# Research Papers — 2026-10 · bundle page 2 of 156 > Full-text markdown of papers #51–100 of 7799 published in 2026-10, newest first. AI-readable mirrors of the full 7799-paper 2026-10 archive (50 papers per file). Canonical page: https://aipod.dev/papers/archive/2026-10/page/2 [← Previous bundle](https://aipod.dev/papers/archive/2026-10.md) [Next bundle →](https://aipod.dev/papers/archive/2026-10/page/3.md) ### Blitz: Accelerating the Diffusion Modules of AlphaFold3 and Boltz-2 Authors: Vinay Saji Mathew, Soundar Kumara, Gretta Kellogg, William K. M. Lai Published: 2026-10-16 Categories: cs.LG arXiv: https://arxiv.org/abs/2610.23703v1 Diffusion-based protein structure prediction models produce all-atom structures in about 200 denoising steps, and that cost has driven a number of few-step samplers. We present Blitz, a few-step distillation framework that uses a geometric noise schedule to distribute steps more evenly across noise scales, then learns those updates through multistep consistency training. Physical steering and target-free candidate selection help reduce the remaining geometric defects at inference. We apply Blitz to AlphaFold 3 and Boltz-2, reducing denoising from 200 steps to eight. Because published few-step results use different benchmarks and scoring procedures, we compare both distilled models with their teachers and other few-step methods under a common evaluation protocol covering structural accuracy, interface quality, ligand placement, and stereochemical validity. Blitz retains teacher-level structural accuracy, with AF3-Blitz outperforming all evaluated few-step baselines and exceeding its teacher on several accuracy metrics. --- ### VISTA: Steering LLMs with Latent User Values Inferred from a Few Survey Answers Authors: Yinuo Xu, Jadie Adams, Brian Hu, Shivani Kumar, David Jurgens, Arslan Basharat Published: 2026-10-16 Categories: cs.CL arXiv: https://arxiv.org/abs/2610.23702v1 Personalizing a language model requires knowing what a person values, which stays largely consistent across tasks. Yet most methods either learn task-specific preferences or place the user's profile in the prompt, where models struggle to act on it. We introduce VISTA (Value-Informed Steering and Targeted Acquisition), which infers a latent profile from users' (full or partial) survey answers, and maps the profile through a hypernetwork to a user-specific steering intervention in a frozen LLM to select answer choices or generate free-text responses. Given the same evidence, VISTA matches or beats in-context profiles on synthetic preferences (Food2D), the World Values Survey (WVS), and PrimeX, without adding prompt tokens. To reduce the cognitive burden of completing full surveys, VISTA also chooses which questions to ask and when to stop: on WVS it stops after 6% of the questionnaire while retaining 94% of the full-history gain, where standard adaptive testing retains at most 42%. The dominant steering direction is an interpretable value axis on which VISTA places unseen users accurately. Our results point toward pluralistic alignment that adapts to individuals at deployable cost. --- ### Equivalence testing and run-count design for stochastic machine-learning experiments Authors: Dmytro Knopov, Nadiya Shvai Published: 2026-10-16 Categories: cs.LG, stat.ML arXiv: https://arxiv.org/abs/2610.23700v1 Claims that two training methods perform equally well often rest on a non-significant difference between two or three runs per method. We develop a protocol, with its power theory, for certifying practical equivalence in per-run metrics such as the AUROC (area under the receiver operating characteristic curve). It rests on a two-level (seed x item) variance model separating between-run dispersion from item-level noise. Equivalence within a pre-specified margin is tested by two one-sided tests (TOST), equivalently by a confidence interval inside the margin; a two-level bootstrap covers composite statistics. A closed-form formula converts a margin and a target power into runs per method; its shifted-t refinement is calibrated against the realised two-level process. Simulations on a 270-cell grid and under non-Gaussian seed laws confirm the formula within its design-relevant range, from five runs per method. For four published runs of a joint energy-based model, the failure to certify equivalence was largely decided by design: with the variance calibrated to those runs, two per method certify with probability near three per cent, whereas certifying 0.01 AUROC at 80 per cent power needs twelve even under exact equality. A second case study on a thousand-seed CIFAR-10 zoo and the MultiBERTs checks the formula against real seed laws (within 0.05 in that range) and resolves comparisons that two or three runs leave undecided. In a two-stage design (pilot, then formula), our experiments suggest that an upper confidence bound on the pilot variance substantially reduces the undersizing seen with its point estimate. --- ### Expressivity of IQP Feature-Map Quantum Kernels via Reproducing Kernel Hilbert Spaces Authors: Valeria Repetto, Stefano Melchionna, Romina D'Aurizio Published: 2026-10-16 Categories: quant-ph, cs.LG arXiv: https://arxiv.org/abs/2610.23692v1 Quantum kernels are among the most promising methods in quantum machine learning, providing a natural way to embed data into high-dimensional Hilbert spaces via quantum circuits and capturing non-linear relationships in data. Their learning behaviour is governed by the function class they represent, which is formally characterized by the structure of the associated Reproducing Kernel Hilbert Space (RKHS). However, a theoretical characterization of this function class, and of how it depends on the choice of feature map and its hyperparameters, is still lacking. In this work, we focus on the Instantaneous Quantum Polynomial (IQP) feature map and derive a Fourier representation of the associated RKHS generators as finite sums of trigonometric functions evaluated on polynomially transformed input variables, with a finite set of accessible frequencies determined by the circuit construction and scaled by the bandwidth hyperparameter. This representation identifies three bandwidth regimes with qualitatively distinct expressive properties from localised, through multimodal, to periodic. Numerical experiments on synthetic datasets with controlled periodic and multimodal structure, across data dimensions and feature-map configurations, reproduce the three regimes in both supervised classification and unsupervised anomaly detection, benchmarked against a classical Gaussian kernel. These results make the relationship between expressivity and data structure explicit for IQP kernels: expressive power translates into a practical benefit only when the accessible frequencies match with the structure of the data, which provides a criterion for selecting the bandwidth when the periodic or multimodal structure of data is known. --- ### HardEval: An Execution-Verified Benchmark for AI-Driven Security Hardening and Compliance Remediation Authors: Manohar Vellala, Soumya Banerjee, Sandip Roy, Ross Gore, Sachin Shetty Published: 2026-10-16 Categories: cs.CR, cs.LG arXiv: https://arxiv.org/abs/2610.23685v1 Implementing Security Technical Implementation Guide (STIG) controls remains a demanding operating-system administration task. Each natural-language security requirement must be translated into a precise and persistent configuration change. These changes span Secure Shell (SSH) configuration, authentication policy, audit rules, and kernel parameters. Existing benchmarks do not measure whether a model can perform this task on a live system. Code benchmarks evaluate isolated programs without persistent system state, while agent benchmarks largely focus on applications rather than operating-system configuration. In this paper, we introduce \textbf{HardEval}, an execution-verified benchmark for evaluating language-model-assisted security hardening. HardEval is built from Red Hat Enterprise Linux 8 STIG controls that fail on an unhardened baseline system. Expert-written reference remediations and Open Vulnerability and Assessment Language (OVAL) checks serve as hidden ground truth for grading; they are never shown to the model, and each task states only the required security outcome while withholding the implementation. The model produces a Bash script, which is executed on a live virtual machine and graded from fail to pass using the corresponding OpenSCAP/OVAL compliance check. Evaluation of fourteen frontier and open-weight configurations reveals large and uneven capability differences across security domains. A recurring failure mode is plausible security code that does not satisfy the exact compliance condition. For four open-weight configurations, multi-run Wilson confidence intervals preserve the observed ordering. HardEval is publicly available. --- ### Beyond the Benchmark: Conditioning, Domains and Failure Modes of Qwen3-TTS Authors: Séverin Baroudi, Yanis Labrak, Petr Motlicek, Hervé Bredin, Ricard Marxer Published: 2026-10-16 Categories: cs.SD, cs.CL arXiv: https://arxiv.org/abs/2610.23678v1 Qwen3-TTS reports state-of-the-art zero-shot voice cloning, but its technical report leaves open how the model should be conditioned, how it behaves outside read speech, and how it fails. We characterise its two cloning interfaces, in-context learning (ICL) and X-Vector, against Chatterbox, XTTS-v2 and F5-TTS on read, meeting, emotional and air-traffic-control (ATC) speech, using eighteen objective measures, an LLM judge, a spoofing countermeasure and a listening test. Neither interface dominates: ICL clones timbre more closely from correctly transcribed references, X-Vector clones timing, and ICL degrades sharply when a short reference transcript is wrong. Domain explains four times more variance than system choice. Qwen3-TTS leads on intelligibility and pitch register but trails on prosody, and the objective measures rank systems differently from listeners. An off-the-shelf countermeasure flags its clones no more often than genuine recordings, and on ATC radio it fails for every system. Code and audio are released. --- ### Agentic Physics-Grounded Video Generation via Memory-Augmented In-Context Reasoning Authors: Xikang Guan, Shaocong Yang, Yanqi Bao, Can Wang, Zheng Gu, Zhengqing Chen, Kai Chen, Jing Huo, Yang Gao Published: 2026-10-16 Categories: cs.CV arXiv: https://arxiv.org/abs/2610.23677v1 Physical plausibility is essential for realistic video generation. Existing methods impose physical guidance before synthesis via simulations or refined prompts, but this can limit dynamic diversity, miss errors that emerge only after generation, and require repeated regeneration when inconsistencies remain. We therefore study a generate-diagnose-repair paradigm that preserves the original generation capability and selectively repairs physical inconsistencies afterward, raising two key questions: what to repair and how to repair? To address them, we introduce PhysRepair, an agentic video generation framework that combines causal reasoning over the observed physical failure with prior repair experience to enable localized and adaptive correction. Specifically, Spatiotemporal Causal Constraints identify the first invalid physical state and its causally affected states, producing a repair specification that guides physically plausible local regeneration while preserving unaffected content. Experience-augmented In-context Learning combines the ongoing repair context with relevant experience from previous repairs, allowing the agent to adapt its repair decisions across diverse physical failures and accumulate completed trajectories for future cases. Experiments on Physics-IQ Verified across solid mechanics, fluid dynamics, optics, thermodynamics, and magnetism demonstrate that PhysRepair consistently improves the physical plausibility of generated videos over existing methods. Our project page is available at https://xxkkxxx.github.io/PhysRepair-Website/. --- ### On the possibilities and limitations of LLM-based multi-agent systems for multi-criteria decision-making in healthcare operations management Authors: Marco Boresta, Tommaso Giovannelli, Sepehr Namirad, Giulia Riccardi Published: 2026-10-16 Categories: math.OC, cs.AI, cs.MA arXiv: https://arxiv.org/abs/2610.23673v1 Multi-criteria models for healthcare operations need weights that express how stakeholders trade off patient waiting time, provider workload, and operating cost, but these weights are difficult to quantify. Teams of large language model (LLM) agents, each prompted to represent a stakeholder, have been used to supply them. We propose a framework for assessing the reliability of LLM-generated preferences as inputs to multi-objective optimization and quantifying how their uncertainty affects decisions. We study explicit preferences by asking agents to rate the importance of the criteria in an outpatient appointment scheduling problem, and implicit preferences by inferring trade-offs from the agents' pairwise choices between emergency department configurations. The computational experiments show that the explicit preferences respond to the clinic's operating conditions and differ across roles, but deliberation among the agents can erase these differences, whereas simple averaging preserves them. The pairwise choices used to reveal the agents' implicit preferences depend largely on which option is presented first, and we find limited evidence that a statistical model using contextual information alongside presentation order predicts the agents' choices on new comparison pairs better than a model using presentation order alone. We then formulate robust weighted-sum optimization models to account for preference uncertainty. In most cases, these models select different schedules or configurations from those obtained using average preference weights. These results indicate that LLM-generated preferences need an audit of this kind before they are used as optimization weights in healthcare decision support. --- ### OwlSight: A Panoramic Monitor for Chain-of-Thought Reasoning Authors: Lewen Yan, Yanxu Zhu, Jie Zhang, Jing Shao Published: 2026-10-16 Categories: cs.AI arXiv: https://arxiv.org/abs/2610.23670v1 Chain-of-Thought (CoT) monitoring offers actionable evidence for oversight AI systems, with the potential to enable early detection and intervention. However, existing approaches are focused on specific tasks, whereas the regulation of AI, particularly reasoning models, requires monitoring a broader range of tasks. We introduce Factor for CoT monitoring, a strategy that supports customizing monitoring requirements in the prompt, which enables a single monitor to cover different tasks by relying on different factors. Furthermore, we introduce two types of factors, fine-grained and coarse-grained (e.g., monitoring whether a specific email address has been leaked versus monitoring privacy leakage), to accommodate diverse user monitoring needs. Based on this, we develop a CoT monitor named OwlSight that relies solely on factor to monitor behaviors related to multiple dimensions, including capability, safety, and trustworthiness. Experiments across 15,971 reasoning trajectories show that OwlSight surpasses general-purpose LLMs by 1.32-30.57 and 11.71-27.81 points at fine and coarse granularity, respectively, and achieves state-of-the-art results on OpenAI's CoT monitorability benchmark. It also generalizes to unseen models' CoTs and remains robust to Factor rephrasings, CoT compression, and obfuscation, offering a promising paradigm for CoT monitoring that can contribute to the oversight of frontier AI. --- ### TUPLE-Bench: Evaluating and Mitigating Privacy-Utility Trade-offs in Typed Decision Models Authors: Yiqu Yang, Qiaoyuan Zheng, Yang Tian, Yu Fan, Elliott Ash Published: 2026-10-16 Categories: cs.CR, cs.AI arXiv: https://arxiv.org/abs/2610.23668v1 Typed decision models often return more than a selected category. They also expose probabilities for each candidate and, sometimes, a separate confidence score. These fields support downstream decisions but may reveal information beyond a recipient's authorization. We introduce TUPLE-Bench (Typed decision models' Utility and Privacy Leakage Evaluation Benchmark), a benchmark of paired synthetic cases in which recipients may learn a coarse operational category but not its fine-grained cause. Across nine models, instructions to withhold the cause fail to protect it: 91.88%-99.93% of responses reveal the cause through direct selection or residual scores. Even when causes are excluded from the candidate set, full outputs enable 62.85%-75.35% recovery on held-out cases, compared with 51.04%-57.64% using only the selected category (50% chance). Full outputs also reduce mean downstream cost in all 27 model-task comparisons. However, confidence fields increase leakage without consistently reducing cost. We further introduce workflow-constrained privacy quantization (WCPQ), which reduces secret recovery while preserving the behavior and cost of specified downstream workflows. Our findings highlight the need to audit decision interfaces field by field and workflow by workflow. --- ### TriRouter: Sparse-to-Dense Electrode-Space Routing for Robust EEG Motor Imagery Decoding Authors: Ying Zhou, Yaoxin, Mengting Liu Published: 2026-10-16 Categories: cs.LG, eess.SP arXiv: https://arxiv.org/abs/2610.23665v1 Motor imagery brain-computer interfaces require reliable decoding from multichannel EEG, but practical acquisition often yields sparse or corrupted electrode montages. We propose TriRouter, an electrode-space routing framework with discriminative-preservation, mask-robustness, and latent-compensation paths. A validation-calibrated router uses observable channel availability to allocate path contributions, preserving complete-channel discrimination while increasing representation recovery under severe missingness. In BNCI2014-001, TriRouter outperforms standard public backbones and a dynamic-spatial-filtering corrupted-EEG baseline. It improves accuracy over the strongest corrupted-EEG baseline by 3.68, 4.02, and 6.46 percentage points at nominal 25%, 50%, and 75% missing settings, and remains effective under structured regional missingness and additional low-channel dataset evaluation. --- ### PCDFusion: Proposal-Context-Detail Bayesian Rendering for Infrared-Visible Image Fusion Authors: Rui Liu, Lin Gu, Hesong Li, Ying Fu Published: 2026-10-16 Categories: cs.CV arXiv: https://arxiv.org/abs/2610.23663v1 Infrared-visible image fusion (IVIF) aims to combine visible and thermal infrared observations for human perception and downstream vision tasks. Visible and thermal infrared observations are physically complementary but unreliable in different ways. Existing deep IVIF methods mainly improve feature interaction and reconstruction, but often lack an explicit local rule for deciding when visible appearance should dominate and when thermal responses should contribute. To address this issue, we reformulate IVIF as local Bayesian posterior rendering, where fusion is controlled by posterior action weights rather than direct modality mixing. In this paper, we propose Proposal-Context-Detail Fusion (PCDFusion), a visible-anchored framework that implements this formulation through three rendering stages. PCDFusion preserves potentially useful thermal responses, updates exposure-degraded visible regions, and renders reliable band-limited infrared structures into the final luminance. Extensive experiments on public IVIF benchmarks show that PCDFusion improves fusion quality, target recovery, and downstream detection and segmentation performance under challenging illumination. The code is available at https://github.com/BIT-RuiLiu/PCDFusion.git. --- ### An Efficient Approach to Beetle Damage Detection of Banana Using Deep Learning Authors: Md. Mehedi Hasan Naeem, Ishtiyak Ahmmad Araf, Md. Arefin Haque Mahir, Tushar Kanti Saha Published: 2026-10-16 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2610.23661v1 Visible surface damage of banana fruits by Basilepta subcostata (the banana fruit scarring beetle) can reduce their market value. Here, manual inspection of Banana fruits is subjective and difficult to scale. This study presents a lightweight framework based on YOLOv8n deep learning model for locating individual bananas and classifying their visible surface condition as healthy or unhealthy. The term beetle damage in the paper title is used operationally for the locally observed scar-like surface condition; the model does not identify a specific beetle species or biological cause. A custom dataset of 2,473 images containing 5,827 annotated banana instances was prepared from local-market imagery in Bangladesh. The reported experiment used an 80/20 protocol with 1,982 training images and 491 test images. A pretrained YOLOv8n model was trained at 640x640 resolution for 50 epochs with a batch size of 4. On the test partition, it achieved 81.36% precision, 74.30% recall, 77.67% F1-score, 87.04% mAP@0.50 and 59.33% mAP@0.50:0.95. A YOLOv8s comparison gave slightly lower F1 and mAP while using a substantially larger checkpoint. The YOLOv8n checkpoint occupied 5.96 MB and required 6.54 ms/image for inference on an NVIDIA Tesla T4. The results support YOLOv8n as a compact baseline for instance-level banana surface inspection within the studied acquisition conditions. --- ### FlexPart: Text-Controlled Semantics-Aware 3D Part Segmentation at Flexible Granularity Authors: Yiqun Zhao, Jiachen Xu, Yihao Zhi, Yunfei Zhao, Zibo Zhao, Xintong Han, Chunchao Guo, Shenghua Gao Published: 2026-10-16 Categories: cs.CV arXiv: https://arxiv.org/abs/2610.23659v1 A 3D asset admits multiple valid part decompositions, while semantics-aware segmentation should reflect both the desired granularity and the parts a user wants to extract. Ideally, a practical segmentation model should strive for a simple principle: what you ask for is what you get. However, realizing this goal is hindered by two core obstacles: the lack of multi-granularity training data and the absence of a holistic control mechanism. We propose FLEXPART, a unified framework for text-controlled 3D mesh segmentation. To establish flexible control, we introduce a holistic control prompt comprising semantic part names with location and shape clauses. To support this mechanism, we utilize an automated Vision-Language Model (VLM) pipeline to construct a large-scale dataset of multi-granularity partitions and their corresponding prompts. Driven by this data, FLEXPART generates a view-independent part-identity field in native 3D space, directly assigning discrete part IDs to mesh faces. To compel the model to respect semantic conditions rather than rely on geometric priors, we train the model using randomized background dropping. Furthermore, an inference-time VLM planner can automatically complete descriptive clauses from 2D renderings, allowing users to simply name the desired parts in ordinary words. FLEXPART with planner-completed prompts outperforms the strongest evaluated baseline by 15% in mIoU, even when that baseline receives ground-truth guidance. On our held-out FLEXPARTBENCH, FLEXPART leads at each granularity and improves average mIoU by 22% over the strongest baseline. With only the requested part names supplied by the user, it also improves IoU by 35% over the strongest click-driven baseline on selective segmentation. --- ### A Language-Informed Loss for Visual Generation Authors: Suchir Agarwal, Keshigeyan Chandrasegaran, Ziyu Chen, Kyle Sargent, Juan Carlos Niebles, Li Fei-Fei, Jiajun Wu Published: 2026-10-16 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2610.23653v1 End-to-end visual generation with pixel-space loss has recently achieved competitive performance without the need for learned image tokenizers. In this work, we study whether language-informed supervision can provide a richer training signal than pixel loss for visual generation. We introduce LILIE (pronounced "lily"), a Language-Informed and Lossless Image Encoder. LILIE is an invertible image transform trained such that a vision-language model can recover captions from noised features. The encoder uses a normalizing flow transformer trained jointly with a vision-language model via next token prediction. We then use the encoder to extract a simple language-informed loss for text-to-image generation. We train JiT models with LILIE loss, reducing gFD by 14 points (66.84 $\to$ 53.01) against a pixel-only baseline on the GPIC dataset. Our code, encoders, and generative models are available at https://suchiragarwal.com/lilie. --- ### In Human Memories. Technology, Creativity and Self-Determination: An Ethical Proposal Authors: Francesco Vitucci, Giuseppe Silvi, Francesco Abbrescia, Anthony Di Furia, Francesco Scagliola Published: 2026-10-16 Categories: cs.CL arXiv: https://arxiv.org/abs/2610.23651v1 This research investigates the role of artificial intelligence (AI) in modern compositional practices, with a particular focus on two case studies: Convergence (2020/21) by Alexander Schubert, for String Ensemble and AI System, and In Human Memories (2024) by Francesco Vitucci, for four cellos and one electroacoustic player, whose installation was created by the Electronic Music School of the "N. Piccinni" Conservatory of Bari. The initial historical survey traces the evolving use of technology in composition, revealing how the search for new expressive possibilities has driven the development of technological tools through experimentation. The appearance of AI marks a fundamental shift: no longer just a tool, AI now poses questions and challenges to the composer. Engaging with AI becomes an exploration of an unknown entity, whose full capabilities are still largely undiscovered. This new interaction leads composers to confront AI not merely as a passive resource but as a participant in the creative process, outlining an ethical path to reclaim technology as a creative tool. In Human Memories reflects this dynamic, demonstrating how AI can serve as both inspiration and a driving force in shaping contemporary music. --- ### JAXtari: High-Throughput and Easy-to-Modify Arcade Learning Environment Authors: Quentin Delfosse, Raban Emunds, Paul Seitz, Sebastian Wette, Jannis Blüml, Daniel Kirn, Dominik Mandok, Kristian Kersting Published: 2026-10-16 Categories: cs.LG arXiv: https://arxiv.org/abs/2610.23650v1 The Arcade Learning Environment (ALE) has been a cornerstone of Deep Reinforcement Learning research, providing a standardized benchmark for evaluating general agents. However, its lack of native support for object-extraction and environment modification restricts research into evaluating whether agents learn robust causal models or rely on brittle visual heuristics. Furthermore, the ALE's dependence on CPU-based emulation creates a significant bottleneck in modern GPU-accelerated training pipelines, limiting both experimental scale and iteration frequency. In this work, we introduce JAXtari, an open-source, high-performance reimplementation of Atari 2600 environments. By leveraging massive GPU parallelization, JAXtari allows PPO agents to finish training on 100M environment steps on a single A100-GPU in under an hour. Beyond speed, it provides object-centric observations and a flexible modification system, allowing researchers to, e.g. develop neuro-symbolic methods or discover novel shortcuts and subsequently train and adapt their agents rapidly, sometimes in only minutes. --- ### Scalable long-range machine learning interatomic potentials with electrical response for atomistic material modeling Authors: Yicheng Chen, Xiaoyun Yu, Hyejeong Song, Xintian Wang, Zihui Li, Zixi Xia, Junmin Chen, Peichen Zhong Published: 2026-10-16 Categories: cond-mat.mtrl-sci, cs.LG arXiv: https://arxiv.org/abs/2610.23643v1 Accurate atomistic simulations of materials under electric fields require long-range electrostatics and electrical-response modeling at length scales beyond the reach of conventional first-principles methods. Yet these capabilities remain difficult to reconcile with the scalability of machine-learning interatomic potentials (MLIPs). Strictly local models enable efficient spatial parallelization but neglect long-range interactions, whereas existing electrostatic extensions introduce global dependencies that hinder distributed computation. Here, we overcome this limitation by analytically separating the long-range electrostatic forces and Born effective charges (BECs) into a global electrostatics calculation and strictly local derivatives of learned latent charges. This formulation preserves the spatial decomposability of equivariant MLIPs while enabling distributed evaluation of energies, forces, stresses, and BECs across different boundary conditions. Implemented as DPA4C-LR on the \texttt{nvalchemi} platform, the framework scales to 5.2 million atoms on four GPUs at approximately twice the computational cost of its short-range backbone. Trained on energies and forces, DPA4C-LR accurately infers BECs and reproduces the field-dependent infrared spectra of water. We further demonstrate its capabilities by recovering the characteristic slim polarization hysteresis of relaxor \ce{Pb(Mg_{1/3}Nb_{2/3})O3}, which is obscured by finite-size effects in smaller simulations; predicting a room-temperature field-driven ionic conductivity of $6.8 \pm 0.7$~mS/cm in amorphous \ce{Li_{0.9}NbO_{0.9}Cl_{4.1}}; and modeling interfacial thermal transport across Si/diamond interfaces with slab-corrected electrostatics. By combining long-range physics, electrical-response inference, and distributed computation, this framework extends predictive atomistic modeling toward device-relevant length scales. --- ### Diversity Without Labels: Task-Agnostic Coreset Selection Across Perception and Action Authors: Julia Hindel, Benjamin Ramtoula, Kyuhwan Yeon, Nick Heppert, Paul Newman, Daniele De Martini, Abhinav Valada Published: 2026-10-16 Categories: cs.CV arXiv: https://arxiv.org/abs/2610.23642v1 While data collection for model training is relatively easy, curating high-quality datasets from large collections is inherently challenging. Coreset estimation methods address this problem by identifying the best data subsets to train models on. However, existing approaches often require proxy model training, extensive annotations, or task-specific tailoring. Unsupervised alternatives typically exploit geometric properties in foundation model feature spaces, but are limited to pairwise image-to-image comparisons or comparisons against a fixed subset. We introduce CoreDNA, an unsupervised, task-agnostic coreset estimation approach that pools neuron activations from vision foundation models into per-image and per-dataset distributions. By comparing each image's distribution to the dataset-level distribution, CoreDNA can capture both low-level visual structure and high-level semantic information across layers. We find that high-quality samples concentrate in the tails of the last layer's activation statistics, which we validate by relating them to training-based coreset methods. Building on these findings, we develop a computationally efficient selection mechanism to construct task-agnostic, high-quality coresets for selective annotation and model training. Across image classification, semantic segmentation, depth estimation, and imitation learning for end-to-end driving and mobile manipulation, CoreDNA matches or outperforms the state of the art, demonstrating unprecedented generalization across perception and action tasks. These results highlight the effectiveness of compact activation-based representations for data selection. --- ### No-Regret Language Generation Authors: Jiaxun Li, Vinod Raman, Ambuj Tewari Published: 2026-10-16 Categories: cs.LG, stat.ML arXiv: https://arxiv.org/abs/2610.23641v1 We introduce no-regret generation, which requires a generator to make a sublinear number of mistakes uniformly over all demonstration sequences and compatible target languages. We also formulate an agnostic variant that allows arbitrary demonstrations that may not be compatible with any languages. For both settings, we introduce tree-based mistake-forcing dimensions that exactly characterize finite-horizon minimax complexity and establish necessary and sufficient conditions for no-regret generation. Our characterizations reveal a rich spectrum of generation rates: up to constant rescaling, every admissible nondecreasing profile is attainable in both the realizable and agnostic settings. Finally, we connect generation complexity to online prediction: for language classes of finite Littlestone dimension \(d\), we obtain realizable mistake bounds of \(O(d\log(1+T/d))\) and agnostic regret bounds of \(O(\sqrt{dT\log(1+T/d)})\). The realizable bound is optimal up to universal constants. --- ### Deep LLM layers expose passage-level context in EEG during silent reading Authors: Ingo Marquardt, Anthilia Alchanat, Priyanka Jain Published: 2026-10-16 Categories: cs.LG, q-bio.NC arXiv: https://arxiv.org/abs/2610.23638v1 In a previous report we showed that a decoder can recover open-vocabulary word-level information from 19-channel EEG recorded during silent reading, by aligning short EEG windows with hidden-state embeddings of the presented word taken from a large language model (LLM). That analysis contrasted two of the model's 33 layers: the context-free input embedding and one mid-depth layer. Here we ask how EEG-decodable structure is distributed over the full depth of the language model. We trained one decoder per layer of Llama-3.1-8B, with and without a causal transformer over the sequence of EEG windows. We distinguish word-level decodability from a context-tracking component. Word-level decodability peaks in early layers and declines steadily towards the last layer. In contrast, the context-tracking component of the retrieval gain rises with depth, both with and without a causal transformer. Unexpectedly, a decoder that sees a single 0.5-s window at a time tracks the text passage too. Two new controls dissect this component. A target-geometry control, a synthetic predictor made of the true target plus noise calibrated to each decoder's word-level accuracy, shows that context tracking is not a by-product of word decoding and target geometry. A control that permutes predictions across runs from either the same or from different sessions shows that most of the observed context-tracking is not caused by an EEG recording session signature. Instead, context tracking mostly follows the text. A single EEG window therefore carries passage-level information that a contextual target can absorb. The results locate word identity in the early layers of the language model, discourse state in the deeper layers, and make the choice of target layer a design dial for EEG language decoders. --- ### PresenceBench: An Exploratory Evaluation of Large Language Model Adaptation to User State Authors: Peña Arroyo Gallego, Mauro Ruiz Moreno, Alejandro Acien, Sergio Romero-Tapiador, Teresa Arroyo-Gallego Published: 2026-10-16 Categories: cs.HC, cs.AI, cs.CL arXiv: https://arxiv.org/abs/2610.23625v1 Presence-aware intelligence is the capacity of an AI system's intelligence layer to adapt appropriately to the functional state of the person it serves (here readiness, flow, and stability) while preserving task quality and safety. We introduce PresenceBench, an instrument for measuring it. PresenceBench evaluates how the intelligence layer uses a supplied state, not how that state is sensed, and breaks the capability into measurable properties. Two are tested here. Receptivity: does the model take up the state and give a better response than with no signal? Specificity: does any improvement depend on the correct state, rather than on extra structured context? A third, adaptation across states, is a required endpoint not measured in this release. A five-arm causal design answers both questions, against each model's own baseline and an equally structured mismatched-state control, scored by a correctness-gated human paired comparison. The exploratory evaluation covered four models across eighteen single-turn, six fixed three-turn and six simulated natural-conversation scenarios. Supplying a synthetic state, raw or translated into a response policy, gave human-preference point estimates above 0.5 for all four models (Claude Sonnet 5, DeepSeek V4 Pro, Gemini 3.6 Flash, GPT-5.6 Terra), with three of eight 95% intervals excluding chance. Specificity was not resolved: no per-model interval excluded chance, and the design cannot yet separate the state values from a generic instruction to adapt. Policy delivery reduced priced cost and output tokens against each model's baseline; these are resource-use changes, not efficiency gains, because correctness non-inferiority was not established. Safety-floor failures were rare (13 of 360 cases), with a non-significant excess under policy delivery. We report the instrument, the frozen results, and the controls a confirmatory study needs. --- ### Deconstructing Multi-Task Active Learning: The Paradox of Gradient Conflict and Orthogonal Decomposition Authors: Juho Jung, Seongjae Kang, Jinyoung Han Published: 2026-10-16 Categories: cs.LG arXiv: https://arxiv.org/abs/2610.23623v1 Existing multi-task active learning (MTL-AL) methods predominantly rely on linear scalarization of heterogeneous objectives such as uncertainty, diversity, and gradient conflict. We argue that this entangled formulation is fundamentally limited, as linear combination fails to capture Pareto-optimal trade-offs under conflicting objectives. We propose a shift from combination to deconstruction and introduce a principled framework that rethinks MTL-AL along three orthogonal dimensions. First, we identify the Conflict Paradox: although the optimization literature treats gradient conflict as interference to be suppressed, penalizing conflict when ranking candidates is measurably worse than not scoring it at all. We therefore decouple the two concerns: the acquisition function does not score conflict, and the resulting interference is resolved by gradient surgery during training. Second, we introduce Orthogonal Decomposition, separating acquisition into Informativeness and Exploration. This two-stage process, informativeness filtering followed by coverage maximization, prevents "diversity dilution" by excluding uninformative outliers before enforcing diversity. Third, we propose Cross-Task Mutual Information (CTMI), a covariance-based statistic of cross-task dependency computed from MC-dropout task entropies. Extensive experiments on COCO, Cityscapes, and NYUv2, including comparisons against recent MTL-AL and data-selection baselines, demonstrate consistent improvements over state-of-the-art methods, highlighting that structural rethinking of acquisition yields larger gains than incremental heuristic design. Code is available at https://github.com/Juho-Jung/multitask_active_learning_official. --- ### Stereo LWIR-RGB Automatic Labeling for Low-Light Camera-Based VRU Detection Authors: Dimitrios Bouzoulas, Pyry Jäppinen, Pejman Habibiroudkenar, Eerik Alamikkotervo, Risto Ojala Published: 2026-10-16 Categories: cs.CV arXiv: https://arxiv.org/abs/2610.23621v1 VRU detection in low-light conditions is a key component in ensuring VRU safety during the most collision prone time of the day. Currently, advances in low-light detection are limited due to the lack of large scale datasets and automated labeling methods. Most autolabeling methods use RGB detections as their labels source, which cannot be replicated in low-light conditions. To fill this gap in this paper we present an automated pipeline for low-light VRU labeling. We use LWIR-RGB image pairs and source our labels from LWIR as VRUs differentiate from the background due to their body temperature. Unlike our previous work that relied on a complex hardware setup to align the two modalities, we perform this experiment using a LWIR-RGB stereo setup, lowering hardware complexity and cost, making the pipeline easier to reproduce. We evaluate method end-to-end on a dataset collected in the capital region of Finland. Models trained on our generated labels beat both off-the-shelf baselines and models fine tuned on a hand labeled subset on F1, mAP@50 and LAMR metrics when tested on a manually labeled dataset, and the result holds for two different LWIR detectors. This shows that low-light VRU autolabeling can be made both accessible and effective without specialized alignment hardware. --- ### GRAFT-EEG: Gauge-Aware Physical Field Transport for Cross-Dataset Motor-Imagery EEG Authors: Ying Zhou, Mengting Liu Published: 2026-10-16 Categories: cs.LG, eess.SP arXiv: https://arxiv.org/abs/2610.23620v1 Cross-dataset EEG decoding often treats montage, reference, and active-sensor changes as representation shift, although they also alter the observation operator. We study target-statistics-blind transfer: fitting receives no target signal, label, or empirical statistic, and electrode coordinates only instantiate a fixed prediction operator. method removes the active-set common-reference gauge and transports each native mask through a template motor leadfield. A source-fitted spectral classifier anchors the decision; a bounded state-space-model (SSM) residual is secondary. We establish restricted identifiability and an operator-perturbation bound. Across five leave-one-dataset-out targets (253 participants), dynamic method achieved 60.865% equal-target balanced accuracy versus 52.684% for strict-clean 5/5 BIOT (hierarchical difference +8.182 points; 95% CI +6.382 to +10.196; positive in every target). On four EEGPT-clean targets, the descriptive means were 61.520% and 59.057%. In four post-hoc exact same-parent tests, FULL exceeded coordinate-basis, matched-spectrum, row-shuffled, and wrong-anatomy controls by 9.959, 9.698, 11.241, and 3.985 points, respectively (all 95% CIs excluded zero; Holm-adjusted P<0.001). A separate 60-participant same-paradigm evaluation retained a positive physical-operator contrast (+4.029 points; +2.338 to +5.773), whereas dynamic--static was inconclusive; its preregistered status is partial confirmation rather than unrestricted external validation. The SSM added +1.059 points (+0.013 to +1.892). Claims remain bounded to five binary motor-imagery domains and the tested operators. --- ### ThetreStageAI: Multi-View Retrieval-Augmented Generation for Structured Bengali Theatre Production Planning Authors: Md. Mehedi Hasan Naeem, Mst. Kamrunnahar Ruma, Shakila Sultana, Nafiza Anjum, Nasreen Jannatun Nemu, Md. Sujan Ali Published: 2026-10-16 Categories: cs.LG arXiv: https://arxiv.org/abs/2610.23618v1 Staging a theatre performance requires coordinating dialogue, actor movement, stage directions, lighting focus, cue timing, and equipment constraints within the same production plan. This work presents ThetreStageAI, a retrieval augmented framework for structured Bengali theatre production planning. Rather than retrieving from one mixed corpus, the system maintains separate scene, actor-blocking, and lighting views. Bengali text is embedded with BGE-M3 and indexed in view-specific Qdrant collections. Retrieved references are combined with a production brief and passed to Gemini, which generates schema-constrained JSON containing characters, scenes, dialogue, stage directions, blocking, lighting, and optional sound cues. Structural and semantic checks are applied before acceptance, with at most one constrained repair attempt. We also construct a human-guided, LLM-assisted synthetic dataset of 500 short Bengali theatre records that expands to 1,500 retrieval documents. Dataset diagnostics confirm consistent identifiers and valid lighting fields while also revealing substantial dialogue repetition. A paired camera-ready evaluation compares single-combined and full multi-view retrieval under matched settings. The source code, dataset, and evaluation materials are publicly available. --- ### GEMS: Molecular Structure Elucidation as Learned Search over Isomer Graphs Authors: Utku Umur Acikalin, Yingheng Wang, Goncalo J. Gouveia, Tyler J. Schwertfeger, Frank C. Schroeder, Carla P. Gomes Published: 2026-10-16 Categories: cs.AI arXiv: https://arxiv.org/abs/2610.23617v1 Identifying small-molecule structures from molecular formulas and tandem mass spectra remains a bottleneck in untargeted metabolomics and natural-product discovery. Even when database retrieval or de novo prediction misses the exact structure, the resulting candidates can provide starting points for exploring nearby structures through a few bond edits. We introduce GEMS (Guided Editing of Molecular Structures), which refines candidates through learned 2-switch bond edits that preserve the molecular formula and each atom's total valence. GEMS uses MIST, a pretrained spectrum-to-fingerprint model, to predict fingerprint probabilities once per spectrum. A neural policy compares each candidate's fingerprint with these predictions, links present features to the atoms and bonds that generate them, and scores all valid edits in one forward pass. The policy is trained on shortest edit paths between structures, so it can be pretrained on PubChem isomers without spectra and then fine-tuned with spectrum-predicted fingerprints. At inference, GEMS searches from multiple initial candidates matching the supplied molecular formula and ranks candidates by fingerprint agreement and multi-start consensus. With the target structure excluded from PubChem initialization, three edit rounds achieve Hit@1/Hit@10 of 29.64%/50.81% on NPLIB1 and 30.22%/41.33% on MassSpecGym, exceeding published baselines under comparable settings. On NPLIB1, GEMS achieves higher identification accuracy in about 3.3 minutes end to end, compared with 26.2 hours for our FRIGID run. GEMS also improves identification from generated candidates, including those refined by FRIGID. Experiments with exact target fingerprints identify fingerprint prediction error as a bottleneck in target recovery and final ranking. Deeper search recovers additional targets without retraining, with larger gains from weaker initial candidates. --- ### The Sign Code: The Hidden Binary Nature of Deep Networks Authors: Niclas Göring, Shuofeng Zhang, Roi Holtzman, Yoonsoo Nam, Ard A. Louis Published: 2026-10-16 Categories: cs.LG arXiv: https://arxiv.org/abs/2610.23616v1 Every neuron in a trained network outputs a number that can be separated into its sign (positive or negative) and its magnitude (absolute size). We call the set of all neuron signs the \textit{sign code}. This raises a key question: does classification depend more on the sign code or on the magnitudes? For overparameterized vision classifiers with sign-preserving activations, the answer is overwhelmingly the sign code. A $k$-nearest neighbor classifier operating only on this binary sign code achieves accuracy comparable to the full model across different architectures, activation functions, and datasets. Moreover, replacing all magnitudes in every layer with random noise while keeping the original signs barely changes residual network performance. In contrast, flipping just 20\% of the signs reduces accuracy to chance. Models generalize well when same-class inputs share similar sign codes. When this overlap is disrupted by a regularizer, training accuracy remains high, but test performance collapses. A margin decomposition helps explain these results: the sign-only contribution to each logit grows as $\sqrt{d}$ in network width, dominating the magnitude residual. The natural unit of analysis for these networks may not be through their continuous weight space but instead through their discrete sign codes. --- ### Beyond Action Prediction: Learning Search-Free Planning in Latent Space Authors: Guanchen Du, Jinhao Du, Yanpeng Sun, Shan Zhang, Na Zhao, Zechao Li Published: 2026-10-16 Categories: cs.RO, cs.AI arXiv: https://arxiv.org/abs/2610.23615v1 Latent world models learn action-conditioned state transitions in a compact representation space, providing a predictive model for visual planning and control. However, planning with these models typically requires costly test-time optimization over action sequences. We propose to replace this online search with a goal-conditioned inverse model that directly predicts actions from current and goal observations. A key challenge is that action prediction alone neither ensures that goal representations encode action-relevant information nor that predicted actions are consistent with the learned latent prediction model. We introduce Hindsight Inverse JEPA Planning (HIJP), a search-free framework that addresses both aspects during training. HIJP relabels future observations along offline trajectories as goals, providing action supervision across multiple temporal horizons. By retaining gradients through the goal representations, this supervision directly shapes the visual representation. HIJP further feeds predicted actions into the latent predictor model and aligns the resulting next-state representations with encoded successor observations, enforcing action-transition consistency. At inference time, the inverse dynamics model and latent predictor models are applied alternately to generate multi-step action sequences without online search. Across four visual control tasks, HIJP achieves 86.8% overall success and a 1.9-16.5x planning speedup over search-based planning. --- ### Trajectory-Triggered Sparse Collaborative Decoding Authors: Amit Ranjan Trivedi, Sruthikesh Surineni, Kyong Ho Lee, Mohit Mittal, Steven Li, Rex Hill Published: 2026-10-16 Categories: cs.CL, cs.LG arXiv: https://arxiv.org/abs/2610.23613v1 Dense assistance methods such as speculative decoding repeatedly verify draft tokens with a target model, keeping target computation coupled to generation. We study sparse collaborative decoding, where a draft model generates most tokens and a target model is invoked only for short corrective segments. The key problem is deciding when to call the target. We show that local uncertainty is a weak trigger: entropy, negative log-likelihood, logit margin, top-$k$ mass, and velocity features give near-chance failure prediction, with AUC 0.50-0.53. In contrast, aggregate trajectory-shape features over the draft confidence path reach AUC 0.86 using a lightweight logistic predictor trained only from draft correctness labels. This induces a trigger-timing tradeoff: early calls have more leverage but weaker evidence, while later calls have stronger evidence but less ability to redirect the trajectory. Used online, learned sparse collaborative decoding improves over the best open-loop schedule by +5.3 to +20.0 accuracy points across five GSM8K draft-target pairs. On MATH-500, it gains 7.0-13.5 points when early prediction and target competence align, but loses when either is weak. On StrategyQA, it improves over open-loop by 5.0-8.3 points, showing transfer beyond arithmetics. --- ### SIT-Eval: Situation-Based Interactive Turn-by-Turn Evaluation of Conversational Agents Authors: Ravil Mussabayev, Saveliy Zavertyaev, Dmitriy Parpulov, Andrey Savchenko Published: 2026-10-16 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2610.23612v1 The rapid proliferation of conversational AI agents necessitates evaluation methods that assess performance across diverse dialogue scenarios, utilizing multiple situational criteria, as appropriate behavior depends on context and encompasses various aspects. Existing dialogue benchmarks typically use static natural conversations, which underrepresent rare situations---a critical safety gap, as low-probability events become frequent failures at production scale. However, limited sample sizes force a choice: either maintain natural dialogues for realism or turn to adaptive situation sampling for comprehensive yet statistically reliable coverage. We introduce SIT-Eval, an adjustable framework for evaluating situation-based conversational agents. We formalize an agent's desired behavior as situational rules: if situation is P, follow rules Q. At each conversation turn, an LLM-interlocutor enacts a uniformly sampled situation. A context-aware LLM-judge leverages all conditionally applicable rules as criteria to evaluate the agent's reply for that situation. The judge is validated using human-verified synthetic golden dialogues. Benchmarking 8 LLMs on 1292 conversation turns reveals failure modes hidden by aggregate metrics: models with similar overall scores vary widely in their adherence to individual rules. All models struggle most with ambiguity clarification and Socratic questioning. --- ### How Many, but How Sure? Class Competition as a Confidence Proxy in Multi-Class Density-Based Counting Authors: Villanelle O'Reilly, Georgios Leontidis, Marc Hanheide Published: 2026-10-16 Categories: cs.LG arXiv: https://arxiv.org/abs/2610.23611v1 Multi-class density estimators can count objects in crowded scenes, but their class-wise outputs offer no indication of how reliably density mass has been allocated. We introduce a training-free confidence proxy based on local competition between the strongest class channels. The resulting prominence measure requires only the predicted density maps, with no retraining or access to model internals. We test whether residual competition is diagnostic of cross-class leakage by attenuating contested density at inference time. On DSACA predictions for the Hicks et al. dataset, this reduces overall MAE from 0.920 to 0.846, suggesting that class competition contains useful information about count reliability. --- ### Not Every Mention Is a Claim: Rethinking Object Hallucination Detection in Vision-Language Models Authors: Hsin-Ling Hsu, Neeraja Kirtane, Kuan-Hao Huang Published: 2026-10-16 Categories: cs.CV, cs.CL arXiv: https://arxiv.org/abs/2610.23610v1 Object hallucination in vision-language models (VLMs) is usually measured by checking each object a caption names against the image. Most benchmarks and detectors alike treat every object mention as a claim that the object is present. This assumption fails when a caption names an object without asserting it, as in "a bench for people to sit on", which mentions people without claiming that any are in the image. We call such mentions non-assertive. Treating them as hallucinations inflates the error rates attributed to models that use such language. To separate them from real hallucinations, we introduce NAME-Bench (Non-Assertive Mention Evaluation Benchmark), 25,497 object mentions from captions that five VLMs generate for COCO images. Each mention is labelled concrete (asserted and present), hallucinated (asserted and absent), or non-assertive, and non-assertive mentions are further divided into four subtypes: depictive, purpose, negation and hedged. All the hallucination detectors that we evaluate flag non-assertive mentions as hallucinations, at rates between 14% and 80%. Adding non-assertive mentions to detector training as negatives largely corrects this on LLaVA-1.5. On the four newer VLMs, a comparable correction takes five to ten times more data, and the detectors then catch fewer real hallucinations. These findings show that current detectors conflate mentioning an object with asserting it, and that this conflation is harder to undo on the four newer VLMs than on LLaVA-1.5. Hallucination evaluation should therefore separate what a caption asserts from what it merely mentions. --- ### ADA 2025: A Multimodal Dataset for Water-Deficit Stress Detection in Sugar Beets Authors: Piper Powell, Robert Franz, Amos Smith, Isaak Ihorst, Joachim Hertzberg, Martin Atzmueller Published: 2026-10-16 Categories: cs.CV arXiv: https://arxiv.org/abs/2610.23609v1 Research and development into artificial intelligence (AI) and machine learning (ML) applications in crop management are hindered by a lack of suitable public datasets. Whereas applications in disease and pest detection benefit from a comparatively higher amount of available data, applications like water-deficit stress detection often have very few if any public datasets available. Moreover, most datasets that are available are single-modality and frequently RGB-only, leaving the early detection and enhanced performance offered by other methods, and particularly by multimodal approaches, out of reach. Datasets with accurate metadata, higher numbers of plants, and smaller intervals between data points are also scarce. To address this, we publicly release ADA 2025, a >500 GB comprehensive multimodal dataset of young sugar beet plants under healthy and water-deficit conditions combining thermal, hyperspectral, RGB, and depth imagery along with relevant environmental and metadata and both relative water content (RWC) and leaf thickness as ground truth sources. Exemplary results showcase the dataset's validity in terms of differences detectable between stressed and healthy plants. --- ### When Magnitude Suffices: Reconstructing Time-Domain-Gated Microwave Signals Without Phase on a Thoracic Phantom Authors: Maciej Cichoń, Patrycja Gołaś, Maksymilian Bartosik, Angelika Walczak, Agata Bielecka-Dąbrowa, Maciej Ślot Published: 2026-10-16 Categories: cs.LG, physics.app-ph arXiv: https://arxiv.org/abs/2610.23608v1 Time-domain gating separates the arrivals that form a microwave S-parameter measurement. It requires the measurement of complex values, so this application almost always uses a vector network analyzer (VNA). We show that, on one calibrated measurement configuration, the gated magnitude the VNA would give can be reconstructed from a magnitude-only sweep of a spectrum analyzer with a tracking generator (SA). We select the gate center for separability of the measured states, correct each instrument once by a per-bin subtraction of a reference state, and train a fully connected network on sweeps from both instruments. The test object is a torso phantom with lung-shaped compartments whose filling takes eleven dielectric states, each held out in turn. On $S_{11}$ the reconstruction from SA input reaches a shape error of 0.08 dB$^2$ with a state ordering of 0.81. This matches the reconstruction from VNA input. On $S_{21}$ it reaches 0.82 dB$^2$ and keeps the ordered groups of the states. An explicit phase prediction performs much worse, because gating needs the phase to about 0.2 rad RMS. Across eleven measurement configurations the dependence of phase on magnitude is strong for reflection and weak for transmission. This explains the difference between $S_{11}$ and $S_{21}$. Where the magnitude carries enough of the phase, a scalar instrument gives the gated response without any phase measurement after installation. On the open UM-BMID breast data set, with whole phantoms held out, the shape error of the network is 30 percent lower than that of a per-bin linear map. --- ### Learning to Discriminate Scene Structures Makes Self-Supervised Depth Learning Scalable Authors: Mengtan Zhang, Yuchuan Cui, Yu Ma, Zizhan Guo, Wei Ye, Rui Fan Published: 2026-10-16 Categories: cs.CV arXiv: https://arxiv.org/abs/2610.23600v1 Self-supervised monocular depth estimation is appealing for its potential to scale with unlabeled data, yet existing methods often degrade when trained on diverse scenes. This study finds that, under self-supervision, models tend to encode dataset-specific structural biases rather than transferable scene geometry, leading to cross-domain optimization conflicts. To address this issue, this study proposes the MaSS framework, aiming to make self-supervised depth learning scalable to diverse data. MaSS introduces a dedicated structural representation pathway explicitly decoupled from contextual features, which is guided by a novel scene-structure discrimination objective to learn generalized, layout-aware structural representations for depth decoding. Trained and evaluated on four diverse datasets, both individually and jointly, MaSS substantially outperforms prior self-supervised methods and is the first to consistently benefit from mixed-domain training. Zero-shot evaluation on eight additional datasets further demonstrates its remarkable generalizability, particularly on those entirely out-of-distribution scenes. The source code is publicly available at https://mias.group/MaSS. --- ### From Trajectories to Differentiated Skills: Cognitive Differentiation for Self-Improving Agents Authors: Yewei Zhao, Derong Xu, Wei Chen, Yin Xu, Peilun Zhou, Xiaohan Wang, Guojun Yin, Wei Lin, Mingjun Xiao Published: 2026-10-16 Categories: cs.AI arXiv: https://arxiv.org/abs/2610.23599v1 Large language model (LLM) agents tackle complex tasks through reasoning and interaction with external environments. Agent skills provide persistent, transferable capabilities that guide task planning, inform contextual decisions, support tool use, and enable executable procedures. Yet existing skill-evolution methods largely organize knowledge around skill artifacts, overlooking the functional heterogeneity within individual skills and its implications for knowledge maintenance and use. Inspired by cognitive science, we propose Cognitive Differentiation in Skill Evolution (CDSE), which evolves a unified skill through four functionally distinct yet coordinated components: Meta, Experience, Tool, and Script. CDSE extracts evidence-grounded semantic gradients from successful and failed trajectories, consolidates them by functional role, and applies component-specific updates and validation with cross-component coordination and adaptive invocation. An analysis of 500 top-ranked skill packages shows that 96.8% contain knowledge spanning at least two functional roles. Across six benchmarks, three agent models, and two execution harnesses, CDSE achieves the highest scores in all 36 settings, improving over the Base agent by 7.6 percentage points and the strongest competing method in each setting by 2.6 points on average. Ablation studies further support the importance of functional differentiation and adaptive invocation. --- ### RL-Guided Temporal Localization for Dual-Channel Retrieval in Long-Horizon Agent Memory Authors: Dabin Sheng, Zhe Wu, Jinming Zhao, Junliang Xing, Yuanchun Shi Published: 2026-10-16 Categories: cs.AI arXiv: https://arxiv.org/abs/2610.23598v1 Time-aware retrieval is crucial for long-horizon LLM agents to effectively leverage ever-growing long-term memory. Existing methods focus on lexical or semantic similarity, leading to target memories being crowded out by temporally mismatched yet semantically similar candidates. Even time-aware approaches inject temporal cues only during memory organization or answer generation, overlooking temporal constraints at the retrieval stage, where candidate selection occurs - the root cause of the displacement. Inspired by the temporal contiguity effect in human episodic memory, we propose TIDER (Temporal Interval-driven Dual-channel Evidence Retrieval), a time-aware agentic retrieval framework. Central to TIDER is a reinforcement learning (RL)-trained temporal interval locator that infers the temporal windows in which the evidence lies. RL optimizes window localization as a discrete, non-differentiable retrieval decision, using composite rewards to convert sparse answer supervision into fine-grained retrieval feedback. These temporal windows first define the temporal retrieval channel; the retrieved candidates are then merged with results from the semantic channel and reranked by proximity to the windows. This temporal grounding enables TIDER to retrieve temporally aligned evidence that would otherwise be drowned out by the volume of semantically similar yet temporally mismatched memories. With a lightweight 1.7B locator, TIDER achieves state-of-the-art accuracy on both long-horizon memory benchmarks - 61.52% on LifeBench (+8.82%) and 78.96% on LoCoMo (+3.66%) - which indicates that temporal cues are crucial for pinpointing target evidence amid a vast pool of semantically similar historical events in long-term daily-life memory. --- ### Boundary Detection for High-Dimensional Point Clouds Authors: Abigail Hickok, Valerio Galanti, Osama Al Dalahmah, Andrew J. Blumberg Published: 2026-10-16 Categories: cs.LG, cs.CG, stat.ML arXiv: https://arxiv.org/abs/2610.23590v1 We introduce two new methods for boundary detection in high-dimensional point cloud data, intended to work in situations with non-uniform density, high ambient dimension, and substantial noise. Our first method fits a separating hyperplane on local neighborhoods after doing local dimension reduction---this method works best when one can reliably estimate the intrinsic dimension of the data. Our second method measures the directional uniformity of local neighborhoods, and has the best overall performance on synthetic data. In addition to pointwise numerical boundary scores, our algorithms yield an estimated normal vector at each boundary point. We validate our methodology on synthetic data as well as two real examples: we analyze spatial transcriptomics data, where we show that we can detect the boundary of a tumor, and we also study LLM activation geometry. --- ### SemPurify: Semantics-Aware Data Purification for Black-box Backdoor Defense Authors: Jiawei Yu, Syed Irfan Ali Meerza, Yi Wu, Amir Sadovnik, Jian Liu Published: 2026-10-16 Categories: cs.CV arXiv: https://arxiv.org/abs/2610.23589v1 Backdoor attacks have emerged as a critical security threat in machine learning, where adversaries embed triggers within poisoned training samples to manipulate a model's predictions on targeted inputs while leaving its general behavior unaffected. This stealthy nature makes backdoor attacks difficult to detect, as the model functions normally until activated by a specific malicious trigger. Current purification-based defenses use image transformations (e.g., blurring) to disrupt trigger patterns but often degrade image quality. Even with advanced restoration techniques like diffusion models, these approaches can still alter visual or semantic information, reducing utility for downstream tasks. To address this, we propose the first semantics-aware data purification framework for black-box backdoor defense, SemPurify, which explicitly incorporates multi-level semantics (i.e., low-level semantic embeddings and high-level image tags) within a text-to-image (T2I) framework to guide image restoration during purification. Using this semantic information, the T2I model subsequently refines the transformed samples, helping preserve visual details and semantic information during restoration. Extensive experiments spanning multiple backdoor scenarios, i.e., classification models, CLIP-style encoders, and generative models, demonstrate that SemPurify reliably removes backdoors while effectively preserving semantic information for downstream tasks. --- ### Enhancing RAG-based Medical VQA with Gated Multimodal Representations Authors: Daniela Jimenez-Pano, Hugo Jair Escalante, Delia Irazú Hernández-Farías Published: 2026-10-16 Categories: cs.CV, cs.CL arXiv: https://arxiv.org/abs/2610.23588v1 Medical Vision-Language Models (Med-LVLMs) have shown strong performance in Medical Visual Question Answering (MVQA), yet they may still produce factually inconsistent responses. Retrieval-Augmented Generation (RAG) can mitigate this issue by incorporating external clinical evidence; however, retrieval is often guided by a single modality, potentially overlooking the joint semantics of medical images and clinical questions. We introduce a multimodal RAG framework that learns a joint query representation through a Gated Multimodal Unit, enabling evidence retrieval to leverage information from both modalities while keeping the underlying Med-LVLM frozen. We evaluate the proposed approach on benchmark datasets. Results show that the multimodal query representation consistently outperforms a reference Med-LVLM and remains competitive with state-of-the-art methods. Our findings demonstrate the potential of multimodal query representations for more effective evidence retrieval in Medical VQA. --- ### Analyzing Occupational Gender Bias across the LLM Training Lifecycle Authors: Manthan Gadhia, Alexander Hoyle, Karolina Stańczak Published: 2026-10-16 Categories: cs.CL arXiv: https://arxiv.org/abs/2610.23587v1 Post-training is widely assumed to reduce the social biases language models inherit from pretraining, yet direct evidence for this assumption remains mixed. In this paper, we examine how gender disparities evolve across the language modeling pipeline. In particular, we study occupational gender bias in OLMo 3, a fully open language model family, across its post-training stages by prompting each checkpoint to generate open-ended narratives about synthetic personas given their occupation and education. We consider two settings, a gender-given setting in which the profile states the persona's gender explicitly and a gender-assumed setting in which it must be inferred by the model. We find that when gender is stated, post-training increases semantic diversity for male-coded profiles, whereas for female-coded profiles the increase is smaller after SFT and not significant after DPO and RLVR. When gender is assumed by the model, post-training affects narratives about female and male personas similarly. The model's own gender assumptions, however, show a consistent female-preferential skew relative to United States labor-force demographics, except for heavily male-dominated occupations. Every successive post-training stage monotonically increases the odds of attributing a female label over the base model. A provenance analysis of the supervised fine-tuning corpus finds only a modest 7.7% female-token surplus, too small to explain the size of these downstream effects. The project code is available under: https://github.com/manthangadhia/llm-lifecycle-bias. --- ### MARS: Memory-Aware Routing for Session-Level LLM Assistants Authors: Minghan Wang, Jinhang Zuo, Zhongxiang Dai, Zhiyong Wang, Fang Kong Published: 2026-10-16 Categories: cs.CL arXiv: https://arxiv.org/abs/2610.23583v1 Long-term LLM assistants must repeatedly decide what prior context to expose and which model should answer as an interaction evolves. Existing memory systems and model routers usually make these decisions separately, although memory construction changes the evidence, distractors, and difficulty of the resulting input. We introduce MARS, a memory-aware routing framework that uses session state as a shared control interface between memory construction and model selection. A Query and Topic Manager (QTM) estimates compact decision-relevant state that guides a fixed state-conditioned memory policy and an online router, enabling each turn to receive an appropriate memory view and a suitable model under a quality-memory-cost trade-off. We evaluate MARS on RealMem and MultiChallenge with 16 system configurations each, covering long-history memory-driven interaction and bounded sessions with structured context interference. On RealMem, MARS preserves 98.2% of full-context GPT-4o's QA while using only 14.5% of its generation API cost, including both generation and controller calls for MARS, and achieves the strongest measured memory-recall and retrieval scores among compared systems. On MultiChallenge, MARS trails the strongest static system overall but approaches its performance at lower recorded cost on a mechanism-focused subset. Together, these results support a favorable quality-memory-cost trade-off for long-history, memory-dependent interaction. --- ### CLEAR: Cooperative Latent Flow for Visual Counterfactual Explanations with Accurate Region Preservation Authors: Yan Zeng, Changlu Guo, Oskar Kristoffersen, Anders Nymark Christensen, Morten Rieger Hannemose, Anders Bjorholm Dahl Published: 2026-10-16 Categories: cs.CV arXiv: https://arxiv.org/abs/2610.23577v1 Visual counterfactual explanations (VCEs) should alter a classifier's prediction while remaining realistic and preserving decision-irrelevant content. Image-space methods provide high-quality counterfactuals but are computationally costly, whereas latent-space methods are more efficient but often less faithful. We identify two sources responsible for this gap. First, latent generation recovers a natural sample from a perturbed representation, whereas classifier guidance pushes it toward a target class, so their update directions may oppose. Second, VAE reconstruction can alter regions that should remain unchanged. We introduce CLEAR, a latent Flow Matching framework for VCEs. CLEAR uses Flow Matching's explicit instantaneous velocity as a reference for the generative direction to decompose classifier guidance into cooperative and conflicting components. It preserves the cooperative component and adaptively attenuates only the conflicting one according to conflict strength. We further analyze conflict frequency and strength throughout generation to reveal how classifier guidance interacts with the generative process. CLEAR also dynamically identifies editable regions, constrains latent modifications, and directly preserves factual content outside them to reduce VAE-induced changes. Across five bidirectional tasks on two datasets, CLEAR retains latent-space efficiency while achieving counterfactual quality comparable to or better than strong image-space baselines, with a 2.5x speedup over the strongest baseline. Project page: https://yan-zen9.github.io/CLEAR/ --- ### Darwin-Claw: Environment-Adaptive Experience Consolidation for Recursively Self-Evolving Agents Authors: Mingze Yin, Xiaohan Wang, Dian Li, Longfei Luo, Jiahong Yan, Yiheng Zhu, Chang-Yu Hsieh Published: 2026-10-16 Categories: cs.AI arXiv: https://arxiv.org/abs/2610.23574v1 While Large Language Model (LLM) agents have demonstrated remarkable capabilities across complex tasks, they predominantly operate in isolated sessions, lacking the ability to continually learn from historical interactions or adapt to user-specific environments. Existing self-evolution paradigms often indiscriminately inject raw interaction trajectories into memory or skill modules, leading to obscured behavioral patterns and compromised update controllability. Furthermore, as foundational LLM reasoning improves, the bottleneck for agentic evolution shifts from accumulating generic knowledge to effectively managing situational experience. To address this, we propose Darwin-Claw, an environment-adaptive self-evolving framework that systematically distills episodic experiences to drive the comprehensive evolution of operational skills and durable memory. Theoretically grounded in an iterative agent-actor and evolver-critic co-adaptation, Darwin-Claw introduces a Hierarchical Experience Consolidation System that extracts fine-grained error attributions and generalizes them into dynamically updated prototype vectors. This structured situational experience serves as interaction patches for context-sensitive perception and explicit optimization signals for skill refinement. Practically, Darwin-Claw is a readily deployable, automated framework compatible with mainstream agent systems without requiring manual curation. Extensive experiments on WildClawBench demonstrate that Darwin-Claw yields substantial performance improvements, establishing a controllable and stable pathway for the continual, user-centric adaptation of autonomous agents in real-world ecosystems. --- ### BMuon: Budgeting Muon's Spectral Geometry for Structured Compressibility Authors: Rida Assalouh, Tom Jacobs, Rebekka Burkholz, Yani Ioannou, Umut Şimşekli Published: 2026-10-16 Categories: cs.LG arXiv: https://arxiv.org/abs/2610.23571v1 Muon has emerged as a strong optimizer for neural network training and admits an interpretation as steepest descent under a spectral-norm geometry. Yet we find that Muon-trained Transformers can degrade sharply when their MLP hidden channels are pruned after training. To reconcile Muon's optimization performance with structured compressibility, we introduce Budget Muon (BMuon), which modifies Muon's spectral geometry so that hidden channels across blocks compete for a shared allocation budget. The resulting oracle is exactly normalized steepest descent under a family of norms that interpolate between Muon's geometry and structured sparse geometries. Our principal sparsity control is an interpretable global channel-allocation budget, rather than a setting-dependent sparsity-penalty coefficient. For linear factorization models, we show that Muon admits minimum-cost exact representations with redundant hidden width, whereas under the BMuon oracle geometry and a sufficiently scarce budget, every minimum-cost representation is width-minimal. For practical training, we derive a minorization-maximization surrogate for the exact allocation problem. On ImageNet-1K with ViT-S/16, BMuon improves over Muon under gradual structured MLP pruning, and its learned allocations provide a more effective pruning criterion than weight magnitude. On a 1B-parameter LLaMA-style model trained on FineWeb, BMuon improves over Muon under post-training MLP channel pruning. --- ### A general recipe for data-adaptive anytime-valid envelopes for iterative algorithms Authors: Soham Bonnerjee, Aaditya Ramdas Published: 2026-10-16 Categories: stat.ML, cs.LG, math.OC, stat.ME arXiv: https://arxiv.org/abs/2610.23570v1 Most convergence guarantees for iterative stochastic algorithms are stated at a fixed iteration count and do not justify stopping at a data-dependent time. Recent work has constructed anytime-valid envelopes that do, but deterministic envelopes are typically too conservative for practical stopping, and existing data-adaptive ones are derived case-by-case, each coupling a specific algorithm to a specific noise model and boundary construction. This paper provides a general recipe for constructing such anytime-valid data-adaptive boundaries for a large class of iterative algorithms. Performance measures of many of these algorithms satisfy a one-step recursion that reduces the entire error analysis to a single general requirement: time-uniform concentration for predictably weighted sums of noise. We formalize this through a strongly adapted sub-$ψ$ condition, and derive off-the-shelf boundary constructions via continuous and discrete mixtures, stitching, and product betting martingales. We show how these boundaries can be directly coupled with a separate algorithmic analysis that derives the one-step recursion to provide a class of readymade anytime-valid envelopes across diverse settings including stochastic gradient descent (SGD), temporal difference learning, and Bures-Wasserstein SGD. In particular, for SGD, our framework either recovers or significantly sharpens every existing envelope we are aware of. Our constructions come with explicit implementation details and are substantiated by numerical experiments. --- ### From Adult Renal Masses to Pediatric Wilms Tumors: Evaluating Cross-Domain Generalization of Kidney Lesion Segmentation on MRI Authors: Sarah de Boer, Myrthe A. D. Buser, Hartmut Häntze, Jawed Nawabi, Lisa C. Adams, Alida F. W. van der Steeg, Marry M. van den Heuvel-Eibrink, Annemieke S. Littooij, Keno K. Bressem, Mathias Prokop, Alessa Hering Published: 2026-10-16 Categories: cs.CV arXiv: https://arxiv.org/abs/2610.23568v1 Deep learning segmentation models for pediatric oncology are often trained on small, single-center cohorts, whereas much larger datasets exist for adult diseases, such as kidney cancer. Whether models trained on adult renal lesions generalize to pediatric Wilms tumors remains unclear. We trained a segmentation model on single-sequence T1- or T2-weighted adult kidney MRI and evaluated it on a held-out adult test set and a pediatric Wilms tumor cohort, where it was compared with a pediatric-specific nnU-Net. Segmentation performance was quantified using the Dice similarity coefficient (DSC) and 95th-percentile Hausdorff distance. On the adult test set, the model achieved a median DSC of 0.86 for renal lesions. On the pediatric cohort, the adult-trained model achieved a median DSC of 0.85 compared with 0.90 for the pediatric-specific model. Despite differences in age, anatomy, and tumor type, the adult-trained model showed promising performance on pediatric Wilms tumors, with only a modest difference in median DSC compared with the pediatric-specific model. However, non-inferiority could not be shown, highlighting the value of pediatric training data. In future work, a fine-tuning strategy could be explored. --- ### Beyond Expected Values: Risk-Sensitive Planning with Distributional Monte-Carlo Tree Search Authors: Quynh Trinh, Trung Nguyen, Nam Nguyen, Tuan Dam Published: 2026-10-16 Categories: cs.LG, stat.ML arXiv: https://arxiv.org/abs/2610.23565v1 Monte-Carlo Tree Search (MCTS) typically stores only scalar mean estimates, which makes it poorly suited for tail-risk objectives such as conditional value-at-risk (CVaR). We introduce Categorical Thompson Sampling with Optimism (CATSO) and Particle Thompson Sampling with Optimism (PATSO), two distributional MCTS algorithms that maintain finite-support empirical backup laws at Q-edges. CATSO represents Q-edge laws with categorical atoms and Dirichlet Thompson sampling, while PATSO represents them with adaptive particles. Both algorithms select actions using the CVaR of a Thompson-sampled Q-edge law plus a polynomial optimism bonus, and propagate scalar continuation values through visit-weighted averages of child Q-edge CVaR scores. For the finite-depth nested CVaR planning target, we prove an $O(n^{-1/2})$ bound on the root value error after $n$ rollouts, which upper-bounds the risk-sensitive simple regret. CATSO adds a fixed-grid discretization bias, while capped PATSO adds a tunable Wasserstein compression bias. Experiments on stochastic planning benchmarks show that distributional Q-edges improve lower-tail return and reduce catastrophic outcomes when mean-optimal and risk-sensitive behavior differ. --- ### LiveSync: Iterative Self-Training for Lip-Sync Generation Authors: Huaishao Luo, Huan Zhang, Guanbo Bao, Yinglu Liu, Youzheng Wu, Xiaodong He Published: 2026-10-16 Categories: cs.CV arXiv: https://arxiv.org/abs/2610.23558v1 Audio-driven lip synchronization aims to match target speech through localized mouth and jaw articulation while preserving subject identity, head pose, and surrounding facial context. Naturally recorded videos provide synchronized reconstruction targets but rarely offer pairs that are matched in identity and pose yet differ primarily in speech-related articulation. Mask-based training enables reconstruction from these videos, but random-reference inputs differ from the aligned self-reference used at inference. We present LiveSync, an iterative self-training framework in which a single mask-based model generates pseudo pairs and is refined for deployment. The model is first trained on a curated real-video corpus augmented with externally generated clips covering under-represented visual conditions. A frozen checkpoint then uses independently sampled audio to generate mouth-edited references from real-video frames, copying pixels outside the edit mask from each source frame. After confidence filtering, these references are paired with aligned masked inputs and original real-frame targets. The retained pseudo pairs are mixed with real training data to refine the same model from its current weights, updating training supervision without adding inference stages. Although we instantiate the framework with a single-pass GAN-based model, the self-training strategy could in principle also be adapted to diffusion-based models. Experimental results show that LiveSync achieves leading synchronization performance on HDTF and competitive performance compared with the baselines on the AIGC-LipSync subset, while maintaining low inference latency. ---