new Japanese versions, and compare performance across languages. For the main evaluation, we standardize execution with the Python interface to OR-Tools to make model outputs comparable and reproducible with open-source software. Our results show that OR formulation ability is largely language-neutral for strong multilingual models; the overall average accuracy difference between English and Japanese is only $-0.3$ pp. Yet error analysis reveals subtle cross-lingual differences, including a pragmatic disambiguation failure in some domains that causes models to output decision-variable values instead of the objective value when the prompt is in Japanese.
arXiv:2607.16307v1 Announce Type: new
Abstract: Diffusion-based text-to-image models often fail on complex prompts involving multiple entities, attributes, and relations, producing object omissions, incorrect attribute assignments, or reversed spatial layouts. Existing training-free methods mainly strengthen token-level attention, but do not explicitly model which attributes belong to which entities or when different constraints should be enforced during denoising.
We introduce \textbf{CoBind}, a training-free framework for stage-aware compositional binding. CoBind parses a prompt into a composition graph of entities, attributes, and relations. It first establishes the global layout using entity-completeness and relation constraints, then binds attributes to their target entities through contrastive cross-entity optimization. Structural guidance is gradually relaxed in later denoising steps to preserve textures and visual details. CoBind also adapts the guidance strength according to the current satisfaction of each constraint, reducing unnecessary latent updates.
CoBind requires no retraining or additional annotations. Experiments on T2I-CompBench++, GenEval, and multiple diffusion backbones show consistent improvements in attribute binding, spatial relations, and complex compositional generation while maintaining competitive visual quality.
arXiv:2607.16245v1 Announce Type: new
Abstract: Gate-level netlists exhibit intrinsic structural properties that influence signal propagation independently of functional simulation. We define a topology-driven structural manipulability score that characterizes node-level structural flexibility using path participation, k-core embedding, symmetry, and centrality. Modeling netlists as directed graphs, we formulate node-level regression to learn this topology-derived score using graph neural networks (GNNs). Experiments on ISCAS85 and EPFL benchmarks evaluate how effectively different GNN architectures approximate this metric across held-out circuits, with hierarchical models yielding the most consistent rankings. Component-level and ablation analyses examine the contribution of individual factors. As an illustrative case study, analysis of Trojan-injected circuits using TrustHub templates reveals statistically distinguishable structural patterns, indicating that topology-based scoring provides complementary structural insight.
arXiv:2607.16766v1 Announce Type: cross
Abstract: Edge inference has emerged as a promising solution for the proliferation of artificial intelligence (AI) services by deploying models at the network edge to circumvent cloud-routing latency. Existing edge inference approaches mainly focused on either cooperative inference to reduce latency or lightweight model design to fit resource-constrained devices. These solutions often address the communication and computation challenges separately, and thus struggle to achieve a balanced trade-off among transmission efficiency, on-device processing cost, and inference accuracy. To bridge this gap, this paper proposes a hybrid-precision task-oriented communication framework for edge inference to holistically balance communication, on-device computation, and utility. In this framework, a binarized front-end is deployed on the edge device to extract and transmit binary features via orthogonal frequency-division multiplexing (OFDM) signals, while a full-precision back-end on the edge server performs the final inference. To ensure model consistency, we introduce an on-device binarization method tailored for split inference and develop an integrated channel-aware transmission scheme featuring subcarrier-based feature calibration. Furthermore, a knowledge distillation (KD)-based training strategy, supported by specialized gradient estimators, is developed to optimize the end-to-end system and inherit semantic knowledge from a full-precision teacher model. Extensive experiments on the large-scale ImageNet dataset demonstrate the superiority of the proposed hybrid system. Our analysis confirms that this design achieves an optimal trade-off among communication efficiency, on-device computational cost, and inference accuracy, outperforming existing edge inference solutions.
arXiv:2607.16308v1 Announce Type: new
Abstract: Cross-domain few-shot semantic segmentation (CD-FSS) has predominantly been formulated as learning domain-invariant representations or improving support-query correspondence. Nevertheless, large domain shifts still make prototype matching unreliable: inconsistent hierarchical responses corrupt the support representation, deterministic prototypes cannot express boundary and appearance ambiguity, and treating prototypes with different reliability equally during optimization weakens foreground-background separation. We therefore propose DAUPNet, a unified framework that reformulates cross-domain prototype matching as uncertainty-aware prototype discrimination. DAUPNet first harmonizes hierarchical support-query features to provide stable evidence, then represents foreground and background prototypes probabilistically, and finally uses their estimated uncertainty to regulate contrastive optimization. On four standard target domains, DAUPNet achieves 72.6% and 76.7% average mIoU in the 1-shot and 5-shot settings, respectively, including substantial gains on the two medical domains. These results demonstrate that modeling prototype uncertainty and incorporating it into optimization provides a robust and interpretable approach to CD-FSS under severe domain shift. The code is available at https://github.com/madness-Lei/DAUPNet
arXiv:2607.17044v1 Announce Type: new
Abstract: Multi-step enterprise agent tasks fail in a characteristic way: single-pass inference has no checkpoint between deciding an answer and committing to it. We study one production system (Leni) whose architecture installs such checkpoints: verification loops (execute, observe, compare, correct) staffed by lightweight task-specialized post-trained models. We evaluate the unmodified production configuration on three public benchmarks stressing distinct failure modes: SpreadsheetBench Verified (silent computation error), BullshitBench v2 (premise confabulation), and the GAIA validation split (cascade error over long tool chains). The full system improves over its frontier base model by +11.0 percentage points on SpreadsheetBench (91.25% vs 80.25%, n=400, p<0.001), +7 to +10 percentage points on BullshitBench (98% vs 91%, n=100), and roughly +15 points on GAIA validation (75.2% pass@1, n=165; 83.0% best-of-k). Our central contribution is a decomposition of that uplift: most of it comes from scaffolding, routing, and specialist models rather than from the verification step itself, whose isolated contribution is small (+1.5 points) but concentrated at the top of the score distribution, where it converts otherwise-failing tasks. We instrument the loop end-to-end, yielding an empirical verifier confusion matrix (catch rate about 0.20, fix rate 0.75, no false-alarm regressions) that grounds a compounding-reliability model. Specialist-swap ablations suggest that the loop's value depends on who observes it: replacing the small trained verifier with the generating frontier model eliminates most rescues. A valid-premise control shows zero over-rejections in 100 expert-level questions.
arXiv:2607.17441v1 Announce Type: new
Abstract: Deepfake generation has raised growing concerns regarding digital media authenticity, misinformation, identity fraud, and public trust. Recent studies show that combining spatial and frequency features leads to stronger detection results than using independently. This paper presents MSCA-FFT, a Fast Fourier Transform (FFT)-based multi-scale cross-attention framework for image-level deepfake detection. The model combines a partially fine-tuned Xception spatial branch with an FFT-based frequency branch. The frequency branch processes the log-scaled FFT magnitude spectrum through shallow convolutional layers, avoiding inverse frequency-to-image reconstruction used in DCT-based pipelines. The spatial and frequency representations are refined by transformer encoders, fused through cross-attention, and passed to an MLP classifier for real/fake prediction. Experimental results show that MSCA-FFT achieves consistently higher performance than the DCT-based state-of-the-art spatial-frequency fusion method and the compared baseline models. The ablation study further indicates that the FFT-based frequency branch provides complementary spectral cues when fused with spatial features. In addition, FFT-based frequency analysis and Grad-CAM/LIME explanations show consistent evidence around manipulation-sensitive facial regions, including the eyes, mouth, nose, and facial boundaries.
arXiv:2607.17924v1 Announce Type: new
Abstract: Multi-agent policy optimization, exemplified by PPO-based methods, is a key branch of cooperative Multi-Agent Reinforcement Learning (MARL). A central design question is how many neighboring agents\footnote{In this paper, "neighbors" refer not only to physical proximity but also to agents whose actions influence one another.} to aggregate in order to effectively utilize global information for cooperation. This decision must be made along two dimensions: in the advantage (which agents' rewards contribute to the credit signal) and in the ratio (which agents' likelihood ratios form the clipped importance weight). Existing methods occupy scattered, underexplored points on these two axes: IPPO treats both separately; MAPPO pairs a team-level advantage with per-agent ratios; HAPPO employs sequential ratios with per-agent advantages; and single-agent reductions operating on factorized joint policies aggregate both into fully joint products. We formalize these two design choices as support matrices $\SA$ and $\SR$, and prove a canonical structure: the expected multi-agent policy optimization objective depends on the pair $(\SA,\SR)$ only through their matrix product $\tS=\SR\SA$. This yields two key consequences: (i) Redundancy: the two support matrices are interchangeable with respect to the signal, meaning neither aggregation pattern is inherently superior.(ii) Variance Ordering: the advantage aggregates rewards as a sum (additive variance with an interior bias-variance optimum at the coupling neighborhood), whereas the ratio aggregates likelihood ratios as a product (multiplicative variance that grows exponentially with support size, with no accompanying bias reduction). The resulting design principle is unambiguous: aggregate neighbors in the advantage, sized to the coupling neighborhood, and keep the ratio per-agent.
arXiv:2607.17531v1 Announce Type: new
Abstract: Test-time collaboration, including self-consistency, best-of-N selection, critic models, and verifier pipelines, is often credited with broadly improving LLM reasoning, yet its gains are uneven and sometimes negative. We ask when training-free collaboration should be expected to help. For a fixed candidate pool, we decompose a selector or verifier's net gain into measurable factors: recoverable mass, verification-signal coverage, conditional selection quality, and harm to already-correct outputs. This reframes collaboration as a candidate-selection problem rather than as an intrinsic property of a multi-agent topology. Across LiveCodeBench, MATH Level-5 hard subjects, and GPQA-Diamond, gains are bounded first by the oracle gap and then by signal fidelity, which we measure directly as candidate-level agreement between verifier verdicts and official labels. On LiveCodeBench, a public-test verifier (MCC 0.825) gains +8.14 percentage points (pp) over a first-sample baseline; a generated-test verifier (MCC 0.248) improves by +2.70pp and is not statistically distinguishable from an LLM selector, but operates at near-zero harm versus the selector's 4.69% harm rate. On MATH, a symbolic answer-equivalence selector beats self-consistency by +4.67pp, while LLM selectors are negative. On GPQA-Diamond, recoverable mass is only 3.03% and 87.54% of candidate pools are answer-identical; a weaker model's pools shrink both further, suggesting that oracle gap is a joint property of task, model, and sampling configuration. Our framework yields a practical pre-deployment diagnostic: estimate the oracle gap, then measure coverage, signal fidelity, and harm before investing in collaboration.
arXiv:2607.17535v1 Announce Type: new
Abstract: Agentic retrieval-augmented generation (RAG) systems increasingly retrieve external evidence and orchestrate tools for knowledge-intensive applications. In Multi-Hop question answering, agents chain facts across documents. Existing defenses focus on content poisoning, which injects false facts, and prompt injection, which embeds directives. We identify a third attack surface: the salience channel, through which fact position, emphasis, framing, and semantic proximity can redirect reasoning even when all retrieved claims are true and no instructions are present. We formalize Salience Induction as truth-preserving edits that redirect Multi-Hop attribute binding while leaving the retrieval trace semantically intact. We define six Salience-Editing operator classes and build an iterative proposer-verifier pipeline under factual and stealth constraints. We also introduce SalientWiki-MH, a decoy-annotated Multi-Hop benchmark. Evaluations across five frontier model families (GPT, Claude, Gemini, DeepSeek, and Qwen) and three agent architectures (ReAct, Reflexion, and tool-calling) show broad generalization. Under a 30% edit budget, Salience Induction achieves an 83.3% attack success rate; the strongest evaluated baseline defense leaves 75.7% post-defense ASR. Untargeted rewriting further reduces attacks only by degrading neutral task success. Our lightweight input-side defense, Salience Normalization, reduces attack success to 15.3% under standard attacks and 23.6% under an adaptive attack. These results show that truthfulness and instruction filtering alone are insufficient: robust agentic RAG also requires defenses against salience-relevance decoupling.
arXiv:2607.17452v1 Announce Type: new
Abstract: Measuring conversational states such as cognitive load and conversational power from multimodal behavior requires characteristic features that are not only predictive but also reliable across task contexts. We present a three-dimensional evaluation framework assessing predictive accuracy, cross-task generalizability, and test-retest reliability, applied to interactional, acoustic, and linguistic features extracted from dyadic conversations during collaborative tasks performed over a video-conferencing platform (AVCAffe dataset; 53 dyads, 9 tasks). Our results show that no single feature family dominates all three dimensions. Linguistic features show the highest predictive accuracy for cognitive load but collapse under cross-task evaluation, revealing sensitivity to task-specific vocabulary. Additionally, acoustic reliability, often reported as evidence of feature stability, degrades once speaker identity is controlled, confirming that standard prosodic features measure vocal characteristics rather than conversational state. Interaction features provide the only genuinely reliable signal, unchanged after speaker normalization. Interestingly, classifying power role remained near chance baseline across all conditions, indicating limitations of task-level aggregated behavior for predicting power role in conversation. Our findings reveal three insights: (1) linguistic features predict best but generalize poorly across task contexts; (2) acoustic reliability collapses to near-zero once speaker identity is controlled, challenging standard evaluation practice; and (3) interaction features provide the only genuinely reliable signal, with floor dominance predicting within-dyad cognitive load asymmetry. These results argue for speaker normalization and multi-dimensional evaluation as prerequisites for context-aware, robust multimodal feature selection in conversational systems.
arXiv:2607.16635v1 Announce Type: new
Abstract: We present TellTale, a text-only approach to ambivalence/hesitancy (A/H) recognition in interview videos, evaluated on the BAH dataset as part of the 3rd A/H Video Recognition Challenge (11th ABAW Workshop, ECCV 2026). Although the dataset provides video, audio, facial crops, and transcripts, TellTale relies on the transcript alone and combines three probability streams. Two text encoders, multilingual-e5-large and mDeBERTa-v3-base, are fine-tuned with parameter-efficient LoRA adapters under a multiple-instance learning (MIL) objective, in which transcript chunks are scored individually and pooled with a smooth maximum so that only the video-level label is needed for supervision. The third stream requires no training: a quantized 14B instruction LLM is prompted, zero-shot, to rate each transcript for A/H. The three probabilities are combined by a weighted average and a single decision threshold, both selected on participant-grouped cross-validated predictions. On the organizer-scored private test set of 152 videos from unseen participants, TellTale achieves a Macro-F1 of 0.7364 and an average precision of 0.7940, compared with 0.2827 Macro-F1 for the official vision-based baseline.
arXiv:2607.16783v1 Announce Type: cross
Abstract: We study thermal transport across copper-graphene-copper interfaces using Non-Equilibrium Molecular Dynamics (NEMD), focusing on the influence of finite domain length and domain configuration, including lattice initialisation and associated graphene wrinkling, on the predicted thermal conductivity and Kapitza resistance. In the literature, NEMD simulations identified trends in the Kapitza resistance of graphene-copper interfaces. However, the simulation outcomes and reliability may depend heavily on configuration choices that are underexplored in the literature. We identify a strong sensitivity of the Kapitza resistance to domain configuration choices that affect the lattice constants and atomic density. We show that two conventional lattice initialisation strategies yield a factor of two difference in the Kapitza resistance, despite differences of only a few per cent in the lattice parameters. This behaviour is accompanied by strain-dependent shifts in the graphene and copper phonon spectra, and by increased phonon overlap at lower strain. Counter to conventional expectations, greater phonon-mode overlap coincides with higher Kapitza resistance, showing that spectral overlap alone cannot capture the interfacial heat-transfer dynamics. We suggest that in lattices initialised with lower residual strain, a damping boundary layer develops near the interface, and increases thermal resistance, as indicated by increased local structural disorder and local spectral broadening over a wider interfacial region. Beyond strain- and density-related effects associated with the lattice constants, Kapitza resistance shows no significant dependence on domain length or boundary temperature enforcement in this study. By contrast, the copper lattice conductivity exhibits clear domain-size and temperature dependence, consistent with phonon mean-free-path limitations and supported by phonon spectral analysis.