arXiv:2605.25881v1 Announce Type: new
Abstract: Dielectric metasurfaces offer a path to high-efficiency optical components at the sub-wavelength scale. In this work, we utilize the wavelength-dependent birefringence of TiO2 nanopillars to create transmissive color filters with switchable spectral responses. We demonstrate that the same physical nanostructure array can function as either an additive (band-pass) or a subtractive (band-stop) filter solely by rotating an external polarizer. The metasurface was fabricated via a modified damascene process, enabling high-aspect-ratio structures with spatial resolutions down to 250 nm. The filters achieve transmission efficiencies of 70% tailorable across the entire visible spectrum, avoiding the ohmic losses typical of plasmonic alternatives. This approach provides a tunable, high-efficiency platform offering a versatile and compact solution for next-generation high-resolution imaging systems.
Science Journals
arXiv:2604.00963v2 Announce Type: replace
Abstract: We show polylogarithmic mixing time bounds for the alternating-scan sampler for positively weighted restricted Boltzmann machines. This is done via analysing the same chain and the Glauber dynamics for ferromagnetic two-spin systems, where we obtain new mixing time bounds up to the critical thresholds.
arXiv:2505.03677v3 Announce Type: replace
Abstract: Learning maps between function spaces with a strong inductive bias is a central challenge in soft computing, especially when training data are scarce and standard deep architectures overfit. We introduce a \emph{neural integral operator} (NIO) framework based on integral equations of the first kind, in which the Urysohn kernel of the operator is parameterized by a feed-forward network~$G_{\theta_G}$ and the latent function is produced by a convolutional encoder~$E_{\phi_E}$, both trained jointly end-to-end via cross-entropy loss. The integral defining the learned operator is approximated by Monte Carlo sampling, which we argue acts as an implicit stochastic regularizer operating at the level of the integrand and complementing parameter-level regularizers such as weight decay and dropout. We benchmark the framework on three real-world spectroscopic classification tasks (FT-IR fruit purees, NIR meat, NIR textiles) of varying size and complexity, against traditional machine learning (decision tree, support vector machine, with and without UMAP) and modern deep learning baselines (FFNN, CNN+FFNN, shallow CNN, transformer). The proposed NIO is consistently among the top two performing models across all datasets and metrics, achieves the best results on the most challenging small-and-complex dataset (Textile), and yields lower performance variance than competing deep models in the small-data regime. The results suggest that operator-learning architectures with stochastic numerical integration are a viable soft-computing strategy for inverse problems in spectroscopy when conventional deep learning approaches are limited by data scarcity.
arXiv:2604.17120v2 Announce Type: replace
Abstract: Varkonyi and Domokos (2006) proved that convex homogeneous bodies with exactly one stable and one unstable equilibrium point exist. Sloan (2023) gave the first analytical parameterization, with radial function $R(\theta,\phi)$ having exactly two critical points on $S^2$.
This is the v2 amendment-of-record of arXiv:2604.17120. v1 claimed Sloan's parameterization does not produce mono-monostatic bodies and reported a 13-member catalog of Fourier/radial extensions certified at ECS=1 via mesh-vertex drainage-basin analysis. Following correspondence with P. L. Varkonyi (BME), an analytical verification suite was built around the Varkonyi-Gauss identity.
Finding 1: Sloan's parameterization does produce mono-monostatic bodies in a strictly-convex sub-regime ($\beta \lesssim 0.036$), where $K_{\min} > 0$ and the identity certifies ECS=1. v1 missed this because its mesh-vertex oracle over-counted on shallow COM-height landscapes. At Sloan's published $\beta=0.05$, strict convexity is lost ($K_{\min}=-0.569$; $K<0$ over 4.01% of surface); the identity's precondition fails. v1's "global surface information" mechanism is replaced by the strict-convexity precondition.
Finding 2: Of v1's 13 catalog instances only Phase-1 ($\beta=0.023149$, $a_1=0.234433$, $k=1$) survives identity-based verification; the remaining twelve were per-$k$ optimizer extrema overshooting the strict-convex boundary. Probing the regime interior verifies further mono-monostatic bodies in $k=2$ and $k=3$ sub-families: the verified set is an open regime in $(\beta, a_1, k)$, not a discrete list.
Finding 3: v1's ECS=1 readings for the 9 radial-family members reflected drainage-basin merging; the $r=0.9993$ gentleness-robustness correlation is retracted.
arXiv:2504.01362v2 Announce Type: replace-cross
Abstract: In this article, we describe the theoretical foundations of the Macaulay2 package ConnectionMatrices and explain how to use it. For a left ideal in the Weyl algebra that is of finite holonomic rank, we implement the computation of the encoded system of linear PDEs in connection form with respect to an elimination term order that depends on a chosen positive weight vector. We also implement the gauge transformation for carrying out a change of basis over the field of rational functions. We demonstrate all implemented algorithms with examples.
arXiv:2604.22948v2 Announce Type: replace
Abstract: History-dependent sampling can reduce long-run Monte Carlo variance by discouraging redundant revisits, but existing schemes typically encode history through empirical measure on finite state spaces, which is infeasible in high-dimensional discrete configuration spaces or ill-posed in continuous domains. We propose Score-Repellent Monte Carlo (SRMC) framework that summarizes trajectory history by a running average of score evaluations in $\mathbb{R}^d$, where $d$ is the dimension of the score and state representation. This history is converted into a surrogate target through an exponential score tilt, indexed with $\alpha$ that represents the strength of repellence in controlling the magnitude of the history-based repulsion. The surrogate family is normalization-free in the standard MCMC sense, yielding a generic wrapper: at each iteration, any base kernel targeting $\pi$ can instead be run on the current surrogate $\pi_{\theta_n}$ while the history is updated online. We analyze the coupled evolution of the history recursion and Monte Carlo estimators using stochastic approximation with controlled Markovian noise, establishing almost sure convergence and a joint central limit theorem. We further identify regimes in which the asymptotic covariance decreases as $\alpha$ increases, with scaling $O(1/\alpha)$, extending the near-zero-variance effect of finite-state history-dependent samplers to general state spaces with constant memory. Experiments on continuous targets and discrete energy-based models demonstrate improved estimator variance and mode coverage, while retaining $O(d)$ memory usage and modest per-iteration overhead.
arXiv:2507.03159v2 Announce Type: replace
Abstract: We present \texttt{MathOptAI.jl}, an open-source Julia library for embedding trained machine learning predictors into a JuMP model. \texttt{MathOptAI.jl} can embed a wide variety of neural networks, decision trees, and Gaussian Processes into a larger mathematical optimization model. In addition to interfacing a range of Julia-based machine learning libraries such as \texttt{Lux.jl} and \texttt{Flux.jl}, \texttt{MathOptAI.jl} uses Julia's Python interface to provide support for PyTorch models. When the PyTorch support is combined with \texttt{MathOptAI.jl}'s gray-box formulation, the function, Jacobian, and Hessian evaluations associated with the PyTorch model are offloaded to the GPU in Python, while the rest of the nonlinear oracles are evaluated on the CPU in Julia. \MathOptAI is available at https://github.com/lanl-ansi/MathOptAI.jl under a BSD-3 license.
arXiv:2605.14334v2 Announce Type: replace
Abstract: We study the problem of secure joint source-channel coding for multimodal semantic sources transmitted over noisy wiretap channels. The source model consists of $m$ modalities (e.g., image, audio, and sensor data), all represented as random variables. The encoder observes independent and identically distributed samples of an arbitrary non-empty subset of modalities. The samples are encoded and transmitted over a discrete memoryless wiretap channel. The legitimate receiver reconstructs all modalities. We extend the rate-distortion-perception problem formulation to multimodal sources. We establish converse and achievability bounds on the fundamental limits of transmission rate, fidelity, and secrecy, under per-modality distortion and perception constraints, and per-subset equivocation constraints. We show that the fundamental limit for secrecy consists of three operationally distinct components: the level of compression, the secret key rate, and the statistics of the wiretap channel.
arXiv:2404.14557v2 Announce Type: replace
Abstract: This study focuses on optimizing the design parameters of a Dual Active Bridge (DAB) converter for use in 350 kW DC fast chargers, emphasizing the balance between efficiency and cost. Addressing the observed gaps in existing high-power application research, it introduces an optimization framework to evaluate critical design parameters,number of converter modules, switching frequency, and transformer turns ratio,within a broad operational voltage range. The analysis identifies an optimal configuration that achieves over 95% efficiency at rated power across a wide output voltage range, comprising seven 50 kW DAB converters with a switching frequency of 30 kHz, and a transformer turns ratio of 0.9.
arXiv:2509.26545v3 Announce Type: replace
Abstract: This work demonstrates that the Lyapunov method can effectively identify the growth rate of a linear time-periodic system describing cold fresh water on top of hot salty water with a periodically time-varying background shear flow. We employ a time-dependent weighting matrix to construct a Lyapunov function candidate, and the resulting linear matrix inequalities are discretized in time using the forward Euler method. As the number of temporal discretization points increases, the growth rate predicted from the Lyapunov method or the Floquet theory will converge to the same value as that obtained from numerical simulations. Additionally, the Lyapunov method is used to analyze the most dangerous disturbance, and we also compare computational resource usage for the Lyapunov method, numerical simulations, and the Floquet theory.
arXiv:2510.08558v3 Announce Type: replace
Abstract: A long-term goal of language agents is to learn and improve through their own experience, ultimately outperforming humans in complex, real-world tasks. However, training agents from experience data with reinforcement learning remains difficult in many environments, which either lack verifiable rewards (e.g., websites) or require inefficient long-horizon rollouts (e.g., multi-turn tool use). As a result, most current agents rely on supervised fine-tuning on expert data, which is challenging to scale and generalizes poorly. This limitation stems from the nature of expert demonstrations: they capture only a narrow range of scenarios, and expose the agent to limited environment diversity. We address this limitation with a middle-ground paradigm we call early experience: interaction data generated by the agent's own actions, where the resulting future states serve as supervision without reward signals. Within this paradigm, we study two strategies of using such data: (1) implicit world modeling, which uses collected states to ground the policy in environment dynamics; and (2) self-reflection, where the agent learns from its suboptimal actions to improve reasoning and decision-making. Evaluation across eight diverse environments and multiple model families shows that our approaches consistently improve effectiveness and out-of-domain generalization, highlighting the value of early experience. Moreover, in environments with verifiable rewards, our results provide promising signals that early experience offers a strong foundation for subsequent reinforcement learning, making it a practical bridge between imitation learning and fully experience-driven agents.
arXiv:2605.24491v1 Announce Type: new
Abstract: Open-source energy system models disaggregate zonal electricity demand to substations through Voronoi-based preprocessing pipelines that combine socioeconomic weighting with auxiliary spatial corrections. Whether the same auxiliary data helps or harms when the weighting component shifts from rule-based to learned has not been investigated. We fix Voronoi partitioning and cross two design axes on metered demand from 1,891 British primary substations: the demand-weighting method and the mechanism through which Nighttime Light (NTL) intensity and substation-proximity signals enter the allocation, giving 15 configurations. Mechanism-isolation experiments further test additive post-correction and random-noise controls to pinpoint the structural cause of any performance reversal. The same auxiliary data reduces RMSE by 41 % on the static base but increases it by 21 % on the GNN base under multiplicative post-correction (p < 0.001 for both); the best static pipeline outperforms the best GNN variant by 19 %. Post-correction on the GNN improves rank-order correlation (p < 0.001) yet worsens absolute error, so correlation-only evaluation masks the calibration penalty. The isolation experiments trace this reversal to the multiplicative correction form under demand conservation constraints, not to signal redundancy; switching to additive post-correction eliminates the antagonism entirely. A transfer check on 13 German primary substations confirms directional replication and shows amplified antagonism where the GNN baseline already explains over 95 % of demand variance. The NTL and proximity signals behind the 41 % static improvement are publicly available at no cost and should be adopted as default corrections in static pipelines; method evaluation should report RMSE and correlation jointly, as the two metrics diverge under post-correction on learned representations.
arXiv:2605.24492v1 Announce Type: new
Abstract: Vision-language models have demonstrated impressive capabilities in general medical visual question answering, yet due to limited interpretability, it remains unclear whether their predictions reflect evidence-grounded clinical reasoning or reliance on spurious priors. We introduce Med-R2 Bench, a hierarchical benchmark aligned with the clinical workflow to evaluate adversarial robustness with visual grounding. We design stepwise QA tasks to assess whether reasoning chains are strictly grounded in visual evidence across the four clinical stages, and employ adversarial perturbations to test robustness against misleading cues. Med-R2 comprises 42,432 images, 31 task categories, and 110,406 QA pairs. Evaluation across 14 VLMs reveals a sequential performance degradation along the four-stage clinical workflow. Adversarial experiments show that models rely heavily on correct prompts to guess answers. Even when provided with explicit visual cues, the models struggle to accurately align textual descriptions. Finally, we demonstrate stepwise fine-tuning using our hierarchical data significantly improves reasoning robustness, highlighting its potential to drive future improvements in evidence-based medical AI.
arXiv:2605.24495v1 Announce Type: new
Abstract: This paper presents Elevator-LIO, a LiDAR-inertial odometry framework designed to achieve continuous robot localization during elevator travel, thereby supporting cross-floor robotic tasks. To address the state-estimation problem in non-inertial frames, Elevator-LIO establishes a decoupled state-estimation model that separately models the robot motion relative to the elevator and the elevator motion itself, and embeds it into a mode-dependent iterated error-state Kalman filter framework. This framework degenerates to conventional LIO estimation in ordinary indoor environments, while enabling the propagation and constrained update of elevator-related states in elevator non-inertial environments, thereby achieving continuous and stable localization. An elevator mode manager detects elevator entry and exit events using LiDAR ranging statistics and estimated states, and introduces event-triggered zero-velocity and zero-acceleration updates when the elevator stops to suppress accumulated vertical drift. In addition, this paper adopts an adaptive voxel downsampling strategy to maintain a stable number of effective points under significant environmental scale changes. We conduct extensive experiments on 20 real-world sequences containing 79 elevator rides, including practical challenges such as large-scale spaces, long vertical travel, dynamic pedestrian interference, and mirror reflections. The results show that Elevator-LIO maintains continuous localization accuracy in all sequences, with terminal height error below 1 cm in 17 sequences. In contrast, existing representative localization systems perform poorly on these elevator sequences. Tests on the Hilti 2022/2023 datasets further show that the proposed method remains competitive in standard indoor scenarios. The project page is available at https://xiaofan4122.github.io/Elevator_LIO_Page/.
arXiv:2605.24497v1 Announce Type: new
Abstract: Large Reasoning Models (LRMs) have demonstrated remarkable capabilities in reasoning and generation tasks and are increasingly deployed in real-world applications. However, their explicit chain-of-thought (CoT) mechanism introduces new security risks, making them particularly vulnerable to jailbreak attacks. Existing approaches often rely on static CoT templates to elicit harmful outputs, but such fixed designs suffer from limited diversity, adaptability, and effectiveness. To overcome these limitations, we propose an adaptive evolutionary CoT jailbreak framework, called AE-CoT. Specifically, the method first rewrites harmful goals into mild prompts with teacher role-play and decomposes them into semantically coherent reasoning fragments to construct a pool of CoT jailbreak candidates. Then, within a structured representation space, we perform multi-generation evolutionary search, where candidate diversity is expanded through fragment-level crossover and a mutation strategy with an adaptive mutation-rate control mechanism. An independent scoring model provides graded harmfulness evaluations, and high-scoring candidates are further enhanced with a harmful CoT template to induce more destructive generations. Extensive experiments across multiple models and datasets demonstrate the effectiveness of the proposed AE-CoT, consistently outperforming state-of-the-art jailbreak methods.
arXiv:2605.24499v1 Announce Type: new
Abstract: The popularity and rapid development of Cloud Computing in recent years has led to a vast number of publications capturing the accumulated knowledge in this field. Due to the interdisciplinary nature and significant relevance of cloud computing research, it has become increasingly challenging to comprehend the overall structure and progress of this field without employing analytical methods. While the evaluation of scientific research has a long tradition in many fields, we have identified a lack of comprehensive scientometric studies specifically focused on cloud computing. This study applies scientometric techniques to empirically examine cloud computing research evolution and current state from a macroscopic perspective. We employed the CiteSpace tool for visual analysis, exploring topics related to cloud computing by retrieving papers published between 2014 and 2023 from the Web of Science Core database. Our approach involved constructing collaboration networks among authors, institutions, and countries to pinpoint the most prolific contributors in each category. Through the analysis of core journal distributions via journal co-citations, document co-citation networks, and clustering analysis, we uncovered the underlying research topics and knowledge structure. The results of this study enhance our understanding of patterns, trends, and other critical factors, offering a foundation for guiding research activities, sharing knowledge, and fostering collaboration in the field of cloud computing research.
arXiv:2605.19245v2 Announce Type: replace-cross
Abstract: Neutral-atom entangling gates are commonly analyzed with a single effective Rydberg-pair state, but near F\"orster resonances the pair manifold contains resonantly coupled interaction channels that change both the control landscape and the achievable fidelity. We develop a two-eigenstate model for this regime and show that when allowing for coupling to both pair states in the resonance, the gate fidelity is bounded by $\mathcal{F}\leq 1-(\pi/2)/(V\tau_R)$, for interaction strength $V$ and Rydberg lifetime $\tau_R$. We construct a gate protocol that saturates this bound in the large-Rabi-frequency limit, improving the existing fidelity limit by approximately $40\%$. We also evaluate common gate protocols near F\"orster resonances and find that retaining the exchange dynamics increases predicted fidelities by up to two orders of magnitude over earlier treatments.
arXiv:2605.24503v1 Announce Type: new
Abstract: As AI-powered compliance monitoring becomes increasingly important in public governance and industrial safety, the ability to provide verifiable evidence and traceable accountability signals is essential. However, existing video anomaly detection datasets focus on event-level binary classification, lacking the rule-driven, explainable analysis required for real-world compliance scenarios. We introduce FoodMonitor, a benchmark for explainable compliance analysis in commercial kitchen surveillance. FoodMonitor comprises 477 video clips with 3,307 violation annotations across a dual-channel design covering both person-level and environment-level violations. Each annotation specifies which rule was violated, what non-compliant behavior occurred, and who committed it with frame-level bounding boxes. We establish a unified evaluation protocol with a two-stage matching mechanism that separately assesses spatial localization and semantic understanding, along with a composite metric ($C_{\text{score}}$) that balances environment and person detection performance. Systematic evaluation of several state-of-the-art multimodal large language models reveals that the best-performing model achieves only 0.360 $C_{\text{score}}$, with spatial localization and fine-grained rule understanding emerging as the primary bottlenecks. Our analysis identifies two distinct failure modes: localization-dominated errors and semantics-dominated errors, providing diagnostic insights for future model development.
arXiv:2605.24505v1 Announce Type: new
Abstract: Digital in-line holographic microscopy enables label-free tracking and metrology, but achieving nanometric sub-pixel displacement accuracy over a wide capture range remains challenging. Frequency-domain registration based on the discrete Fourier transform (DFT) is globally stable and tolerant to large shifts, yet it suffers from sub-pixel quantization and interpolation artifacts that limit precision near zero displacement. In contrast, spatial-gradient refinement such as Lucas-Kanade (LK) can reach very high sub-pixel accuracy, but it is strongly initialization-limited and prone to divergence outside a narrow convergence basin. Here we propose a Hybrid Spectral-Spatial Domain (HSSD) framework that resolves this trade-off by combining the global robustness of a DFT-based coarse estimator with LK refinement of the residual motion. The DFT stage provides reliable initialization and substantially extends the capture range, while the LK stage suppresses the precision floor by reducing interpolation and quantization errors characteristic of standalone frequency-domain methods. We validate HSSD using numerical simulations and experiments in a transmission in-line holographic imaging system, achieving nanoprecision displacement measurement and stable tracking across a wide range of displacements and defocusing conditions. This hybrid strategy enables reliable nanometric localization in regimes where standalone DFT or LK methods either fail to converge or saturate in accuracy.
arXiv:2605.24516v1 Announce Type: new
Abstract: Mixed-motive scenarios are ubiquitous in real-world multi-agent interactions, where self-interested agents often defect for immediate rewards, overlooking the potential of altruistic cooperation to improve long-term gains and collective welfare. Peer punishment can deter defection, but as costly second-order altruism, its persistent imposition may undermine the punisher's interests. Existing approaches often struggle to effectively implement punishment to promote cooperation. To balance the efficacy and cost of punishment, we propose Adaptive Punishment for Cooperation (APC), a distributed method that determines punishment intensity based on both a dynamic punishment probability and the severity of defection. This dynamic probability substantially reduces costly and ineffective punishment while also promotes cooperation. To accurately assess defection and its severity, we use a defection awareness module, whose learning is guided by game reward. Theoretical analysis and empirical results show APC performs effectively in iterated public goods game. Empirically, APC also significantly outperforms existing baselines across sequential social dilemmas, learning rational and effective punishment policies that foster cooperation by strategically deterring defection.
arXiv:2605.24517v1 Announce Type: new
Abstract: CLI agents are the closest thing language models have to an embodied setting: the model emits commands, the terminal executes them, and the returned stream -- stdout, errors, files, logs, and traces -- records the consequences. We argue that this stream is a supervision signal, but standard agent RL discards it: GRPO-style training updates action tokens with sparse outcome-level rewards while ignoring environment responses already in the rollout. Failed rollouts provide little policy-gradient signal despite containing rich evidence about how the environment responds. We introduce ECHO (Environment Cross-entropy Hybrid Objective), a hybrid objective that combines the standard policy-gradient loss on action tokens with an auxiliary loss that trains the policy to predict environment observation tokens resulting from its own actions. ECHO reuses the same forward pass as GRPO, requires no additional rollouts, and turns terminal feedback into dense supervision for all rollouts. ECHO doubles GRPO pass@1 on TerminalBench-2.0: Qwen3-8B improves from 2.70% to 5.17%, and Qwen3-14B from 5.17% to 10.79%. ECHO also produces policies that better predict terminal dynamics, even on trajectories they did not generate: across held-out rollouts, it sharply reduces environment-token cross-entropy while GRPO alone barely changes it. From base Qwen3-8B, ECHO matches expert-SFT-then-GRPO performance on held-out terminal tasks without expert demonstrations, and recovers roughly half of the expert-SFT initialization benefit on TerminalBench-2.0. In some settings, the environment prediction loss alone enables verifier-free self-improvement, allowing policies to improve on unseen OOD tasks by learning only from environment interactions. Together, these results suggest that environment observations are not merely context for future actions, but a dense, on-policy supervision signal already present in every rollout.
arXiv:2605.24518v1 Announce Type: new
Abstract: The quadratic complexity of self-attention in Transformer models remains a significant bottleneck for processing long sequences and deploying large language models efficiently. For this approach, there has been significant research into Sparse Attention, and Deepseek Sparse Attention has combined various methods of creating segments of tokens to reduce the time complexity. This paper introduces a novel approach, Grammatically-Guided Sparse Attention, which constrains attention computations based on the grammatical roles of tokens. By leveraging Parts-of-Speech (POS) tags, attention masks are dynamically generated that enforce linguistically coherent connections between tokens, reducing the computational graph without sacrificing essential linguistic dependencies. Two masking strategies are proposed and evaluated: a hard mask that strictly allows only predefined grammatical interactions, and a soft mask that biases attention towards these interactions. The experiments, conducted on the SST-2 sentiment classification task using a DistilBERT-like architecture, demonstrate that Grammatically-Guided Sparse Attention maintains comparable accuracy to full attention while significantly reducing the theoretical computational overhead. Preliminary results show accuracy values of 0.8200 for hard masking and 0.8165 for soft masking, closely matching the 0.8200 of full attention, providing a path towards more efficient, interpretable, and linguistically-informed Transformer architectures.
arXiv:2605.24524v1 Announce Type: new
Abstract: In non-invasive neural language decoding, results can be inflated by sources that are not stimulus-evoked neural evidence: decoder priors, embedding-based metrics, and non-neural structural nuisances such as signal duration. The methodological challenge is therefore attribution: a reported gain is more informative when it can be traced to a specific source. We recast stimulus-locked MEG-to-audio retrieval as an auditing framework that separates apparent performance into three sources - structural shortcuts, window-level stimulus-locked evidence, and cross-window contextual aggregation - and provides a diagnostic for each. Signal-blind Gaussian noise reaches 66.3% Rank@1 (R@1) under variable-length decoding but collapses to near chance once fixed-duration windows and stimulus-identity splits are enforced, isolating structural leakage. Under these controls, fixed-window retrieval recovers measurable MEG-audio discriminability, while an oracle sentence-bucket diagnostic shows that 95.7% of Top-1 errors select the wrong sentence, localising the residual bottleneck to sentence-level competition. We audit this contextual source with Group Context Bias (GCB), an inference-time additive logit bias that pools sentence-consistent evidence across windows while leaving the base retrieval scores and candidate pool fixed. Used as a score-space intervention, GCB makes the contextual source measurable: R@1 shifts from 44% to 52% on Gwilliams and from 22% to 29% on MOUS under the same fixed setting. GCB is auditable under this design: its effect collapses under random-grouping perturbations and vanishes when local evidence is attenuated in MEG or is near chance in EEG, supporting its use as a controlled source-attribution intervention. These results suggest that brain-to-language performance should be source-attributed, not merely reported.
arXiv:2605.24526v1 Announce Type: new
Abstract: Interactive assistance systems typically provide feedback after an action has been completed, supporting error recovery but not preventing the error itself. We present TRAFA, a real-time predictive feedback system for procedural tasks that intervenes before errors are committed. TRAFA operationalizes predictive feedback through a Track-Forecast-Act framework that tracks hand and object state, forecasts user motion conditioned on scene context, and triggers feedback when a predicted action is likely to violate task constraints. We instantiate this pipeline in a sequential assembly setting and evaluate it through both technical benchmarking and a controlled user study against conventional reactive feedback. Our results show that predictive feedback improves task accuracy and efficiency while maintaining a comparable number of feedback events. These findings position feedback timing as a key dimension in system design and show how real-time anticipation can be integrated into interactive systems to prevent errors before they occur.
Learnable Shape Prototypes with Occlusion-Geometry-Guided Injection for Amodal Instance Segmentation
arXiv:2605.24533v1 Announce Type: new
Abstract: Amodal instance segmentation aims to predict the complete object mask including occluded regions that lack pixel-level observations and must be inferred with the aid of shape priors. Existing methods acquire shape priors through fixed-capacity encoding spaces or expensive generative models, and inject them uniformly across all spatial positions without adapting to the varying prior demand between visible and occluded regions. In this paper, we propose a gated reliability-adaptive shape prior framework, which introduces a shape prior memory module that combines learnable prototypes via cross-attention to produce instance-adaptive shape priors through weighted prototype combination rather than generation. A spatial adaptive reliability gate then employs the signed distance field of the visible mask to modulate injection intensity at each position according to its occlusion depth, preserving reliable features in visible regions while directing shape compensation toward occluded areas. Experiments on two mainstream amodal instance segmentation benchmarks demonstrate that the proposed method outperforms existing approaches under multiple evaluation settings, improving the mean intersection-over-union over occluded regions by over 11 percentage points on one of the two benchmarks under the standard setting, while using approximately one-third of the total parameters. Linear probing analysis further reveals that the visible-mask cross-attention module implicitly encodes occlusion geometry into visual token representations, explaining the effectiveness of the proposed module decomposition.