# Research Papers — 2026-10 · bundle page 4 of 156 > Full-text markdown of papers #151–200 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/4 [← Previous bundle](https://aipod.dev/papers/archive/2026-10/page/3.md) [Next bundle →](https://aipod.dev/papers/archive/2026-10/page/5.md) ### Routing Policy Optimization for Online Diffusion Reinforcement Learning Authors: Qiuhong Shen, Yiding Yang, Yuxin Zhang, Xinwei Huang, Minxuan Lin, Bo Liu, Hao Kang, Haotian Yang, Yizhi Wang, Chongyang Ma, Xinchao Wang Published: 2026-10-15 Categories: cs.CV arXiv: https://arxiv.org/abs/2610.22282v1 Reinforcement learning can align visual generators with reward signals beyond their pretraining objectives, but existing score-function methods typically introduce stochasticity by perturbing the reverse denoising trajectory, leading to costly rollout-based optimization. We instead ask whether stochastic exploration can be introduced inside the generator while preserving its original sampler. Routing Policy Optimization (RPO) answers this question by turning attention routing into a stochastic exploration action. At a few selected denoising levels, each query block keeps its own block and samples further key blocks in proportion to the generator's native attention mass, or keeps dense attention. The sampled route perturbs the velocity of that transition, and a Gaussian likelihood around the dense velocity scores the perturbation with a score-function objective. RPO further uses a controlled two-rollout estimator: two rollouts share the same prompt, initial latent, routed denoising levels, and dense anchor layers, differing only in their sampled attention routes. Their reward difference provides an advantage signal for the routed transitions. Because the likelihood differentiates the dense velocity itself, the learned updates remain effective after routing is removed. At inference, RPO restores dense attention and retains the original deterministic sampler. Experiments on text-to-image and text-to-video generation show that RPO improves generation quality while achieving nearly 4x higher training efficiency than Flow-GRPO-style methods. Please see our project page at https://velox-world.com/rpo for visual results. --- ### LEASH: Uncertainty-Aware Weak-to-Strong Escalation and Oversight for Computer-Use Agents Authors: Qingni Wang, Changdae Oh, Yuzhe Yang, Sharon Li, Jiachen Yang, Ang Li, Xin Eric Wang Published: 2026-10-15 Categories: cs.AI arXiv: https://arxiv.org/abs/2610.22280v1 Computer-use agents increasingly tackle long-horizon tasks, where running a frontier model at every step is expensive and a small open-weight model alone often fails. Weak-to-strong escalation offers a middle ground where a weak agent acts by default and hands control to a strong one only when needed. The core challenge is deciding when to escalate, which requires an uncertainty signal cheap enough to check at every step. However, even after handoff, the strong agent can still get lost, but a black-box API that the strong agent is usually served exposes no internals to monitor. To this end, we propose LEASH, a training-free framework that unleashes the strong agent only when the weak agent's latent uncertainty signals it is needed, and keeps it on a leash afterward. LEASH has three layers: (1) initial task uncertainty for early routing, (2) step-level uncertainty for runtime escalation, and, after handoff, (3) a shadow observer for continued oversight of the strong black-box agent. Here, we use latent uncertainty as a fuel for all three layers. Specifically, we compare the weak model's hidden states with those from its past successful and failed trajectories, and a state that looks more like the failures signals higher uncertainty. During weak-agent execution, our uncertainty estimator reuses activations already produced during action generation, so it needs no extra forward pass or dedicated monitor training. Experiments on OSWorld across four contemporary model backbones show that LEASH consistently improves over standalone weak agents and outperforms random escalation by 6.55--11.94 points at matched escalation rates. With Qwen3.5-35B-A3B, LEASH achieves 73.66% task success at $0.4117 per task, matching always-on Claude Opus 5 (72.90%) while reducing inference cost by 19.2%. --- ### MJev-Spatial: System One Spatial Intelligence via Native Attention Matching Authors: Luozheng Qin, Yuhan Zhu, Changhai Zhou, Qian Qiao, Jun Gao, Yuhua Zhou, Harry Zhang, Xiangyu Zeng, Yuxuan Liu, Kaijie Chen, Andrew Chen, Pony Ma Published: 2026-10-15 Categories: cs.CV arXiv: https://arxiv.org/abs/2610.22276v1 Spatial intelligence lets an agent perceive and act in the physical world, and many of its judgments must be made at interaction rates: which spatial relation holds, how far an object is, or where to place it. Vision-language models (VLMs) produce these answers by free-form text generation, paying for autoregressive decoding and answer parsing even though the output space is compact: one of a few options, a scalar, or a 2D point. In light of this, we introduce MJev-Spatial, a System 1 multimodal decision model that tackles all three types of spatial understanding tasks within a unified classification framework. Specifically, MJev-Spatial treats each candidate as a class and scores it using the backbone's pretrained query and key projections, without adding parameters. The candidates are the given options for choice tasks, or value bins and image regions refined through coarse-to-fine classification for numerical and point prediction, respectively. At inference, each classification decision takes a single forward pass and generates no answer tokens. To train on heterogeneous spatial data efficiently, we supervise all coarse-to-fine stages jointly with teacher forcing in one forward pass, and combine this with length-aware global batching and cross-GPU load balancing, raising training throughput by 30.8% over the original batching strategy. Built on Qwen3.5-4B and trained on 12.3M question-answer pairs converted from existing spatial datasets, MJev-Spatial achieves the highest overall score (65.64%) among eight tested models, ranks first on seven of twelve benchmarks, and outperforms its unmodified autoregressive base model on all twelve, while running 2.5 times faster than the fastest autoregressive baseline and 5.6 times faster than the base model. --- ### Librarian Agents: Training LLMs to Curate Long-Context Workspaces that Anticipate Future Queries Authors: Anikait Singh, Teresa Zhang, Yoonho Lee, Roshen Sanjay Nair, Aviral Kumar, Chelsea Finn Published: 2026-10-15 Categories: cs.LG, cs.AI, cs.CL arXiv: https://arxiv.org/abs/2610.22273v1 Reasoning over long document contexts remains a challenge for large language models (LLMs). Standard retrieval methods are inherently reactive, bottlenecking agents with complex, cross-document reasoning at inference time, while precomputed structures like knowledge graphs often fall short because they rely on rigid heuristics. Instead, can we allow agents to perform anticipatory directed curation: learning to proactively reorganize unstructured corpora prior to inference to efficiently serve diverse future queries. We build on a key asymmetry in modern agentic harnesses (e.g., Claude Code): while LLM agents are adept at navigating existing filesystems to solve an immediate task, they are rarely trained to construct reusable workspaces for future use. To instantiate this capability, we introduce Librarian Agents, a framework in which a builder agent transforms raw documents into a hierarchical filesystem of synthetic files and reference links, which forms a workspace tailored for a downstream answering agent. We formulate workspace construction as a Markov decision process and optimize the builder policy via reinforcement learning to directly maximize downstream answering accuracy. Empirically, Librarian Agents outperforms the strongest retrieval and precomputation baselines by 33 percentage points on BrowseComp-Plus, amortizes upfront construction costs across repeated queries, and generalizes zero-shot to unseen domains on BRIGHT with a 31-point average accuracy improvement. --- ### Unsupervised Clustering of Public Transport Vehicles Using Deep Morphological Features and Color Histograms Authors: Juan Irving Vasquez Published: 2026-10-15 Categories: cs.CV arXiv: https://arxiv.org/abs/2610.22270v1 Surveillance cameras along urban roads produce large volumes of unlabeled video that could support the monitoring of public transport, but grouping the observed vehicles into meaningful categories without manual annotation remains challenging, since buses from different routes often share the same body structure and differ mainly in their color schemes. This study presents an unsupervised clustering methodology for public transport vehicles that combines morphological features extracted with a frozen DINOv2 Vision Transformer and handcrafted HSV color histograms through weighted early fusion, followed by UMAP projection and K-Means clustering. The method was evaluated on 7,972 bus images extracted from about 1,460 hours of video of an arterial road in Mexico City, with the goal of isolating the buses of the city's public bus network (RTP) without supervision. With a fusion of 65\% morphological and 35\% color features, the pipeline produced 14 clusters that separate distinct routes and vehicle types, and it increased the precision of the target cluster with respect to morphological features alone. These results show that color complements self-supervised features for route-level clustering, provided that it does not dominate the representation. --- ### X-RIC: An Agentic RAN-Intelligent Controller Benchmark for Enhancing Reinforcement Learning Authors: Adhwaa Al-Chaab, Ayman Younis, Dario Pompili Published: 2026-10-15 Categories: cs.AI, cs.AR arXiv: https://arxiv.org/abs/2610.22269v1 The O-RAN ALLIANCE introduces architectural flexibility through disaggregation and AI-driven control. While the RAN Intelligent Controller (RIC) enables centralized optimization, current AI approaches lack reasoning capabilities and explainability. We propose eXplainable RIC (X-RIC), a novel framework integrating agentic AI into O-RAN architectures. X-RIC combines reactive Reinforcement Learning (RL) with strategic agentic reasoning through: (i) automated Behavioral Cloning (BC) using domain heuristics for sample-efficient training, (ii) a Telecom Language Model (TLM), a transformer-based reasoning module providing contextual abstraction and interpretability, and (iii) closed-loop co-adaptation within the Near-RT RIC control loop enabling continuous policy refinement. To ensure deployment readiness, we introduce BARX (Benchmarking Agentic Reasoning for RAN), a domain-specific evaluation framework assessing agentic control across latency, compliance, and determinism metrics. We validate X-RIC on SLA-aware network slicing in simulation and O-RAN-compliant testbed deployments. X-RIC achieves 89% operational cost reduction while maintaining SLA compliance, converging 75% faster with 50% fewer training steps and 46% lower variance. --- ### Beyond Pairwise Repulsion: Collective Diversity via Determinant Maximization in Diffusion Sampling Authors: Kwanyoung Lee, SeungJu Cha, Yebin Ahn, Sungho Koh, Dong-Jin Kim Published: 2026-10-15 Categories: cs.CV arXiv: https://arxiv.org/abs/2610.22267v1 Strong guidance in text-to-image (T2I) diffusion models concentrates sampling toward high-density modes of the conditional distribution, collapsing independently sampled noise vectors onto nearly identical visual prototypes. This limits utility for tasks that demand distributional breadth, most critically synthetic data augmentation. Diversity, however, is a collective property of the whole batch: a set can preserve every pairwise distance yet still collapse onto a low-dimensional subspace. Existing sum-of-pairwise noise-optimization and repulsion objectives respond only weakly to this higher-order failure. We propose DetNO (Determinant Noise Optimization), which formulates diverse generation as RKHS volume maximization on the initial noise, maximizing the log-determinant of an RBF Gram matrix over one-step Tweedie predictions. This set-level functional equals the squared batch volume; its repulsion is strongest when the batch is most collapsed and self-attenuates as diversity is achieved. An anchor-based fidelity term keeps the optimized noises near plausible regions, and Lagrangian relaxation gives a two-term objective with a single hyperparameter $λ$ governing the diversity-quality trade-off, without training or sampler modification. Across three backbones (SD1.5, SDXL, SD3), DetNO achieves leading intra-batch diversity in VendiScore and Mean Self-Similarity, and a downstream few-shot classification experiment confirms that broader mode coverage brings statistically significant accuracy gains for synthetic data augmentation in the low-shot regime. Fixing the pipeline and varying only the batch objective, both RBF-based aggregators outperform InfoNCE, and the log-determinant attains the best downstream accuracy on both benchmarks. Its diversity advantage over InfoNCE widens as the batch grows. Code is available at https://github.com/mobled37/DetNO. --- ### SCOPE: Semantic Bounded Credit Optimization for Prompt Engineering in Text-to-image Generation Authors: Yufei Jiao, Jinfa Huang, Zeyue Xue, Tianyu Xie, Yuexiao Ma, Rongrong Ji, Xiawu Zheng, Jiebo Luo Published: 2026-10-15 Categories: cs.CV arXiv: https://arxiv.org/abs/2610.22264v1 Reinforcement learning (RL) enables language models to improve text-to-image (T2I) generation by rewriting prompts for a frozen image generator. Existing methods typically apply a single signal derived from the image to individual subword tokens, although objects, counts, attributes, and relations form coherent semantic spans. We propose SCOPE, an objective based on Group Relative Policy Optimization (GRPO) that aligns update weighting with this structure. The rewriter jointly generates an enhanced prompt and its semantic partition. SCOPE pools shifts in policy log probabilities within each span to derive a shared coefficient for its token losses while retaining a single advantage for each completion. It requires neither separate span rewards nor additional image renders. Across Qwen-Image, SD3, and FLUX.1-dev, SCOPE achieves average relative gains over vanilla GRPO of 9.48% on GeneVal and 7.62% on T2I-CompBench. Ablation studies show that both semantic boundaries and pooling within spans contribute to these gains, supporting semantic structure as a practical inductive bias for prompt engineering with RL. --- ### Baselines and Score Centering in Policy Gradient Estimation Authors: Shuo Han, Ziqing Hu Published: 2026-10-15 Categories: cs.LG arXiv: https://arxiv.org/abs/2610.22263v1 Reinforcement learning has become a widely used approach to post-training large language models. In a typical implementation, learning is off-policy: The policy generating the responses differs from the policy being trained because of training-inference mismatch or policy staleness in asynchronous training. This discrepancy is commonly addressed by importance sampling, whose ratio is usually regularized in practice through truncation, masking, or clipping. Under regularized importance sampling, the importance-weighted policy gradient is no longer invariant to constant shifts in the reward. This invariance is recovered by using the covariance between the reward and the importance-weighted score as the estimand, which coincides with the policy gradient when the ratio is not regularized. This note develops this covariance perspective for both sequence-level and token-level importance sampling. The covariance perspective reveals the roles of the group-mean baseline and score centering in policy gradient estimation. Both perform centering in covariance estimation, but they differ in the mean being subtracted: The group-mean baseline uses a sample mean, whereas score centering uses a population mean. The two lead to distinct unbiased covariance estimators with different variances. Whether both forms of centering should be used simultaneously depends on whether importance sampling is applied at the sequence level or at the token level. With sequence-level importance sampling, score centering adds no benefit once the group-mean baseline is used; with token-level importance sampling, using the group-mean baseline without score centering generally introduces bias when the importance ratio is regularized. --- ### Topology-Inspired Fine-Tuning for Wind-Field Super-Resolution Authors: Arjun Dadhwal, Brian Summa Published: 2026-10-15 Categories: cs.LG arXiv: https://arxiv.org/abs/2610.22262v1 Scientific super-resolution (SR) models are typically optimized with pointwise reconstruction losses, yet scientific interpretation often hinges on extrema, connected regions, and multiscale scalar structure. We characterize this tension with pretrained CNN and GAN wind-field SR models. Pointwise metrics, bottleneck persistence-diagram (PD) distance, and merge-tree (MT) distance favor the CNN, whereas order-2 Wasserstein PD distances favor the sharper GAN, so even different distances on the same topological descriptor reward different properties. Motivated by this disagreement, we fine-tune the pretrained CNN with differentiable topology-inspired auxiliary losses on scalar speed, gradient magnitude, soft high-speed level sets, and local maxima, and with fixed-pair supervision, which constrains predicted values and persistence at ground-truth persistence-pair endpoints located offline with the Topology ToolKit. A loss study of 16 fine-tuning configurations, each compared with a matched standard fine-tuning control that uses only the original reconstruction loss, on a fixed 168-field benchmark (with topology evaluated on a fixed 160 by 160 subregion of each field), shows that, among the scalar-field losses, gradient supervision provides the strongest PD improvement but does not improve mean MT distance, whereas fixed-pair supervision gives the largest and most consistent joint improvement in PD and MT distances. The final objective combines these two signals. Relative to the pretrained CNN, it lowers mean bottleneck, $W_{2,\infty}$, and $W_{2,2}$ distances by 31%, 25%, and 27% (and by 32-35% relative to the control, which on its own worsens all three), lowers all three on 150 of 168 fields, and gains 1.30 dB PSNR, about half the gain of standard fine-tuning. Mean MT distance also improves, but on only 91 of those 150 fields, indicating that PD and MT capture complementary structure. --- ### FPG-Seg: Annotation-free Retinal Vessel Topology Segmentation Authors: James Willoughby, Irina Voiculescu Published: 2026-10-15 Categories: cs.CV arXiv: https://arxiv.org/abs/2610.22259v1 State-of-the art approaches to retinal vessel segmentation typically require extensive annotations, a highly time-consuming process, reliant on clinical expertise. We present FPG-Seg, an annotation-free pipeline for retinal vessel segmentation combining classical vesselness filtering, Bayesian deep learning and topology-aware graph reconstruction to segment retinal vessel topology without manual labels. Tested across four widely used benchmark datasets for retinal vessel segmentation, FPG-Seg attains a clDice of $0.74{-}0.78$, representing $86{-}98\%$ of the performance of an equivalent fully-supervised baseline, despite using no manual annotations for training or hyperparameter tuning. As part of this work, we also provide, to our knowledge, the first performance benchmark of an annotation-free vessel segmentation method on FIVES. --- ### Denoising Textures is a Different Ballgame Altogether! Authors: Saptarshi Mandal, Oindrila Haldar, Chirayata Bhattacharya, Chandra Sekhar Seelamantula, Zhiwei Xu, Thrasyvoulos N. Pappas, Gaurav Sharma Published: 2026-10-15 Categories: eess.IV, cs.CV arXiv: https://arxiv.org/abs/2610.22258v1 The human brain effortlessly deciphers colors and textures from visual stimuli and reconstructs the world around us, yet preserving such details remains challenging in image denoising. While significant progress has been made in generic image denoising, the problem of texture denoising has remained relatively under-explored. Fine details, such as intricate patterns and repetitive structures, often suffer from over-smoothing even in deep learning based approaches for denoising. In this paper, we analyze the behavior of state-of-the-art (SOTA) denoisers on texture datasets and show that performance improvements are possible only by fine-tuning the denoisers on the texture datasets. Surprisingly, fine-tuning on texture datasets result in the models catastrophically forgetting the original dataset. These findings highlight the need for specialized treatment of textures in denoising. Since denoising quality directly impacts downstream tasks such as generative modelling, we argue that dedicated texture aware denoisers are essential for faithful texture reconstruction and synthesis. --- ### BitSparse: Compact Activation Storage for ReLU Feed-Forward Networks Authors: Max Zhu, Bohang Sun, Pietro Lio, Jialin Yu Published: 2026-10-15 Categories: cs.LG arXiv: https://arxiv.org/abs/2610.22256v1 Reducing memory use in large language model (LLM) training is an ongoing challenge. Activation checkpointing saves memory by recomputation, while lossy compression changes the stored activations. The reemergence of sparse $\relu$ and $\relu^2$ activations creates a third opportunity: losslessly compressing the activation cache itself. We present BitSparse, a GPU-resident lossless compression format for sparse activations. Compression and reconstruction run in custom Triton kernels fused into an autograd function, shrinking the activation cache while recovering it exactly for the backward pass. Bfloat16 (BF16) and FP8 activations are supported. Optional sign-bit packing further increases compression up to close to the theoretical limit. We analyze its compression ratio and compare it to the theoretical optimum. BitSparse lowers per-token VRAM by 31\% and 30\% Modded-NanoGPT and Nemotron-H respectively, with overheads of no more than 1\% at practical sequence lengths. Additionally, a new preallocated buffer system avoids host synchronization during training when variable size tensors are allocated. --- ### Explore Once, Evaluate Many: Multiple Policy Evaluation for Stochastic Linear Bandits Authors: Spyros Dragazis, Yilei Chen, Aldo Pacchiano Published: 2026-10-15 Categories: cs.LG, stat.ML arXiv: https://arxiv.org/abs/2610.22252v1 We study how to evaluate $K$ target policies simultaneously in stochastic contextual bandits with linear reward structure. Given a known $d$-dimensional feature map and reward-free access to policy queries, we propose PRISMA (Policy evaluation through Regularized Information Sharing and Mixture Allocation), an algorithm that estimates every target policy's value to accuracy $\varepsilon$ with probability at least $1-δ$. The algorithm uses $\widetilde{\mathcal{O}}\!\left( \frac{(L^2S^2+β^2)(1\veeρ^\star)}{\varepsilon^2} \cdot \log(K) \right)$ context draws and reward observations, where $ρ^\star\le d$ is an instance-dependent measure of the difficulty of designing a shared sampling distribution. Here, $L$, $S$, and $β$ bound the feature norm, reward parameter norm, and reward noise, respectively, and the suppressed factors are logarithmic in $K$, $d$, and $1/δ$. Our method combines reward-free planning, a regularized sampling design, and a single ridge regression estimator shared across all target policies. An adaptive doubling procedure selects the regularization scale without knowledge of $ρ^\star$. The guarantees require no numerical lower bound on second-moment eigenvalues and have no explicit dependence on the cardinalities of the context and action spaces. Policy-query costs are accounted for separately, with $K$ queries per shared planning context. --- ### Not X, but When: Diagnosing a Mandarin LLM Writing Tic through Contextual Appropriateness Authors: Yihang Peng, Yi Jing, Dewen Liu, Weihan Meng, Ying Liu Published: 2026-10-15 Categories: cs.CL arXiv: https://arxiv.org/abs/2610.22249v1 Recurring expressions in model-written text are often treated as signatures of machine authorship, yet most of them serve legitimate functions. Whether such an expression is a tic depends on whether the model knows when it is called for. This judgment of appropriateness, central to communicative competence, remains largely unexamined for language models. We study it through the Mandarin corrective construction bushi X, ershi Y ("not X, but Y"). Drawing on Gricean pragmatics and corrective negation, we introduce a decision tree that classifies uses as substantive, optional, or defective by testing whether the context makes X a live alternative and whether rejecting X is informative. Five commercial models use the construction 2.5 to 9.2 times as densely as a human reference. Of 307 annotated uses, only 1.30% are defective while 52.8% are optional, that is, correct but unneeded, so models fall short in appropriateness far more often than in correctness. Several reward models favor the construction even over quality-matched variants, an LLM judge favors it even where the context does not license it, and both supervised fine-tuning and preference optimization tend to raise its frequency in open-weight models. We hope this work encourages the development of models attentive to the appropriateness of their expressions. --- ### Disagreement-Guided Weak-to-Strong Routing Authors: Sima Noorani, Shayan Kiyani, George Pappas, Hamed Hassani Published: 2026-10-15 Categories: cs.AI arXiv: https://arxiv.org/abs/2610.22245v1 Small language models are increasingly deployed on edge devices and robotic systems to meet cost and latency constraints, yet their limited capabilities can make reliable deployment challenging. Stronger models can provide assistance when needed, but deciding when to route a weak-model response to a stronger model often relies on signals of response correctness. In open-ended and agentic tasks, however, obtaining such signals can require costly ground-truth supervision that is difficult to obtain at scale. We revisit the foundations of weak-to-strong routing and ask whether routing decisions can instead be guided by signals that do not require ground-truth supervision. We first characterize the optimal weak-to-strong routing policy, while allowing for multiple weak-model responses generated through local inference. Building on this characterization, we derive a principled approximation to the optimal routing based on model-to-model comparison. In particular, we introduce a notion of model disagreement that can be measured directly from model-generated responses, without ground-truth supervision. Importantly, the approximation becomes increasingly accurate as the strong model becomes a better proxy for the ground-truth distribution. Across agentic and single-turn tasks, our approach effectively navigates the trade-off between routing cost and response accuracy without ground-truth supervision. Moreover, even when ground-truth correctness signals are available for training, disagreement-based routing can outperform methods trained directly to predict correctness --- ### Language Is a Security Surface for Agent Guards Authors: Hasan Demirkiran, Jens Ernstberger Published: 2026-10-15 Categories: cs.CR, cs.AI, cs.CL, cs.LG arXiv: https://arxiv.org/abs/2610.22244v1 Step-level guards decide, before each tool call, whether an agent's next action is safe given the user's request and the trajectory so far. Users write those requests in many languages, while guard prompts and the benchmarks on which purpose-built guards are trained and evaluated are English. We ask whether a guard's verdict depends on the language of the request. We build LinguaGuardBench from the 1,220 labelled steps of TS-Bench AgentDojo-Traj: we translate only the user instruction into German, French, Turkish, Swahili and Amharic, and keep every tool call, observation and label byte-identical, so any change in a verdict is attributable to language alone. Six English rewordings and a whitespace placebo serve as controls, and managed guards are run three times to measure their noise. Across 19 guards, from frontier APIs to 4B open models and four purpose-built agent guards, most open, small and purpose-built guards change verdicts under translation beyond their own noise floor, and the effect grows with the rarity of the language: among the nine guards below 13B parameters, the mean placebo-adjusted flip rate rises from +1.5 pp on French to +4.4 pp on Amharic. Purpose-built guards are the most sensitive. The direction of the damage tends to follow guard size: open guards of 12-30B parameters block more benign steps (up to +5.0 pp false positives), and every significant loss of recall is in a guard below 10B (up to -7.4 pp). Frontier guards shift by at most about one point; the strongest, Gemini 3.8 Flash, returns identical verdicts with and without the request. The damage is concentrated in a few instructions and, in a post-hoc analysis, falls on steps where the agent's reasoning restates the English request, consistent with guards reacting to inconsistency between channels. English paraphrase moves every guard less than translation does. We release the benchmark and all 489k verdicts. --- ### Shortcut Sensitivity in Agent-Safety Evaluations Authors: Hasan Demirkiran, Jens Ernstberger Published: 2026-10-15 Categories: cs.AI, cs.CR, cs.LG arXiv: https://arxiv.org/abs/2610.22243v1 AI agents take actions through tools. Some actions are sensitive, so guards that supervise them must themselves be evaluated reliably. Yet the data used to evaluate a guard can include cues that make labels predictable without measuring the intended safety construct, and those cues can also influence guard decisions. We introduce a three-level shortcut-auditing framework and audit completed LinuxArena trajectories and two TS-Bench-derived step datasets. We distinguish label predictability, observational guard association, and paired decision sensitivity. In LinuxArena, shorter average shell commands strongly predict trajectory labels. In AgentDojo-Traj, exact-tool lookup falls from F_1=.901 with mixed-domain folds to .495 under domain holdout. In AgentHarm-Traj, released TS-Guard F_1 falls from .902 under row weighting to .719 with equal interaction weighting, while balanced accuracy rises. We then test the strongest safe-row marker association in a seven-configuration screen on a frozen 64-step AgentDojo-Traj cohort using repeated, interleaved Haiku calls. Rewriting the <INFORMATION> marker lowers blocking by 8.33 percentage points, while opaque tool aliases raise it by 11.46 points. All unsafe calls remain blocked; the changes concern false positives against inherited labels. Both effects remain nonzero after subtracting the formatting-placebo effect, with exploratory 95% ranges excluding zero. The same edits also change measured accuracy against inherited labels. This provides direct evidence that benchmark representation can materially change a guard's decisions and measured evaluation performance. The Haiku-sized effects do not reproduce in Luna or deterministic Q8 TS-Guard. We recommend domain group holdouts, evaluation-unit metrics, and placebo-controlled paired interventions. --- ### SurvICL: a Foundation Model for in-context Survival Analysis Authors: Julie Alberge, Linus Bleistein, Judith Abécassis, Marine Le Morvan, Gaël Varoquaux, David Holzmüller Published: 2026-10-15 Categories: stat.ML, cs.LG arXiv: https://arxiv.org/abs/2610.22242v1 Tabular foundation models (TFMs) now outperform gradient-boosted trees in classification and regression. Yet in survival analysis, which predicts when an event occurs from censored data, often on small clinical cohorts, we find that no existing TFM beats the half-century-old Cox model. We address this issue by introducing SurvICL, a TFM that predicts survival curves in-context. SurvICL is pretrained on purely synthetic datasets generated from a novel flexible proportional-odds prior that outperforms other prior types in our ablation study. Unlike most other survival models, SurvICL is trained with a plain quantile regression loss, which is possible because its synthetic pre-training data generator can provide uncensored event times. Additionally, we introduce the largest survival benchmark to date spanning 103 datasets. On the 63 datasets held out from all design decisions, SurvICL outperforms 20 tree-based, deep, and foundation-model baselines without any tuning: it wins 82.9\% of pairwise comparisons in integrated Brier score, against 71.3\% for the runner-up, the Cox model, and is the first model to rank above Cox on small datasets. --- ### MMSearch-OPD: Easy Made Hard, Hard Made Learnable for Long-Horizon Multimodal Search Agents Authors: Huanyao Zhang, Jiepeng Zhou, Jiaye Lin, Ontai Tang, Jingtian Hao, Mugeng Liu, Ru Peng, Lang Mei, Xiaohan Yu, Liyan Liu, Chong Chen, Wentao Zhang Published: 2026-10-15 Categories: cs.CV arXiv: https://arxiv.org/abs/2610.22240v1 Multimodal deep search is moving from short queries to long-horizon investigations that continually acquire evidence. However, training such agents remains constrained by limited challenging training questions and insufficient process feedback. We introduce MMSearch-OPD, combining controllable data refinement with joint reinforcement learning (RL) and on-policy distillation (OPD). Our Refine pipeline performs targeted expansion of reasoning chains derived from existing simple multimodal questions. Parameterized transformations control search depth and cross-modal evidence requirements while reusing existing image--question pairs. For training, answer rewards and lightweight behavior penalties guide RL, while the reward gap modulates token-level teacher guidance on imperfect student trajectories. An exponentially decaying distillation coefficient reduces teacher dependence as training proceeds. Experiments show gains from data refinement and further improvements from joint training after SFT, indicating stronger long-horizon multimodal deep search capabilities. --- ### Query-Efficient Rejection Sampling for Parallel Masked Diffusion Authors: Shiyuan Zhang, Xuheng Li, Qiwei Di, Quanquan Gu Published: 2026-10-15 Categories: cs.LG, math.PR, stat.ML arXiv: https://arxiv.org/abs/2610.22238v1 Masked diffusion models (MDMs) accelerate generation by unmasking multiple tokens in parallel, but sampling these tokens independently ignores their correlations and biases the output. Anari et al. (2026) corrected this bias with rejection sampling, decomposing the log acceptance probability into a sum of token-level terms, but the evaluation requires one score query per unmasked token. We propose the Single-Increment Rejection Sampler (SIRS), which provides an estimate of the log acceptance probability from a single randomly sampled increment at the cost of one query. SIRS then converts a Poisson number of clipped estimates into accept/reject decisions through a Bernoulli factory adapted from Chen et al. (2026). Since each query returns the conditionals of all remaining tokens in the block, we further Rao-Blackwellize each increment by averaging over these tokens at no extra cost, thereby reducing clipping. We show that the KL-divergence between the data distribution and the SIRS output is bounded by a clipping term plus the model's training error. With exact scores, parallel decoding pays the full within-block dependence, whereas SIRS pays only the part beyond the clipping range. --- ### Training-Free On-Policy Distillation Authors: Bangji Yang, Jiajun Fan, Zhaoyu Zhu, Hongbo Ma, Ge Liu Published: 2026-10-15 Categories: cs.LG, cs.AI, cs.CL arXiv: https://arxiv.org/abs/2610.22236v1 On-policy distillation consolidates specialist models into one student at the price of rollouts, teacher passes and gradient steps for every teacher. We show that the consolidation does not need training: with forward passes only, the capabilities of separately tuned specialists can be aggregated into one model that is, on average, more accurate than either specialist alone. When specialists share a base model, distillation has an exact layer-wise solution: each linear layer of the fused model must reproduce every specialist's output on the inputs that specialist receives while reading its own answers, a least-squares problem whose solution needs only the second moments of those inputs. One forward pass over a few hundred rollouts provides them. We call this training-free on-policy distillation (\method). What the rollouts contain matters more than the solver: answers truncated in repetition loops are few among a mathematics specialist's rollouts but carry a large share of their tokens, and removing them lifts the fused model on AIME25 from task arithmetic's level to the specialist's at matched precision. On Olmo~3 7B, a few GPU-hours of forward computation per specialist yield a generalist that keeps most of each specialist's gain on each of seven mathematics and code benchmarks. --- ### OS-Helix: Unifying GUI Interaction and Code Execution in Native Computer-Using Agents Authors: Bowen Yang, Kaiming Jin, Zichen Ding, Zhenyu Wu, Zehao Li, Zhoumianze Liu, Kanzhi Cheng, Fangzhi Xu, Yanan Sun, Qiushi Sun, Yu Qiao Published: 2026-10-15 Categories: cs.LG, cs.AI, cs.CV arXiv: https://arxiv.org/abs/2610.22234v1 Real desktop automation is neither purely visual nor purely programmatic. Agents must understand user-facing interfaces while efficiently manipulating files, scripts, and system state. However, native GUI/code coordination cannot be obtained by prompt-level tool composition. In our preliminary experiments, naively adding a code tool degrades a GUI agent, revealing that GUI/code routing must be internalized through training on high-quality hybrid data. We present OS-Helix, an end-to-end trained multimodal computer-using agent that unifies GUI actions and code execution in one action space. To construct the required supervision, we introduce application-artifact composition, pairing desktop applications with authentic file artifacts to instantiate executable Ubuntu states, and a decoupled instruction/evaluator generation pipeline that produces environment-grounded tasks whose final states can be verified regardless of the solution modality. We cold-start the policy with SFT on filtered hybrid data and further improve it through online RL in live desktop environments. For RL, we propose Step-balanced GSPO, which aggregates importance weight by action step rather than raw assistant tokens to prevent long code actions from dominating heterogeneous-action optimization. On OSWorld-Verified, OS-Helix-27B and OS-Helix-35B-A3B achieve relative improvements of 35.0% and 33.7% over their matched Qwen3.5 hybrid baselines, while OS-Helix-27B-RL further improves by 42.9% to reach 66.44% success. Ablations provide evidence consistent with contributions from learned GUI/code coordination, scalable verifiable supervision, and action step balanced optimization, while demonstrating cross-platform transfer on WindowsAgentArena. Our project homepage is at https://os-copilot.github.io/OS-Helix/. --- ### WaveLLM: Wavelet-Domain LLM Entropy Modeling for Lossless Image Compression Authors: Yue Yu, Bowei Kang, Changsheng Gao, Dong Liu, Xiaoyan Sun Published: 2026-10-15 Categories: cs.CV arXiv: https://arxiv.org/abs/2610.22232v1 Lossless image compression fundamentally relies on accurate entropy modeling, driving a recent shift toward large language model (LLM)-based methods. However, existing approaches mainly emphasize the modeling capacity of LLMs while directly serializing images into pixel or residual sequences, leaving heterogeneous signal patterns entangled within a single prediction stream. In this paper, we argue that effective entropy modeling depends not only on model capacity, but also on the predictability of the input representation. Motivated by this insight, we propose WaveLLM, a wavelet-domain LLM-based entropy model for lossless image compression. By decomposing an image into multi-level wavelet subbands, WaveLLM converts a heterogeneous prediction problem into a set of more structured and statistically homogeneous subproblems that are more amenable to autoregressive probability estimation. To bridge numerical coding symbols with the LLM's native text token space, we introduce a dedicated numerical interface featuring value embeddings and a coefficient prediction head. Furthermore, a task-adaptive prompting strategy is designed to enable a single shared model to adapt to diverse subbands within a unified framework. Built on Llama-3.2-3B with LoRA adaptation, WaveLLM achieves 2.51, 2.24, and 2.27 bpsp on Kodak, CLIC.pro, and DIV2K, respectively, outperforming the PixelCNN-based codec using the same wavelet transform by 0.43, 0.36, and 0.46 bpsp, and surpassing recent LLM-based lossless compressors on all three benchmarks. --- ### Architecting for Hardware Diversity: Meta's End-to-End Co-Design for High-Performance AI Training on AMD GPUs Authors: Haoyu Zhang, Arshad Rahman, Adnan Aziz, Ajith Prabhudev, Amin Farmahini, Ben Rhoads, Benson Ma, Brian Shi, Cen Zhao, CK Luk, Dingming Wu, Emad El-Haraty et al. Published: 2026-10-15 Categories: cs.DC, cs.AR, cs.LG arXiv: https://arxiv.org/abs/2610.22230v1 Meta deploys a heterogeneous fleet comprising NVIDIA GPUs, AMD GPUs, and proprietary MTIA accelerators to reduce costs and mitigate vendor lock-in. However, the extensive history of NVIDIA-specific optimizations hinders performance portability to other hardware. This paper presents our methods for optimizing alternative accelerators, using AMD Instinct MI350X versus NVIDIA B200 for recommendation model training as a case study. We propose three reusable principles. First, hardware-model co-design is essential: rather than directly porting models from NVIDIA GPUs, revising model structures and software recipes allows full exploitation of accelerator-specific strengths (e.g., MI350X's 1.6x HBM capacity relative to B200). Second, our analytical performance model enables zero-warmup optimal kernel selection across diverse datatypes and matrix shapes in production models, drastically simplifying tuning. Third, increasingly mature agentic workflows have lowered the barrier to porting from the CUDA ecosystem; we extend these workflows for microarchitecture-aware kernel optimizations and custom shim generation. Compared to the directly ported baseline, these methods deliver up to 1.5x improvement in end-to-end training throughput. To our knowledge, this is the first hyperscaler report on productionizing AMD GPUs for large-scale training. Our contributions have been upstreamed to the open-source PyTorch, Triton, and ROCm ecosystems. --- ### How Much Physics Can Vision Learn from Audio? Probing and Distilling Damping Structure Authors: Zainab Afolabi, Vadim Selyutin, Zhao Gao, Babs Khalidson, Mohamed Ghoneim Published: 2026-10-15 Categories: cs.CV, cs.LG, cs.SD arXiv: https://arxiv.org/abs/2610.22221v1 Physical understanding is crucial for decision making, yet vision models struggle to infer properties such as damping that govern how materials behave. Audio encodes such properties directly, because an object's acoustic response to impact reveals its damping through measurable decay. We probe frozen encoders for the quality factor $Q$, a dimensionless measure of how many oscillations an object completes before its ring dies away, and report how well each embedding orders impacts by $Q$ using the Spearman rank correlation $ρ$, for which $ρ=1$ is a perfect ordering and $ρ=0$ an unrelated one. Audio embeddings reach $ρ=0.73$ and retain most of that advantage on materials withheld from training, while a single video frame taken at the moment of impact, encoded by CLIP, supports only $ρ=0.28$. Having established where the signal lives, we ask how much of that gap vision can recover. We distill a frozen audio teacher into a frozen CLIP backbone through a trained projection head and find that this recovers a useful fraction of the audio advantage. Because pooled correlations on this task are dominated in practice by category identity, we also evaluate under a leave-one-material-out protocol in which the held-out material is withheld from the projection head as well as the probe. Distillation with a four-frame input improves the correlation there from 0.234 to 0.338 (+0.053 over the matched four-frame baseline, 14 of 17 materials), a gain that no category-level strategy can produce. Widening the visual input to four frames spanning the impact supplies half of that improvement at no acoustic cost, which indicates that part of what a single frame misses is visible dynamics. --- ### Skill-Decoupled On-Policy Expert Distillation for Unified Vision-Language Models Authors: Yuqi Xiong, Chunyi Peng, Zhipeng Xu, Qing Yang, Zhenghao Liu, Zulong Chen, Yu Gu, Ge Yu Published: 2026-10-15 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2610.22217v1 Reinforcement learning (RL) can elicit specialized visual capabilities in vision-language models (VLMs), yet jointly optimizing heterogeneous objectives often introduces cross-capability gradient conflicts. Training experts independently mitigates this interference, but direct reward optimization may entangle specialized knowledge with reward-hacking biases. This paper proposes Skill-Decoupled On-Policy Expert Distillation (SCOPED), which trains capability-specific experts and extracts explicit skills through reward-guided reflection. During on-policy distillation, these skills guide expert supervision while remaining unavailable to the unified student, encouraging it to internalize procedural knowledge without explicit skills at inference time. Across eight multimodal benchmarks, SCOPED consistently outperforms the backbone and mixed-capability RL, demonstrating the effectiveness of skill-decoupled on-policy expert distillation for integrating independently acquired capabilities. Further analyses reveal interference during joint capability optimization and suggest that skill augmentation promotes more consistent expert trajectories, greater attention to task-relevant visual evidence, and better preservation of specialized expertise in the unified student. All code and data are available at https://github.com/NEUIR/SCOPED. --- ### FutureLift: Lifting Trajectory Proposals into Bidirectional Future Graphs for Motion Forecasting Authors: Daniel Grimm, J. Marius Zöllner Published: 2026-10-15 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2610.22215v1 Accurate motion prediction is critical for safe autonomous driving, yet even strong forecasters leave residual errors in their multimodal proposals. Existing refiners compress complete proposals or update their segments sequentially, preventing the full proposed future from serving as explicit relational context for correction. We introduce FutureLift, a backbone-agnostic refiner that represents each multimodal proposal as a heterogeneous future graph, with spatially grounded temporal nodes connected to agent histories, map elements, alternative modes, and, for joint prediction, other agents future trajectories. Bidirectional temporal edges allow later evidence to revise earlier segments in a single pass, and the updated nodes predict residual coordinate and confidence corrections. Requiring only proposals, scores, and scene context, FutureLift provides a common refinement interface for structurally different forecasting backbones. On Argoverse2 and the Waymo Open Motion Dataset, it consistently improves frozen proposals and outperforms native and specialized refiners across marginal and joint forecasting. On the WOMD motion-prediction leaderboard, FutureLift ranks first in minADE (0.5049m) and minFDE (0.9942m) among the compared standard-input methods, without LiDAR or ensembling, while achieving a competitive 0.4874 Soft mAP. --- ### GNN-Guided Local Refinement for Efficient Layout Decomposition with DSA-MP Authors: Yu Feng, Liyi Zhou, Haoyu Wang, Yiming Xu, Hao Geng, Shaozhi Deng, Ningsheng Xu, Shisheng Xiong Published: 2026-10-15 Categories: cs.AI arXiv: https://arxiv.org/abs/2610.22214v1 Multiple patterning lithography (MPL) has been widely used to improve lithography resolution in advanced nodes, but it faces cost limits as feature sizes shrink. Directed self-assembly (DSA) offers a cost-effective complement, extending the resolution of optical lithography to sub-10 nm and beyond. Integrating DSA into MPL (DSA-MP) further enhances the resolution of contact/via patterning. Nevertheless, growing design complexity exposes the limits of existing decomposition methods, which struggle to balance solution quality and runtime, and neglect edge-to-edge spacing constraints of DSA templates, compromising manufacturability. To address these issues, we introduce a new conflict graph model capturing the previously overlooked constraints and develop a local refinement strategy powered by the graph neural network, enabling efficient and scalable layout decomposition. Experiments demonstrate that our method achieves higher solution quality with faster convergence than state-of-the-art methods, while maintaining robustness across layout scales, design styles, technology nodes, and varying mask counts. --- ### Generalization Guarantees of Randomized Block Coordinate Descent for Nonconvex Learning Authors: Ruixi Hu, Liang Wu Published: 2026-10-15 Categories: cs.LG arXiv: https://arxiv.org/abs/2610.22213v1 Randomized block coordinate descent (RBCD) is a popular optimization algorithm for many large-scale machine learning problems. The existing theoretical analysis of RBCD mainly focuses on its convergence behavior from an optimization perspective, while its generalization behavior remains less understood for nonconvex problems. In this paper, we establish high-probability generalization guarantees for RBCD with nonconvex objectives. In the general nonconvex setting, we derive bounds on the gradients of both empirical and population risks, requiring only standard smoothness assumptions. With polynomially decaying step sizes, we tune the iteration number to balance optimization and generalization, thereby quantifying how early stopping leads to the optimal convergence rate for the population gradients. In the gradient-dominated setting, we provide bounds in terms of functional suboptimality, which nearly match the optimistic rates in convex and low-noise settings. Finally, we develop dimension-independent bounds for loss functions of a specific structure, making the generalization guarantees applicable to high-dimensional problems. --- ### A restart rule for genetic algorithms based on Schnabel census Authors: Anton Eremeev, Yulia Zakharova Published: 2026-10-15 Categories: cs.NE arXiv: https://arxiv.org/abs/2610.22211v1 A new adaptive restart rule for Genetic Algorithms (GAs) is considered. The rule is based on the Schnabel Census method, originally developed for statistical estimation of a size of animal population. In this paper, the Schnabel Census method is applied as a heuristic to estimate the number of different solutions that may be visited with positive probability, given the current distribution of offspring. The rule consists in restarting a GA as soon as the maximum likelihood estimate reaches the number of different solutions observed at the recent iterations. We demonstrate how the new restart rule can be applied in a GA with steady-state replacement scheme, using three different combinatorial optimization problems as examples. Computational experiments on well-known benchmarks show a statistically significant advantage of the GAs with the new restarting rule over the original versions of GAs. The new restart rule also tends to be superior to the well-known rule, which restarts an algorithm when the current iteration number is twice the number of iterations till the current best incumbent was found. --- ### TRACE: Topology-Aware Reconnection via Prediction-Adaptive Connectivity Enforcement for Thin-Structure Segmentation Authors: Junho Moon, Ikbeom Jang Published: 2026-10-15 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2610.22210v1 Thin, elongated structures are particularly sensitive to connectivity errors, as even a few local segmentation mistakes can disrupt otherwise continuous paths. Existing skeleton-based objectives typically supervise fixed centerline locations and do not account for alternative local routes that may restore connectivity. We propose TRACE, a topology-aware, prediction-adaptive objective that uses a target skeleton to define required local connections. For each connection, TRACE identifies the easiest valid reconnection under the current prediction and jointly supervises its bottleneck with the corresponding skeleton point. The supervision thus adapts to the evolving prediction rather than remaining fixed to the centerline. We further introduce a topology-aware skeleton carrier that preserves both foreground and class-specific topology, enabling reliable definition of these connections. Extensive experiments across six 2D and 3D benchmarks show that TRACE improves connectivity and topology while maintaining strong segmentation accuracy with low training overhead. Code will be released upon acceptance. --- ### Beyond Downstream QA: Rethinking How We Evaluate Agent Memory Authors: Sebastian-Antonio Toma, Jean de Kernier, Mike Ply, Oliviu Matei Published: 2026-10-15 Categories: cs.CL, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2610.22209v1 Agent memory benchmarks evaluate whether an agent's responses reflect correct knowledge, measuring downstream question-answering accuracy as a proxy for memory quality. We show this evaluation paradigm is incomplete: it conflates memory correctness with retrieval and generation quality, producing both false positives where correct responses mask broken memory and false negatives where incorrect responses blame correct memory for retrieval failures. We introduce FEB-30, a benchmark of 30 fact-evolution trajectories spanning 6 diagnostic categories and 3 difficulty levels, constructed via paradigm inversion: documents are generated from known fact trajectories, yielding deterministic ground truth without human annotation. Each trajectory defines how a fact should evolve across multi-author artifacts (documents, emails, conversations), enabling direct inspection of the memory store after each ingestion step. Applying FEB-30 to a representative memory system, we demonstrate that intrinsic and extrinsic evaluation diverge by an average of 16 percentage points, with 44% of trajectories exceeding meaningful divergence thresholds. Diagnostic decomposition reveals that the system correctly classifies memory operations (create, update, reinforce, supersede) only 20% of the time, a failure entirely invisible to extrinsic evaluation, and that failure modes differ systematically by category and difficulty. We release the benchmark, document generation pipeline, and scoring framework to support reproducible evaluation of agent memory systems. --- ### Decentralized LLM Workflows via Sequential Auctions Authors: Ian Gemp, Shubhendu Trivedi, Adam Fisch, Fantine Huot, Michael Kaisers, Mirella Lapata, William W. Cohen, Alekh Agarwal, Kate Larson, Jacob Eisenstein Published: 2026-10-15 Categories: cs.GT, cs.AI, cs.MA arXiv: https://arxiv.org/abs/2610.22201v1 We present a scalable framework for solving multi-step tasks through a decentralized market of self-interested agents. Prior theoretical work demonstrated that agents organized through sequential auctions can converge to a globally optimal policy in a Markov decision process (MDP), but relied on the restrictive assumption of the existence of perfect clones as a model of perfect competition. In reality, exact clones drive agent profit to zero, violating the natural risk-reward incentives of participants. Conversely, we empirically demonstrate that the opposite extreme, absence of competition, also leads to market failure and value extraction collapse. Between these two extremes, we investigate how heterogeneous agents can survive in markets that span the critical middle ground of imperfect competition. To be profitable, agents learn to strategically shade bids to secure positive individual profit margins, while still ensuring that the marketplace functions effectively enough so that collaborative tasks posed by market consumers can still be solved. We formulate and evaluate four bidding algorithms (Mirror Descent, Thompson Sampling, Bayes-Optimal, and PPO) across three auction protocols (1st-Price, 2nd-Price, and Credit-Conserving Vickrey). Across a diverse set of domains, we demonstrate that imperfect competition successfully sustains necessary agent profit margins while maintaining high task success rates at low costs. --- ### OpenGCUA: Toward a Fully Open Stack for Generalist Computer Use Authors: Hanzhang Zhou, Long Li, Quyu Kong, Lei Wang, Panrong Tong, Long Chen, Tianyu Xia, Gongjie Zhang, Xu Zhang, Chen Liu, Jianan Zhang, Chenglin Cai et al. Published: 2026-10-15 Categories: cs.CV arXiv: https://arxiv.org/abs/2610.22200v1 Computer use has reached a turning point: the focus is expanding from GUI-only interaction to systems that complete tasks on users' computers through visual and programmatic interfaces. We define a generalist computer-use agent (GCUA) as a system in which the model chooses how to observe, act, and organize its work across visual and programmatic interfaces. Yet frontier implementations, such as GPT-6 Astra's computer-use system, remain largely closed, limiting systematic study of their design. We introduce Open Generalist CUA (OpenGCUA), an ongoing project to develop and open the full stack for GCUAs with the community. This report presents the first release of OpenGCUA. Code execution is a key feature of our agent design: it lets intermediate results determine subsequent operations and enables models to apply their existing programming capabilities directly to computer interaction. We also propose harness-to-model decision transfer, moving selected observations and task-state decisions from fixed harness rules to the model. On OSWorld-v2, our full configuration improves task scores across all eight evaluated models, including GPT-6 Astra and Claude Opus 5, raising mean partial score by +19.6 points and reducing reported steps per task by an average of 53.0% relative to batched GUI-only execution. We release our implementation and prompts, evaluation trajectories from all eight models, and preliminary versions of a real-device harness and a cross-device SDK. The harness addresses real-device challenges not fully captured by benchmarks, while the SDK provides a common programming interface across computers, mobiles and browsers. Across four real macOS tasks under our real-device harness, selected Qwen3.8-Max runs use 19.4 times as many reported reasoning tokens as GPT-6 Astra, highlighting substantial differences between models in real-device execution efficiency. --- ### Hy-Memory-30B-A3B: Attributed Reinforcement Learning for Long-Term Memory Writing Authors: Tianxiang Fei, Jiazheng Kang, Mao Zheng, Xingyan He, Xiang Yu Published: 2026-10-15 Categories: cs.CL arXiv: https://arxiv.org/abs/2610.22196v1 External-memory systems extend language-model applications beyond a context window, but their quality depends critically on the model executing the write path: deciding what to retain, how precisely to represent it, and how new evidence should update persistent state. Existing frameworks typically place a general-purpose large language model behind framework-specific prompts. This raises serving cost, and a writer specialized for one protocol does not transfer directly to another; a unified model that can operate across established memory frameworks remains largely unexplored. We present Hy-Memory-30B-A3B, a compact long-term-memory writer trained across Mem0, MemOS, and Hy-Memory. Memory On-Policy Distillation teaches framework-specific write protocols on student-generated states; Memory RL then optimizes downstream answerability using dense per-chunk probes, attributed cross-chunk rewards, over-storage controls, and chunk-level GRPO. Carrying policy-written banks across windows keeps training aligned with deployment. On fixed 20% benchmark subsamples, Hy-Memory-30B-A3B ranks first on all four evaluated harnesses, including held-out EverMemOS, with a 55.49% macro average versus 54.62% for Claude Opus 5. --- ### HTN-Orchestrate: A Hierarchical Task Network Framework for Multi-Agent Orchestration Authors: Yanchen Deng, Jianguo Huang, Yao Long Teng, Yixuan Li, Handuo Zhang, Ning Wang, Liangcai Su, Xinyu Wang, Bo An Published: 2026-10-15 Categories: cs.AI, cs.CL arXiv: https://arxiv.org/abs/2610.22195v1 Orchestrating multiple agents on a complex task requires reconciling local subtask execution with global task planning. Dynamic dispatch approaches provide execution flexibility but do not explicitly model subtask hierarchies or dependencies, while predefined workflows impose this structure but fix the decomposition before execution begins. Therefore, we introduce HTN-Orchestrate, a two-layer framework that separates these responsibilities through a hierarchical task network (HTN). The Planner maintains subtask hierarchies and dependencies, deferring decomposition until prerequisite results become available, while local Orchestrators schedule workers to execute each leaf subtask. Results propagate upward under \textbf{AND} semantics, and accumulated findings inform a rebuilt task network when the current decomposition cannot resolve the task. On five benchmarks spanning web research, scientific reasoning, and research workflows, HTN-Orchestrate outperforms ReAct, Agent Team, and DeepSeek Harness using the same base model, achieving 89.73% accuracy on BrowseComp and 73.44% on SUPERChem, exceeding Agent Team by 2.45 and 4.69 percentage points, respectively. --- ### AutoSFT: An Agentic System for Automated Data Research in Supervised Fine-Tuning Authors: Bingxiang He, Pengqi Sun, Shihao Peng, Shaoxuan Li, Huiyu Xiao, Hongjian Shi, Zixuan Fu, Ziqing Qiao, Yuxin Zuo, Cheng Qian, Xinqiang Yu, Jiayi Li et al. Published: 2026-10-15 Categories: cs.AI, cs.CL arXiv: https://arxiv.org/abs/2610.22194v1 Language model agents increasingly automate post-training, but when given the full training stack they converge on supervised fine-tuning and spend most effort revising training data. Existing systems give the agent one kind of control, selecting rows from a fixed pool or calling a generator, whereas a human data engineer adjusts proportions, prunes, repairs, replaces and synthesizes the data. We design a data action space for these changes and ask whether an automated system can make the full range rather than one of them, and what structure that requires. We first run this loop with a single agent, and it reaches only one action, adjusting category proportions, while its context fills with execution output and is discarded mid-search. We therefore introduce \textbf{AutoSFT}, a multi-agent system where a meta agent separates execution from analysis, gathers proposals from independent angles, and calls a synthesis subagent for capabilities the pool cannot cover, with the model, the trainer, and the budget fixed within each comparison. Adding each component in turn helps: the executor halves context loss, the analysis subagents lift the best IFEval by $+6.5$, and synthesis adds $+3.4$ IFEval and $+2.9$ IFBench. The same loop transfers to code, where analysis adds $+3.7$ MBPP+ over a single agent that does not improve. Every result comes from runs that retrain from the base model each round, so each component's effect is individually attributable. Our findings suggest that automating SFT data research depends more on covering diverse data actions than on a stronger single action. --- ### Terminal-RSI: Recursive Self-Improving Terminal Agents via Recursive Self-Rewrite and Policy-Refine Authors: Junyao Yang, Zongxia Li, Yucheng Shi, Zhongzhi Li, Ruhan Wang, Ruiwen Zhou, Haitao Mi, Shuicheng Yan, Leowei Liang Published: 2026-10-15 Categories: cs.AI arXiv: https://arxiv.org/abs/2610.22192v1 Converting verifier outcomes into useful policy updates and partial progress into complete task resolution is crucial for terminal agents. We present Terminal-RSI, which alternates Recursive Self-Rewrite (RSR) of verified trajectories with critic-inheriting Recursive Policy-Refine (RPR). We formulate Turn-Boundary Skip-Observation GAE (TB-GAE) to retain token-specific baselines while applying decay across interaction turns. Family Dense Reward (FDR) calibrates semantic progress across task families; Success-Prioritized Reward (SPR) separates strict success from bounded partial credit. A Warm-up Critic Model, recursive critic handoff, and stage-specific learning rates coordinate the two refinement phases. The five-stage sequence from Qwen3.8-27B base model through RSR-1, RPR-1, RSR-2, and RPR-2 yields Terminal-RSI-27B: 19.84% on Terminal-Bench 4.0, 86.52% on Terminal-Bench 2.1, 16.95% on Terminal-Bench 3.0, and 79.00% on Terminal-Bench Hard. On LHTB, process reward rises from 0.21 to 0.37 and strict completion from 0/46 to 5/46; SWR100 success doubles from 3.00% to 6.00%. By recursively coupling verified trajectory rewriting with critic-inheriting policy refinement, each retained stage consolidates discovered behaviors and progressively advances the agent from partial local repairs to strict terminal verification. --- ### Attention over noisy examples: generalization and memorization in the exponential Hopfield network Authors: Francesco Alemanno Published: 2026-10-15 Categories: cond-mat.dis-nn, cs.LG, cs.NE arXiv: https://arxiv.org/abs/2610.22191v1 In Hebbian learning from examples a Hopfield network of $N$ neurons is not supplied with the $K$ archetypes it has to retrieve but with $M$ blurred examples of each, of quality $r$, and whether it can infer them depends on the load $α=K/N$ (interference among them) and on the dataset entropy $ρ=(1-r^2)/(Mr^2)$ (how little the examples say about them). We transfer this protocol to the exponential Hopfield model, whose zero-temperature update is a step of softmax attention with the unlabelled examples as keys. A signal-to-noise analysis gives in closed form the overlap between the output and an unseen archetype, $m^{(1)}={\rm erf}(1/\sqrt{2ρ_{\rm eff}})$: the majority-rule law of the pairwise model at small load, with $ρ$ replaced by an attention entropy $ρ_{\rm eff}$ built from the weight the softmax gives to the correct class and the number of examples it spreads it over. Two consequences follow. First, in a window of inverse temperatures $β$ that widens as $\sqrt N$ the interference between archetypes is exponentially small in $Nβ$: the load drops out and the number of archetypes that can be learnt grows as $\exp(c\sqrt N)$. Second, learning fails differently: in the pairwise model through a spin-glass phase, in the exponential model through memorization: the softmax concentrates its weight on the training example closest to the input and returns it in place of the archetype. Seen from the input the softmax weights are those of the random energy model, and memorization is its freezing transition, at $β\sqrt{N(1-r^2)}\simeq\sqrt{2\ln M}$; it is measurable on a head whose keys are examples as the number of examples carrying weight. On MNIST the diagnostic follows the random-energy prediction; on fashion-MNIST it detects sub-clusters in the classes, and the best accuracy lies beyond the memorization threshold, where the softmax reads out a local sub-prototype. --- ### Change the Number, Change the Answer: Assessing Rating Scale Sensitivity in LLM-Based Text Scoring Authors: Wonjeong Jo, Gyuwan Kim, Ming Shan Hee Published: 2026-10-15 Categories: cs.CL, cs.AI, math.NA arXiv: https://arxiv.org/abs/2610.22187v1 Large language models (LLMs) are increasingly used in computational social science to assign scalar scores to social and psychological constructs. However, studies use different rating scales, and whether their scores remain comparable across scale designs is understudied. To address this gap, we define and systematically investigate scale sensitivity, the variation in LLM-generated scores caused by rating scale design. We distinguish numerical sensitivity, changes in normalized scores, from semantic sensitivity, changes in score interpretation. To assess both, our experimental design crosses 2 formats (continuous vs. ordinal), 4 ranges (3, 5, 11, and 21 points), and 2 signs (signed vs. unsigned) across 3 tasks (political leaning, sentiment, and toxicity), 2 prompting strategies (zero-shot and rubric-based), and 5 LLMs. We find that changing one design element shifts normalized scores by about 0.10 on average and changes the category of roughly 20% of texts. Sensitivity is greatest for toxicity, the only unipolar construct examined. Also, rubric-based prompting does not reduce sensitivity compared with zero-shot prompting. We provide practical guidance for designing more reliable LLM-based text scoring and highlight the need for robustness checks across alternative scales. --- ### EVA-Cap: Optimizing Audiovisual Video Captioning via Event-Centric Alignment Authors: Jiapeng Shi, Qiuxia Hou, Jiaqi Leng, Junke Wang, Yanhao Zhang, Haonan Lu, Ziyi Ye, Zuxuan Wu Published: 2026-10-15 Categories: cs.CV, cs.CL arXiv: https://arxiv.org/abs/2610.22186v1 Recent advances in Omni Language Models (OLMs) provide a unified solution for audiovisual video captioning. Nevertheless, existing approaches predominantly treat captioning as an unstructured text generation process, neglecting the inherent semantic structure of videos and leading to suboptimal fine-grained audiovisual alignment. To bridge this gap, we propose EVA-Cap, a novel framework that decomposes captions into atomic audiovisual events and transforms unstructured text supervision into event-centric alignment. We train EVA-Cap via Supervised Fine-Tuning (SFT) and Group Relative Policy Optimization (GRPO) on a self-curated dataset, EVA-Data, which is built with event-centric quality control. During the GRPO phase, the model is explicitly optimized against three event-centric alignment targets: global event completeness, intra-event attribution, and inter-event synchronization. Extensive experiments demonstrate that EVA-Cap significantly outperforms existing methods in event-centric evaluation (e.g., +9.0% average gain on ChronusAV), while achieving competitive performance on general audiovisual captioning benchmarks (e.g., +3.3% on WorldSense). --- ### Whole-Body Dexterity Authors: Conor Mc Gartoll, An Dang, Pranay Thangeda, Arjun Gupta, Carolina Higuera, Nima Fazeli, Mustafa Mukadam, Manikantan Nambi Published: 2026-10-15 Categories: cs.RO, cs.AI, cs.LG, eess.SY arXiv: https://arxiv.org/abs/2610.22184v1 Current humanoid dexterous manipulation remains largely confined to end-effectors and simple grasps, falling short of humans' ability to engage the entire body by bracing and sliding bulky objects against the arms and torso. We introduce Whole-Body Dexterity (WBD), a general framework that uses the robot's body and appendages as contact surfaces to grasp and manipulate objects. Our sim-to-real pipeline first generates a dataset of feasible whole-body grasps, then trains reinforcement-learning experts on a general any-to-any repositioning objective by sampling initial and goal grasps from the dataset, and distills these experts into a single policy that uses proprioception and object geometry. We show that the distilled policy transfers zero-shot to a real Unitree G1, where it repositions a basketball and a cardboard box to regions around the torso and completes a long-horizon task that stows a box under one arm, carries it from a shelf to a table, and places it. The controller operates interactively, steered by coarse motion references from a human teleoperator. Across real-world whole-body manipulation tasks, the distilled controller completes 90% of trials, compared with 18% for direct retargeting and 4% for a generalist whole-body controller. --- ### LAMP: Layerwise Adaptive Mixing of Predictors for Training-Free Few-Shot Vision-Language Adaptation Authors: Zhipeng Ye, Feng Jiang, Qiufeng Wang, Hao Li Published: 2026-10-15 Categories: cs.CV arXiv: https://arxiv.org/abs/2610.22182v1 Few-shot adaptation often relies on an encoder's final global representation, leaving complementary information in intermediate layers underused. Combining these features is challenging because their dimensions can differ and their usefulness varies across tasks. We propose LAMP, Layerwise Adaptive Mixing of Predictors, a training-free framework that combines global and regional information from multiple encoder layers. A single forward pass per view supplies all descriptors. Global and support-calibrated regional kernel predictors independently fit the same label residuals. Labeled validation data select their mixture in class-score space, preserving native feature dimensions without learned projections. LAMP-F further adapts text-anchored class prototypes, using a cached response to update the fixed-kernel residual predictions exactly. Experiments cover eleven datasets and ViT and ResNet backbones. LAMP achieves state-of-the-art mean accuracy among the compared published training-free ResNet-50 methods at 4, 8, and 16 shots, reaching 78.73% at 16 shots. Averaged across five support budgets, LAMP improves over a paired global-feature-only baseline by 1.14 and 1.59 percentage points on ViT-B/16 and ResNet-50; LAMP-F raises these gains to 1.96 and 2.54 points. LAMP-F reaches 16-shot accuracies of 83.57% and 79.10%, respectively. Controlled experiments identify multi-layer integration as the main source of improvement and examine pooling, calibration, validation labels, and computational cost. These findings show that combining layer-specific predictions can exploit complementary features from frozen encoders to improve few-shot classification. --- ### LCT: Losslessly Compressed Tensors with GPU friendly execution for Training Authors: Max Zhu, Bohang Sun, Pietro Lio, Jialin Yu Published: 2026-10-15 Categories: cs.LG arXiv: https://arxiv.org/abs/2610.22179v1 Lossless tensor compression could reduce memory usage during training, but training requires tensors to be repeatedly re-encoded. We introduce Losslessly Compressed Tensors (LCT), a GPU execution framework designed for this repeated encode-decode regime. LCT exploits the natural redundancy in floating point data to reduce storage for weights, activations and optimizer states. Highly optimized GPU kernels implement Huffman coding with precomputed codebooks for low compression and decompression overhead, essential for training. Our analysis characterizes LCT's storage costs and adaptation to magnitude shifts. LCT achieves a compression ratio of up to $1.455\times$, and is demonstrated in training and finetuning settings. On a 162M-parameter NanoGPT, LCT reduces memory usage by $24.6\%$ for a $7.2\%$ step-time increase. In Qwen3-4B fine-tuning, LCT saves up to $23.8\%$ of peak memory, enabling a $3.5\times$ larger context length in training on the same GPU. --- ### REKI: Registering Knowledge in VLAs Authors: Juliette Marrie, Paul Pacaud, Adrien Ramanana Rahary, Shizhe Chen, Moritz Böhle Published: 2026-10-15 Categories: cs.RO, cs.CV, cs.LG arXiv: https://arxiv.org/abs/2610.22178v1 We introduce REKI, a 2.4B vision-language-action model (VLA) pretrained exclusively on public data that sets a new Pareto frontier between performance and model size: on LIBERO-Plus, RoboTwin Clean2Rand and RoboCasa365, it is outperformed only by models about twice its size. REKI consists of a vision-language model (VLM) followed by a flow-matching action expert, and is pretrained on a curated mixture of public robot, egocentric human and vision-language data. Additionally, we endow it with a memory that gives the policy access to its recent past at negligible cost: a few register tokens in the action expert summarise the output of the VLM at each frame and persist in the expert's cache. This memory improves success in simulation and on real robots, and makes oscillations between poses about four times rarer. REKI uses a simple model design and training paradigm, and is optimised to be efficient at training and inference. We release the full code across data curation, training, inference and evaluation, deployment, and all pretrained and finetuned model weights: https://github.com/kyutai-labs/reki. --- ### Motion- and Gravity-Aware Feedforward 3D Reconstruction Authors: Shiyao Qian, Yuan Ren, Dongfeng Bai Published: 2026-10-15 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2610.22177v1 Feed-forward transformer models have recently achieved strong performance in multi-view 3D reconstruction, but they still rely primarily on appearance-based attention and often struggle in dynamic scenes with weak geometric cues. In this work, we introduce a motion- and gravity-aware framework under the Learning Using Privileged Information (LUPI) paradigm. During training, residual optical flow and gravity direction are provided as privileged signals to guide geometric representation learning; during inference, the model uses standard visual inputs only. We integrate these priors into transformer backbones with lightweight architectural changes, preserving feed-forward efficiency. Experiments on standard benchmarks show consistent gains in reconstruction accuracy and robustness, especially in scenes with dynamic objects, ambiguous structure, and challenging camera motion. Our analysis further shows that the optimal injection stage is backbone-dependent: middle-layer integration is generally preferable for VGGT, while late-stage integration is more effective for π3. These results suggest that LUPI offers a practical way to inject motion and physical priors into feed-forward 3D reconstruction without increasing test-time complexity. --- ### Mosaic: Multi-Session Agentic Reinforcement Learning Authors: Yilong Xu, Minrui Xu, Xinyu Zhou, Xiang Long, Haochen Tan, Zhijiang Guo Published: 2026-10-15 Categories: cs.CL, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2610.22174v1 Most existing agentic reinforcement learning methods assume single-session trajectories, where interaction proceeds within one continuously accumulated context. However, agentic techniques such as context compaction and subagent forking may produce trajectories spanning multiple sessions, creating a mismatch between the single-session training assumption and inference-time execution. Therefore, we introduce Multi-Session Agentic Reinforcement Learning, a more general setting that supports an arbitrary number of sessions within a trajectory. Under this setting, a widely adopted method is to broadcast trajectory-level advantages to all sessions, but this does not distinguish contributions from various sessions or guide session allocation within trajectories. To address this issue, we propose Mosaic, a multi-session agentic RL method with critic-free hierarchical optimization. At the trajectory level, it incorporates session-usage signals into advantage estimation. At the session level, it shapes each session's advantage according to the estimated marginal contribution. Experiments on BrowseComp with Qwen3.5-4B show that Mosaic outperforms broadcasting in both context compaction and subagent forking scenarios while using fewer sessions on average. Further analyses show that our method achieves improvements in session utility and allocation, which we suggested is a promising target for future advances in multi-session agentic RL. --- ### Lexical Reference Frames: Interpreting LM Residual Streams in Native Semantics Authors: Jian Gu, Hongyu Zhang, Chunyang Chen, Aldeida Aleti Published: 2026-10-15 Categories: cs.CL, cs.LG arXiv: https://arxiv.org/abs/2610.22169v1 The vocabulary of a language model is its native semantics: the embedding and the unembedding fix the meaning of a state at the two ends of the residual stream. Between them, existing methods either reuse a fixed lens or treat the states as a point cloud, and neither reads a state the way the model does. We argue that every hidden state has a native reference frame, induced by the layers that read it next and adaptive to the context. Lexical reference frames approximate it by reading a state through the model's own remaining layers with a stated number of rounds of attention between positions and a typical attention output for the rest, without fitting to the model's output. No lens, with fixed directions per layer, can exceed the square root of the native frame's coherence, 0.30 and 0.35 at mid-depth of SmolLM3-3B and OLMo3-7B, whereas the first-order frame, LexFrame(1), reaches 0.97 and 0.97 and reads a state within 0.16 nats of the model's output. Its coefficients attribute predictions more faithfully than attribution patching, its lexical basis steers a prediction while sparing other tokens, and its confident readings would let an early exit skip 0.45 and 0.42 of the layers, where lenses skip almost none. In this way, lexical reference frames make every hidden state of the model readable, attributable and steerable. --- ### CLUE: Uncovering Social Context for Pragmatic Understanding Authors: Chengfei Wu, Zhaoqing Wu, Dan Goldwasser Published: 2026-10-15 Categories: cs.CL arXiv: https://arxiv.org/abs/2610.22167v1 Understanding a conversation requires connecting what is said to the speakers' relationships, expectations, and earlier exchanges. Prior work shows that supplying relevant context can improve models' social judgments, but much of this information may already be present in the conversation. This motivates training models to discover the cues that matter and use them to support an interpretation. We propose CLUE, a two-stage training pipeline that combines supervised evidence reasoning with GRPO guided by Grounded Information Utilization ($\mathrm{IU}_{\mathrm{G}}$). This reward credits new, evidence-supported social inferences while accounting for reasoning effort. We train on a small set of conversational examples to assess effectiveness and generalization across tasks, languages, and genres. At inference time, the trained model reasons over the supplied interaction in a self-contained manner. Compared with base Qwen, CLUE achieves relative gains of 17.3% in-domain and 9.3% out-of-domain in average weighted F1, with higher F1 on four of five out-of-domain evaluations. These gains are accompanied by over 80% fewer reasoning tokens on average in both domains. Human evaluation further finds that CLUE more often grounds its explanations in specific conversational content, with closer alignment between its quoted evidence and human-selected context. ---