# Research Papers — 2026-09 · bundle page 73 of 266 > Full-text markdown of papers #3601–3650 of 13298 published in 2026-09, newest first. AI-readable mirrors of the full 13298-paper 2026-09 archive (50 papers per file). Canonical page: https://aipod.dev/papers/archive/2026-09/page/73 [← Previous bundle](https://aipod.dev/papers/archive/2026-09/page/72.md) [Next bundle →](https://aipod.dev/papers/archive/2026-09/page/74.md) ### SparSP: Exploiting Communication Sparsity for Sequence-Parallel Video DiTs Authors: Desen Sun, Xinrui Zhong, Yuke Wang, Sihang Liu Published: 2026-09-26 Categories: cs.DC, cs.LG, cs.PF arXiv: https://arxiv.org/abs/2609.32197v1 Diffusion Transformers have become the dominant architecture for video generation. Their substantial computational cost motivates scaling inference across multi-GPU servers, yet efficient scaling remains challenging on commodity GPUs connected via PCIe, whose bandwidth is limited. Although sparse attention substantially reduces computation, its implications for communication remain underexplored. This paper argues that attention sparsity should be treated as a communication primitive. We present SparSP, an efficient sparse sequence parallel communication system that co-designs token distribution, communication routing, and asynchronous execution for sparse video diffusion models. First, Dependency-Aware Placement distributes sequence blocks according to diffusion models' sparse attention patterns. Second, Demand-Directed KV Routing transfers KV blocks directly to requesting GPUs without intermediate relays. Third, a Decoupled Transfer Runtime separates communication from GPU computation to reduce resource contention and maximize effective bandwidth. Our evaluation shows that SparSP improves attention performance by 1.38- 1.5$\times$, achieves an average 1.17$\times$ (up to 1.69$\times$) end-to-end speedup across three representative servers and three video diffusion models, and reduces communication volume by 12.54-23.05%. Moreover, we achieve an average 1.53-1.76$\times$ bandwidth improvement over NCCL. --- ### The Judge Is Not Its Twin: Post-training makes a model's writing more predictable but barely moves its taste, as a judge, toward predictable writing Authors: Arman Nik Khah, Arvin Bahreini Published: 2026-09-26 Categories: cs.CL, cs.LG arXiv: https://arxiv.org/abs/2609.32196v1 Language models are now routinely graded by other language models. If post-training makes a model's own writing more predictable, it may also teach the same model, acting as a judge, to reward predictable writing, so that progress on creativity would be invisible to automated evaluation. We follow two open model families, OLMo-2 and Zephyr (7B parameters each), through their public training stages and measure every stage twice, as a writer of short stories and as a judge of pairs of stories. As writers, the models drift as feared: each family's fully trained model finds the stories of its base, supervised fine-tuned (SFT) and preference-trained (DPO) stages progressively more familiar, in all ten prompts, and a model from the other family finds the trained stories 7.0 to 9.2 percent (OLMo-2) and 24 to 25 percent (Zephyr) less surprising per token. As judges, they barely move toward predictable writing. Asked which story is better, a question every judge can use to tell a story from its own words in scrambled order, no trained judge's estimated tilt toward the more predictable story grows by as much as one point in the probability of picking it. A post hoc one-sided 95% upper bound on that growth is 2.7 points on an average pair, about the size of the untrained OLMo-2 judge's own tilt. Asked which is more creative, no trained judge's estimate favors the predictable story more than its base's does. Training instead strengthens a preference for longer stories when the question is creativity, and for one answer slot, and it breaks "more creative" as a question: trained judges asked it no longer reliably prefer a story to its scrambled words. A follow-up could not build pairs that differ in predictability but not in quality, because the routes that made this writer's stories less predictable also broke some of them, often enough to fail a quality floor set in advance. --- ### High-Probability Guarantees for SGD under $β$-Heavy-Tailed Gradient Noise Authors: Qijun Tong, Masahiro Ikeda, Ryota Kawasumi Published: 2026-09-26 Categories: stat.ML, cs.LG arXiv: https://arxiv.org/abs/2609.32195v1 Stochastic gradient descent (SGD) is widely used to train machine learning models, but subsampling the training data introduces noise into its updates. The strength and applicability of high-probability guarantees therefore depend critically on how the tails of gradient noise are modeled. Reports of heavy-tailed gradient noise in deep learning motivate relaxing the bounded-noise and sub-Gaussian assumptions commonly used in high-probability analyses of SGD. We use Young functions from Orlicz space theory to describe noise tails in a common framework. We model SGD gradient noise by adopting a Young function that preserves the finiteness of all polynomial moments while allowing tails heavier than sub-Weibull, including lognormal distributions. The resulting class is called $β$-heavy-tailed, with $β$ controlling the tail heaviness. We establish concentration inequalities for $β$-heavy-tailed noise and combine them with a uniform bound on the difference between empirical and population gradients along the SGD trajectory to obtain high-probability bounds on optimization and population-risk stationarity for smooth nonconvex losses under trajectory assumptions. The bounds are not restricted to a particular learning-rate decay rule and make explicit the effects of noise tails and learning-rate schedules. Under the Polyak-Łojasiewicz condition, we bound the risk at the last iterate. We also analyze SGD with gradient clipping under the $β$-heavy-tailed noise model. --- ### Devol-ONE: One Autoregressive Mixture of Transformers to Unify Vision-Language-Action and Latent World Modeling Authors: Hongyi Cai, Yi Herng Ong, Tingshiuan C. Wu, Lim Chiew Hui, Hanxia Li, Kehong Guo, Sze Yuan Cheong Published: 2026-09-26 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.32193v1 Vision Language Action (VLA) models condition actions directly on current visual and language context, without an explicit account of how the scene evolves under candidate actions. World Action Models (WAM) attempt to address this limitation by predicting future states, but existing designs keep prediction and policy learning architecturally separate, connecting them only through the predicted output, whether through pixel space video generation or a latent forecasting module trained independently of the policy. We present Devol-ONE, a Mixture of Transformers architecture that unifies vision language understanding, latent world dynamics prediction, and action generation within a single autoregressive framework. Instead of encoding vision language tokens once and feeding them to the action expert, Devol-ONE runs autoregressive prediction jointly across a vision language stream and a V-JEPA pretrained dynamics stream, attending to the vision language key-value cache at every layer to forecast future latent states under language guidance. The action expert is in turn shaped continuously by semantic reasoning and predicted physical dynamics rather than by a fixed representation computed in advance. Extensive experiments are conducted on LIBERO, LIBERO-PLUS, RoboTwin2.0 along with real-world evaluation on Flexiv single-arm and dual-arm setups. Ablation studies show the effectiveness of dynamic stream prediction and layer-wise unified attention to validate our model architectural coherency. --- ### CoMemBench: Benchmarking Collaborative Memory Boundaries across Multi-Agent Workflow Topologies Authors: Sen Zhao, Ruiqi Kong, Zuyu Zhang, Lifeng Shen, Xinyu He, Ding Zou, Xu Zhang, Qinghua Zhang Published: 2026-09-26 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.32192v2 Multi-agent workflows require task-relevant information to be shared across agents, while irrelevant, stale, unverified, or incompatible information must remain isolated. We call this task-conditioned scope of information a collaborative memory boundary. Workflow topology determines which intermediate artifacts are applicable to which downstream workers and when they cease to be valid, thereby providing a structural stress dimension for sharing and isolation. Existing memory benchmarks primarily evaluate retention and retrieval, whereas multi-agent benchmarks emphasize coordination and end-to-end completion, leaving topology-conditioned memory boundaries largely unmeasured. We introduce CoMemBench, an execution-grounded benchmark for collaborative memory sharing and isolation across multi-agent workflow topologies. It constructs 800 composite workflows across four domains from source-grounded dependency graphs, with node-local specifications, verifiable artifact handoffs, native evaluators, and matched isolation challenges. CoMemBench measures workflow completion, verified node progress, required-handoff reliability, isolation robustness, and token cost. Experiments reveal a sharing-isolation trade-off: broader context improves information availability but can weaken isolation, while system rankings shift across topologies and artifact violations. --- ### Evaluating Single and Multi-Omics Based Explainable Artificial Intelligence (MOXAI) for Molecular Subclass Classification of Adult-Type Diffuse Gliomas Authors: Md Zahangir Alom, Quynh T. Tran, Breuer Alexandar, Brent A. Orr Published: 2026-09-26 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2609.32190v1 DNA methylation (DNAM) profiling has emerged as a powerful diagnostic tool for classifying brain and solid tumors. However, existing computational models typically analyze methylation and copy number variation (CNV) data separately, failing to capture the complementary information their integration could provide. Moreover, current classification models lack mechanisms for within-class risk assessment analogous to traditional tumor grading, and no established explainability method can attribute classification decisions to specific genomic loci. In this paper, we present MOXAI (Multi-Omics Based Explainable AI), a deep learning framework that integrates DNA methylation and copy number data from methylation arrays to classify molecular subtypes of adult-type diffuse gliomas, alongside single-modality variants for comparison. Using a cohort from The Cancer Genome Atlas (TCGA), we trained ResNet50, DINOv2, and Graph Attention Network (GAT) models on methylation data alone, copy number data alone, and combined multimodal data. We further developed explainable AI (XAI) methods based on class activation maps (CAMs) and gradient-weighted CAM (Grad-CAM) to identify the specific CpG sites, genes, and chromosomal regions most relevant to each classification decision. The multimodal model achieved up to 92.98% cross-validation accuracy, outperforming models trained on CNV data alone. DINOv2 showed the strongest generalization, reaching 94.25% accuracy (confidence >0.9) on independent validation sets. XAI results aligned with established molecular features of adult-type diffuse glioma subtypes, confirming the biological interpretability of the framework. --- ### ScopeIF: Improving Scope-Aware Precise Instruction-Following in Large Language Models via Graded Reward Modeling Authors: Bosi Wen, Yilin Niu, Xiaoying Ning, Ying Zhang, Hongning Wang, Minlie Huang Published: 2026-09-26 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.32189v1 Precise instruction-following is a fundamental ability of large language models (LLMs), requiring their outputs to strictly satisfy objective constraints in input instructions. In complex application scenarios, these constraints often possess diverse scopes that govern specific response segments rather than the entire output. However, existing optimization methods often neglect constraint scope during data construction and rely on binary per-constraint rewards, yielding limited data diversity and sparse supervision for complex constraints. To this end, we propose ScopeIF, a novel training framework for scope-aware precise instruction-following. We first introduce a unified schema that factorizes objective constraints into three decoupled dimensions: Scope, Target, and Range. Grounded in this schema, we construct ScopeInstruct, a large-scale instruction dataset with diverse scope-aware constraints, and combine tool-grounded verification with graded reward modeling to quantify the violation degree of each constraint, providing dense supervision for policy optimization. Extensive experiments demonstrate that ScopeIF consistently outperforms existing methods, particularly on complex scope-aware constraints, while preserving general capabilities. Notably, it enables optimized Qwen3-4B and 8B models to rival or surpass strong frontier models such as Gemini-2.5-Pro and DeepSeek-V3.2, establishing an effective paradigm for advancing scope-aware instruction-following. Our code and data are available at https://github.com/thu-coai/ScopeIF. --- ### Presence Is Not Faithfulness: Figurative Vehicle Intrusion in Text-to-Image Generation Authors: Xiaoyu Ma, Chen Yang, Hao Chen Published: 2026-09-26 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2609.32188v1 Text-to-image (TTI) models increasingly generate high-quality images from natural-language prompts, yet figurative language exposes a failure: a vehicle that should guide the depiction of a tenor may instead be rendered as a visible object. We call this failure Figurative Vehicle Intrusion: the intruding content is textually licensed, but it is assigned the wrong visual role, showing that visual presence is not always faithfulness and that presence-oriented evaluation can miss such errors. To study it systematically, we introduce Vehicle Intrusion and Semantic Tenor Assessment (VISTA), a multilingual benchmark of figurative prompts organized by Figurative Form and Mapping Mechanism. We further propose V-Score, a diagnostic question-answering metric that evaluates role-aware figurative faithfulness in generated images. Evaluations on recent high-performing TTI models show that vehicle intrusion persists across languages and figurative categories. As a lightweight mitigation, we introduce VISTA-Guard, which partially reduces vehicle intrusion and suggests a practical path toward more figuratively faithful TTI generation. All resources will be released publicly. --- ### Rethinking Cross-Channel Importance in Time-Series Forecasting Authors: Yong-Hoon Choi, Kwang-Hyun Park, Youngjin Cho Published: 2026-09-26 Categories: cs.LG, cs.IR arXiv: https://arxiv.org/abs/2609.32187v1 Cross-channel modeling is central to multivariate time-series forecasting, yet channels that are statistically related, predictively useful, and actually used by a trained forecaster are often treated as if they defined the same notion of importance. We show that they need not coincide. Cross-channel dependency structures change substantially across future offsets, and horizon-adaptive source selection improves a controlled Ridge predictor in 21 of 32 dataset--prediction-length conditions, with a mean gain of $5.16\%$. This selected-set signal also transfers to a matched nonlinear predictor. Yet imposing the same horizon-specific source logic on iTransformer yields only 11 of 20 wins and a mean gain of $0.208\%$, with little alignment between controlled and neural gains. Functional interventions further show that strong forecasters use cross-channel information, while their source-reliance rankings agree little with controlled utility or with one another across iTransformer, TimesNet, and a cross-channel TimeMixer. As a constructive consequence, bounded post-hoc support improves a frozen channel-independent forecaster in 12 of 16 dataset--horizon conditions, with a positive aggregate bootstrap interval. Cross-channel importance should therefore be interpreted relative to the forecasting mechanism and question that define it: related $\neq$ useful $\neq$ used. --- ### Contamination, Prior, or Evidence? Decomposing and Training Evidence Use in Whole-Slide Vision-Language Models Authors: Wenhao Zhang, Zhongliang Zhou, Shiyuan Zhang, Yiqing Yang, Pinqiao Wang, Lehan Yang, Hanyin Wang, John Kang, Sheng Li Published: 2026-09-26 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2609.32185v1 Pathology vision-language models (VLMs) are conventionally evaluated by accuracy, but accuracy alone does not measure evidence use: it may conflate dataset contamination, prior knowledge, and image evidence. In a motivating study of lymph-node metastasis prediction, we found that most public pathology VLMs showed minimal differences when changing from feeding the models with whole-slide images, an annotated lesion, or no image at all. To better understand the specific features leveraged by these models, this paper presents two contributions aimed at disentangling these factors. First, we present CleanSlide, a TCGA-based VQA benchmark designed to eliminate image- and question-side contamination. It contains 149K audited multiple-choice questions over 9,985 slides, with patient- and tissue-source-disjoint splits. Every question is audited for option shortcuts, stem leakage, cross-split duplication, and blind solvability. Second, we propose Pair-DPO, a preference loss over counterfactual slide pairs from the same question and source. By controlling for shared confounding factors, Pair-DPO cancels out the question-attributable signal and leaves image evidence as the source of preference. Specifically, each pair consists of two real slides with opposite, verified findings, introducing neither editing artifacts nor unverified labels for diffuse or graded features such as invasion, necrosis, and tumor grade. Experiments show that our method gains 15.29% from image evidence on the CleanSlide, compared with 2.81% for the best published model. On the external CPTAC and BCNB cohorts, our method achieves accuracies of 57.6% and 59.0%, outperforming all other evaluated models by 9.7% and 3.4%, respectively. We will release the benchmark and code. --- ### AI Harness: Certification under Proposal-Conditioned Information for Foundation-Model Agents Authors: Hailin Zhong, Shengxin Zhu Published: 2026-09-26 Categories: cs.AI, eess.SY arXiv: https://arxiv.org/abs/2609.32184v1 Foundation-model agents are often modeled as policies over an observed state. In deployed systems, however, a runtime may intervene only after the model has emitted a semantic proposal, making the proposal both an action candidate and a decision-time observation generated by a history-conditioned process. We show that collapsing this structure into a state-only proposal envelope can preserve proposal coverage while destroying certifiability. In a finite robust interface, the viability kernel of the collapsed model is contained in the physical projection of the history-augmented kernel, and the collapse is lossless exactly when every proposal-conditioned collapsed fiber retains a common robust-safe intervention. This gap can be maximal even with constant-size proposal and history alphabets. The same common-action condition yields a dual result: observing the current proposal can restore robust feasibility when it separates latent modes requiring incompatible interventions. We extend these one-step results over time using exact finite beliefs and standard safety and reachability fixed points, separating indefinite operational viability from finite worst-case verified progress. Controlled model-in-the-loop tests reproduce the predicted obstructions when telemetry or effect verification is removed or intervention authority is restricted. Thus, our contribution is not a new fixed-point calculus, but a characterization of when proposal--history correlation at the model--tool boundary is necessary for certification. --- ### Scalable In-Domain Self-Supervised Foundation Model for Dense Representation Transfer in High-Resolution Plant Imaging Authors: Junlin Guo, Sharmin Majumder, Isaac Lyngaas, John Lagergren, Xiao Wang Published: 2026-09-26 Categories: cs.CV, eess.IV arXiv: https://arxiv.org/abs/2609.32183v1 High-resolution plant imaging enables detailed characterization of plant morphology, but dense scientific analysis remains limited by costly pixel-level annotations, large image pixel dimensions, and substantial variation in imaging conditions. This work proposes a scalable in-domain self-supervised pretrained foundation model for high-resolution, high-pixel-dimension multi-species plant imagery. A masked autoencoder with a ViT backbone is pretrained on more than 10 million multi-view plant image tiles using distributed training. Following scalable pretraining, the learned foundation-model representations are comprehensively benchmarked across fine-grained dense prediction and coarse global feature recognition, with particular emphasis on limited supervision and realistic downstream imaging conditions. This work focuses on the domain gap of existing foundation models in dense feature representation and transfer. Through extensive experiments involving limited annotations, cross-view variation, and resolution degradation, the in-domain FM achieves a Mean Dice of 0.8686 and a Pooled Dice of 0.8959, outperforming an MAE counterpart pretrained on large-scale natural-image data by 0.0694 and 0.0613, respectively. The results further indicate that increasing pretraining scale produces consistent improvements in dense feature transfer. Overall, these findings suggest that scaling in-domain self-supervised pretraining can reduce the domain gap and improve transferable dense representations for high-pixel-dimension scientific imaging. --- ### KeyRec: Bounded Visual Memory for Streaming and Long-Video Understanding Authors: Zihan Chen, Xuejian Rong, Xiaojuan Wang, Boqing Gong, Adi Zicher, Yael Pritch, Nikhil Karnad Published: 2026-09-26 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2609.32182v1 Vision-language models are increasingly used to understand long videos and continuous streams. However, dense visual tokens accumulate with video duration, making long-context inference prohibitively expensive. Existing training-free visual-token selection methods reduce this cost by retaining informative tokens, but may lose coherent event evidence and fail to distinguish detailed recent observations from long-range history. We propose \textbf{KeyRec}, a training-free framework for constructing bounded visual memory. During query-agnostic writing, KeyRec preserves fine-grained recent observations in a visual cache and organizes historical evidence into a structured event bank. Candidate events are proposed according to their novelty relative to previously stored events and maintained through an online add--merge--evict update. When a question arrives, a text-only router adaptively allocates a fixed readout budget between recent and event memory, without reprocessing historical frames. KeyRec operates on model-facing visual embeddings and supports both modular encoder--projector VLMs and the encoder- and projector-free NEO-ov architecture. Across four streaming and long-video benchmarks and three VLM backbones, KeyRec achieves the best compressed performance in 13 of 15 settings using only 10\% of the dense decoder-facing visual-token budget. It outperforms the strongest compressed baseline by 2.21--18.37 points on real-time questions, achieves the best compressed result in five of six long-video settings, and performs best in every NEO-ov 2B setting. --- ### Binaural Audio-Visual Instance Segmentation Authors: Saijun Wang, Guanfeng Tang, Hongbo Zhao, Zhicheng Lei, Yutong Zhang, Wei Ye, Rui Fan Published: 2026-09-26 Categories: cs.CV, cs.SD, eess.AS arXiv: https://arxiv.org/abs/2609.32180v1 Audio-visual segmentation (AVS) aims to segment sounding objects at the pixel level by integrating auditory and visual cues. However, existing methods are predominantly developed under the monaural setting and primarily rely on cross-modal semantic correspondence, which limits their ability to distinguish visually similar instances of the same semantic class. In contrast, humans naturally exploit binaural hearing, where interaural differences and direction-dependent acoustic filtering introduced by the head and pinnae provide physically grounded spatial cues for accurate sound source localization. Motivated by this observation, we introduce binaural audio-visual instance segmentation (BiAVIS), a new task that leverages synchronized binaural audio and video frames to segment sounding instances. To advance research on this task, we establish two benchmarks by manually annotating an existing binaural audio-visual dataset and collecting a new real-world dataset, BiAVIS-Bench, in more challenging and diverse scenarios. We further propose a BiAVIS model, which leverages an audio-only sound source localization network to learn spatial and semantic priors for sounding instances from binaural audio. A query-level audio-visual fusion strategy is subsequently introduced to inject these informative priors into the instance segmentation decoder. Extensive experiments conducted on the two proposed benchmarks demonstrate the superior performance of the BiAVIS model over previous monaural AVS methods, especially in resolving instance-level intra-class ambiguity. On the more challenging BiAVIS-Bench, the proposed BiAVIS model outperforms the best-performing monaural baselines by 17.22\% in mAP and 7.71\% in FSLA, respectively. --- ### Analytic-Walk Rotary Positional Encodings for Graphs Authors: Jiaqing Xie, Yuxin Wang, Xipeng Qiu Published: 2026-09-26 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.32178v1 Rotary position encodings make attention sensitive to relative position, but extending them to graphs requires choosing how graph structure enters the rotation. Previous works assign each node a rotation from spectral coordinates, so the rotary factor between two nodes depends only on their endpoints and cannot distinguish the routes connecting them. We introduce \textit{Analytic-Walk Rotary Positional Encodings} (AW-RoPE), which place the rotations on edges and sum the transported features over all walks, so contributions along different routes can reinforce or cancel. An exact variant evaluates the complete sum by a differentiable linear solve, and a sparse variant truncates it at a finite depth. We prove forward and parameter-derivative truncation bounds at fixed inputs and parameters. Both variants act on projected queries and keys, and the sparse recurrence also augments message-passing networks. Across five synthetic tasks both variants reduce nRMSE by $15$--$58\%$ relative to the strongest baseline, and on real superpixel, peptide and OGB benchmarks the sparse recurrence attains the best mean on every dataset with Performer kernels and on twelve of thirteen datasets with GIN. Analysis shows that AW-RoPE can distinguish routes whose only cue is how two endpoints are connected, while node-wise rotary encodings cannot. --- ### Federated 3D Gaussian Splatting for Large-Scale Scene Reconstruction at Wireless Edge Authors: Guanlin Wu, Chao Hu, Pu Chen, Juyong Zhang, Han Hu, Shuguang Cui, Jie Xu Published: 2026-09-26 Categories: cs.CV, cs.DC arXiv: https://arxiv.org/abs/2609.32177v1 Three-dimensional (3D) Gaussian splatting (3D-GS) has emerged as a promising technique for large-scale scene reconstruction due to its high rendering efficiency and fidelity. However, the training of large-scale 3D-GS models at wireless edge faces various technical challenges including the limited communication, computation, and graphics processing unit (GPU) memory resources at edge devices, the structural inconsistency issue across local models hindering their effective aggregation, as well as privacy leakage risks associated with raw visual content and camera parameters. To address these challenges, this paper proposes a novel resource-efficient federated learning framework for efficiently training 3D-GS models of large scenes under severe resource constraints. First, we propose an on-device model lightweighting mechanism that adaptively selects and prunes Gaussian points to balance the rendering quality and training efficiency. In this mechanism, we quantitatively evaluate the importance of different Gaussian points at each device to facilitate the pruning, and use a novel importance-to-latency ratio criterion to determine the number of pruned Gaussian points under GPU memory and computation/communication latency constraints. Furthermore, we develop a 3D-GS model recovery mechanism that restores structural consistency across local 3D-GS models without accessing private camera parameters, enabling their effective aggregation towards a global model. Finally, extensive experiments show that our approach significantly accelerates convergence, maintains high rendering quality, and reduces training latency compared to state-of-the-art federated 3D-GS baselines. --- ### Efficient Support Recovery of Mixtures of Sparse Linear Classifiers with Less Measurements Authors: Xiaxin Li, Arya Mazumdar Published: 2026-09-26 Categories: cs.LG, cs.DS arXiv: https://arxiv.org/abs/2609.32176v1 The support recovery problem in mixture of linear classifiers intends to identify which features actually matter when data is generated by a mixture of several linear decision rules. In particular, the aim is to recover the support (nonzero coordinates) of $l$ unknown $k$-sparse vectors from sign measurements. Each measurement is generated by selecting one of the $l$ vectors uniformly at random, and returning the sign of its inner product with a chosen measurement vector. In this paper, we propose adaptive and non-adaptive schemes that significantly improve upon prior results by reducing the number of measurements and achieving sublinear decoding time simultaneously. In particular, our adaptive constructions substantially reduce measurements compared to existing approaches, while also lowering decoding complexity from super-quadratic to sublinear in the ambient dimension. We further provide a non-adaptive scheme that improves previous measurement bounds while maintaining efficient decoding. Overall, our approach yields a more efficient trade-off between sample complexity and decoding time for support recovery in mixture models than previously known methods. --- ### OneFixer: High-Quality and Consistent One-Step Autoregressive 3DGS Refinement for Driving Scenes Authors: Boseong Jeon, Junhyeop Lee, Juhan Cha, Hayoung Kim Published: 2026-09-26 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2609.32175v1 Autoregressive video diffusion is a promising render-time fixer for 3D Gaussian Splatting (3DGS) in autonomous-driving simulation, but deployment demands high visual quality and temporal consistency at low latency. This is especially hard for one-step causal generation, where each imperfect prediction immediately becomes context for subsequent frames. Existing approaches stabilize rollouts through staged training with multiple modules and rollout-aware regularization, yet one-step quality still falls short of what deployment requires. We introduce OneFixer, a one-step autoregressive video-diffusion fixer trained in a single task-specific adaptation stage. Our key idea is a deployment-matched shared rollout: the model's own one-step predictions serve as the causal context for flow matching, exposing training to deployment-time errors, while the same rollout receives direct pixel-space perceptual supervision to preserve fine detail. Because the predictions optimized for current-frame quality are exactly those reused as future context, fidelity and autoregressive robustness are learned jointly, without bidirectional-to-causal conversion or teacher-student distillation. OneFixer further exploits cues that driving simulation readily provides, lane geometry and dynamic-agent states, to improve geometric fidelity. On Waymo and proprietary driving scenes with 900-frame rollouts, OneFixer achieves the lowest FVD, LPIPS, and DISTS among all baselines at one step, with temporal consistency matching or exceeding multi-stage DMD pipelines. Under identical backbone and conditioning, it matches a multi-stage DMD-with-Self-Forcing pipeline in under half the GPU-hours and keeps improving beyond its plateau. In closed-loop simulation with a driving policy, OneFixer reduces the collision rate by a third relative to raw 3DGS rendering. Project page: https://onefixer-web.vercel.app/ --- ### Noisy Test-Time Reinforcement Learning for Code LLMs Authors: Xikai Yang, Hieu Trung Nguyen, Dunyuan Xu, Yuzhi Zhao, Jinpeng Li, Wenao Ma, Pheng-Ann Heng Published: 2026-09-26 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.32172v1 Large language models (LLMs) have demonstrated remarkable performance across various code-related tasks. However, unlike carefully curated datasets that are typically high-quality and error-free, real-world user instructions are often vague and error-prone, posing significant challenges to the robustness of code LLMs. Furthermore, robustness-oriented fine-tuning relies on paired clean-noisy samples, which are costly to curate and require sophisticated noisy simulation techniques. To address these challenges, we propose the Noisy Test-time Reinforcement Learning framework (NTRL-Code), which enables robust self-evolution of code LLMs using only unlabeled noisy data during the testing stage. Specifically, NTRL-Code uses conservative self-denoising to obtain a cleaner semantic anchor for target estimation, and employs an abstract-syntax-tree (AST)-based structural aggregation mechanism to estimate a proxy target from multiple candidate programs. The policy is then optimized on the original noisy prompts with a hybrid reward that combines format validity, code similarity, and anti-repetition signals. Extensive experiments on three benchmarks, each incorporating character-level, word-level, and paragraph-level perturbations, demonstrate that NTRL-Code yields robust and consistent improvements, stabilizing the predictions of various base models. Our code is available at https://github.com/Xikai97/NTRL-Code. --- ### Uncertainty-Aware Selection of Online Algorithms with Simulator Ensembles Authors: Yongyi Guo, Zifan Xu, Ziping Xu, Kelly W. Zhang Published: 2026-09-26 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2609.32170v1 The performance of online reinforcement learning depends critically on design choices, especially those that affect exploration. These choices are often selected by fitting a simulator to offline data, evaluating candidate algorithms in that simulator, and deploying the best-performing one. The simplest Plug-In selection rule simply selects the best performing algorithm on the fitted simulator, making evaluations unreliable when the offline data used to fit the simulator are limited. We investigate Uncertainty-Aware selection, which forms an ensemble of simulators---for example, obtained by bootstrap resampling---and selects the online algorithm with the best average performance across the ensemble. While ensemble-based approaches have been used to mitigate distribution shift and facilitate sim-to-real transfer, we formally show that this approach can mitigate the effects of limited data when fitting the simulator and theoretically has significant regret gains compared to Plug-In selection in multi-armed bandits. We also empirically investigate the Uncertainty-Aware selection approach in deep RL experiments on robotic control tasks that involve selecting reward-shaping hyperparameters, and show that it leads to more reliable selection and improved online performance. --- ### Uncertainty Quantification of Next Generation Reservoir Computing with Applications to Memory-Driven Dynamical Systems Authors: Livia Popa, Sumanta Basu, Martin T. Wells Published: 2026-09-26 Categories: stat.ML, cs.LG arXiv: https://arxiv.org/abs/2609.32169v1 Nonlinear dynamical systems with memory arise across science and engineering, yet uncertainty quantification for efficient forecasting methods such as Next Generation Reservoir Computing (NGRC) remains underdeveloped. We study Bayesian ridge and conformal prediction intervals for NGRC and characterize when their uncertainty estimates agree or differ. In low dimensions, their asymptotic widths are governed by different summaries of the residual distribution, so agreement depends on residual shape rather than dimensionality alone. In high dimensions, regularization introduces a further tradeoff between estimation variance, shrinkage bias, and posterior uncertainty, leading to an explicit transition between regimes where Bayesian intervals are wider or narrower than conformal intervals. We extend these results to quadratic NGRC feature maps and give sufficient conditions for transferring the analysis to temporally dependent forecast windows. Simulations and real-data experiments support the theoretical predictions and illustrate how residual distribution, regularization, dimensionality, and distribution shift affect interval calibration and efficiency. These results provide a principled framework for choosing and interpreting uncertainty quantification methods in reservoir-based forecasting. --- ### CAFE: Counterfactual Prediction via Fast Posterior Estimation Authors: Xinyan Han, Xiaoyu Lin, Hao Zou, Xingxuan Zhang, Bo Li, Peng Cui Published: 2026-09-26 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.32167v1 Counterfactual prediction estimates an individual's outcome under an alternative intervention given their factual observations. Such outcomes are generally not identifiable from observational data without additional assumptions. Even within the class of fully observed additive noise models (ANMs), different causal graphs can generate the same observational distribution yet imply different individual counterfactual outcomes. Predictions based on a single estimated graph ignore this structural uncertainty. We therefore target a Bayesian counterfactual posterior predictive distribution that combines predictions from plausible SCMs. We introduce CAFE (\textbf{C}ounterf\textbf{A}ctual Prediction via \textbf{F}ast Posterior \textbf{E}stimation), an amortized inference framework that directly approximates the Bayesian counterfactual posterior predictive distribution. We pretrain a transformer-based model on synthetic counterfactual tasks generated from a diverse prior over ANMs. Given an observational dataset, an individual's factual observations, and an intervention, CAFE approximates the corresponding posterior predictive distribution in a single forward pass. Experiments show that CAFE accurately predicts individual counterfactual outcomes in identifiable settings and approximates the posterior predictive distribution when structural uncertainty induced by observationally indistinguishable causal graphs exists. Strong performance in realistic manufacturing and viticulture settings further demonstrates its empirical robustness beyond the assumptions of the training prior. --- ### PastForward: Faster On-Device GUI Agents via Computational Experience Reuse Authors: Taehwan Park, Changmin Lee, Hayeon Lee, Taesik Gong Published: 2026-09-26 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.32166v1 Running GUI agents on edge devices can keep sensitive screens and interaction histories local, but the computational cost of inference at every action step makes deployment challenging. Existing GUI agent systems either perform full vision-language model (VLM) inference at each action step or reuse coarse-grained knowledge matched to prior tasks. However, dynamic mobile environments and user tasks make it difficult to fully utilize prior task executions without additional fine-tuning or task-specific offline exploration. To address this challenge, we present PastForward, a system that accelerates GUI agents through validated, fine-grained reuse of computational experience accumulated during ordinary task execution. During decoding, PastForward retrieves prior output sequences as device-adaptive multi-token proposals and verifies them in a single VLM forward pass. Across action steps, it uses prior GUI transitions to begin next-step inference while the device executes the current action, retains the early computation only when the predicted screen matches the observed screen, and carries reusable KV states forward. We evaluate PastForward on AndroidWorld workloads derived from real mobile usage patterns using multiple VLM backbones across server and edge platforms. On device, PastForward achieves action-step latency speedups of 1.63-2.36$\times$ while maintaining task success rates. --- ### FOCUS: Fixed-Confidence Online Causal Learning Using Sequential Adaptive Interventions Authors: Haijie Xu, Chen Zhang Published: 2026-09-26 Categories: stat.ML, cs.LG arXiv: https://arxiv.org/abs/2609.32165v1 We study fully online fixed-confidence causal discovery without any historical observational data. Starting from zero samples, the learner sequentially selects interventions to recover both the causal DAG and its edge weights under a linear-Gaussian structural equation model. We establish an instance-dependent lower bound for any $(ε,δ)$-correct algorithm and propose \textsc{FOCUS}, which adaptively allocates interventions through an online max--min game. A key contribution is a computable concentration inequality for the accumulated KL divergence involving causal parameters shared across interventions. We prove that \textsc{FOCUS} is $(ε,δ)$-correct and that its expected stopping time matches the lower bound in its $Θ(\log(1/δ))$ dependence up to an instance-dependent constant. Experiments demonstrate improved structure and edge-weight recovery and confirm the predicted stopping-time trend. Our codes are available on https://anonymous.4open.science/r/FOCUS_code-76E5 --- ### PQR3D: Progressive Query Refinement over Reference-Conditioned Temporal Windows for Multi-View 3D Object Detection Authors: Hui Ye, Yudong Liu, Yiran Chen, Rajshekhar Sunderraman, Shihao Ji Published: 2026-09-26 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.32163v1 Temporal context is essential for camera-only multi-view 3D object detection. Existing streaming detectors maintain and propagate query states from one frame to the next, requiring sequence-aware training and chronological inference. We propose PQR3D, which performs progressive query refinement within referenceconditioned temporal windows. This design enables random frame sampling and independent inference without persistent query memory. Within each window, PQR3D progressively transfers motion-aligned high-confidence queries from earlier timestamps toward the target frame. We further introduce masked selfattention to regulate interactions among regular, propagated, and denoising queries while keeping denoising supervision isolated from detection queries. In addition, a stage-decoupled anchor embedding injects position before self-attention and size, orientation, and velocity afterward, reducing interference from temporally inconsistent attributes. With a ViT-L backbone, PQR3D sets a new state of the art on the nuScenes test set, achieving 71.6 NDS and 64.9 mAP. Source code is available at https://github.com/huiyegit/PQR3D --- ### PruneForget: Joint Unlearning and Pruning of Vision Models Authors: Yu-Shan Tai, Amber Yijia Zheng, Raymond A. Yeh Published: 2026-09-26 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.32162v1 Machine unlearning and model pruning are increasingly coupled in the real world. Models must support unlearning requests, e.g., for safety concerns, while also meeting requirements in latency and memory budget. Until recently, existing works have studied each aspect as an independent problem, e.g., running unlearning and pruning sequentially. In this work, we show that unlearning and pruning are naturally aligned and should be solved jointly to be made aware of each other. Intuitively, parameters that encode information of the unlearned samples are natural pruning targets, as unlearning and pruning both call for the "deletion" of such parameters. We propose PruneForget, a method that uses the unlearn set as a guide for pruning, so that unlearning and pruning mutually benefit each other. Extensive experiments on image classifiers and generative models show that PruneForget removes the influence of the unlearned samples while producing a more compact model with reduced inference cost. It achieves a negligible performance gap relative to an oracle that retrains from scratch for unlearning and then prunes. --- ### Typed Decision Models: An Early Evidence Audit and Evaluation Checklist Authors: Lijuan Tang, Yuemeng Zheng Published: 2026-09-26 Categories: cs.CL, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2609.32160v1 Typed decision models (TDMs) return probability distributions over caller-defined options without generating text. TypeSafe released Jev, a commercial typed decision model, on 15 September 2026, and a small body of evaluation and replication work appeared within days. We review 28 papers posted between 19 and 24 September and relate their findings to earlier work on label-probability classification, constrained decoding, reranking, calibration, and model cascades. In this early literature, the typed readout itself has not shown an independent accuracy advantage over comparable label-probability readouts. Jev's clearest gains are in latency and cost, while accuracy gaps remain on harder tasks. In practical deployments, confidence is often used to decide when to defer to a stronger model or a human. We use recurring weaknesses in these studies to derive a 14-item evaluation checklist for future TDM work. Because the evidence covers only the first nine days after the release of one hosted model, the review should be read as an early evidence map rather than a settled assessment of the model class. --- ### CausalDriveBench: Evaluating Causal Reasoning in Vision-Language-Action Models for Autonomous Driving Authors: Narendiran Chembu, Navvrat Rao, Shreedhar Shreeshail Kodate, Gayatri Srujana Banda, Arko Sarkar, Abhinav Khanna, Rajarshee Das, Umesh Kanala, Siddarth Khandelwal, Kumar Aman, Aish Dubey, Kaustubh Beedkar et al. Published: 2026-09-26 Categories: cs.CV, cs.RO arXiv: https://arxiv.org/abs/2609.32157v1 Vision-Language-Action (VLA) models for autonomous driving produce natural-language reasoning alongside predicted trajectories, but whether this reasoning reflects the causal structure of the scene remains untested. We introduce CausalDriveBench, an evaluation framework grounded in Pearl's Causal Hierarchy (PCH) that tests causal reasoning in driving-specific VLAs through structured visual question answering (QA) and alternative-trajectory prediction. To this end, we construct causal scene graphs over nuScenes that distinguish causally active, dormant, and distractor entities, separating perceptual salience from causal relevance. The benchmark spans all four rungs of PCH (association, intervention, and counterfactual along with causal discovery) for QA generation. For the higher rungs, we additionally provide reference trajectories under specified scene modifications, enabling action-level verification that complements reasoning-level evaluation. In total, the benchmark contains 7,285 verified causal QA pairs and 1,000 counterfactual trajectories derived from nuScenes. We evaluate 10 driving-specific VLAs and 3 general-purpose VLMs, and report three findings. First, the best model reaches only 70.6% QA accuracy, and 4 of 13 models score below random chance. Second, comparing each driving VLA to the general-purpose VLM that shares its language backbone, the cost of driving fine-tuning ranges from 2 to 34 percentage points on causal QA, with post-training design explaining the spread. Third, causal QA and trajectory accuracy are statistically uncorrelated across models: under counterfactual prompts, predicted trajectories either over-react or collapse onto the observed-scene baseline. Taken together, these results show that neither fluent rationales nor accurate observed-scene trajectories constitute evidence of causal understanding. --- ### AquaBEV-Nav: Learned BEV Occupancy for Underwater Navigation and Exploration Authors: Trung Tien Dong, Zhenqi Wu, Sahasra Kondapalli, Jiayi Wu, Yi Sheng, Xiaomin Lin Published: 2026-09-26 Categories: cs.RO, cs.CV arXiv: https://arxiv.org/abs/2609.32156v1 Safe underwater exploration requires a robot to understand where surrounding structures are located and which regions are available for motion. Existing vision-based underwater exploration systems commonly obtain this information indirectly by estimating monocular depth, unprojecting the geometry into 3D space, and accumulating it into a 2D bird's-eye-view occupancy map. This reliance on intermediate depth estimation is particularly problematic underwater, where scattering and wavelength-dependent attenuation degrade visual cues and limit the reliability of monocular depth estimates. We introduce AquaBEV-Nav, an underwater exploration framework that bypasses explicit monocular depth estimation through direct bird's-eye-view occupancy prediction. Built upon the CORAL hierarchical exploration framework, AquaBEV-Nav replaces its depth-based perception front end with AquaBEV. Given a single RGB frame, AquaBEV maps visual features into a learned polar representation, performs causal reasoning along the range dimension, and reconstructs local Cartesian occupancy without relying on intermediate depth prediction. The resulting occupancy map is accumulated into CORAL's persistent spatial memory, providing spatial context for VLM-based high-level planning and collision constraints for dynamics-aware local trajectory generation. Across ten simulated reef environments and six occupancy backbones evaluated under a single protocol, AquaBEV-Nav reaches 37.48 structure IoU and 53.2 target IoU, 88.95% closed-loop coverage with zero collisions. --- ### A Unified Optimism-Agnostic Framework for Linear Bandits over Spherical Action Sets Authors: Arda Güçlü, Subhonmesh Bose, John R. Birge Published: 2026-09-26 Categories: stat.ML, cs.LG arXiv: https://arxiv.org/abs/2609.32149v1 Linear bandits model sequential decision-making problems with noisy rewards that are linear in the decision variable, where an agent must simultaneously learn about an unknown parameter that governs the mean rewards, while maximizing (expected) rewards over time. Two prominent algorithmic families--upper confidence bound (UCB) and Thompson sampling (TS)--achieve a balance of exploration (to estimate said parameter) and exploitation (utilization of knowledge about it) across time. The quality of estimation of that parameter depends on the eigenvalues of a design matrix. In this paper, we begin by showing that if the inference quality obtained from exploration, encoded in the minimum eigenvalue of the design matrix, grows $\gtrsim \sqrt{t}$ with time $t$, while actions remain sufficiently concentrated for exploitation, then an algorithm produces optimal high-probability $\mathcal{O}(\sqrt{T}\log T)$-regret rate over a time-horizon $T$ for spherical action sets. This analysis is algorithm-agnostic and follows an alternative route to the classical optimism-based elliptical-potential argument for regret analysis. Then, we illustrate that variants of UCB and TS satisfy the inference and concentration properties and in turn, enjoy optimal regret rate. In effect, our results provide a modular framework that can be used to analyze linear bandit algorithms and explicitly connect quality of parameter estimation to optimal regret accumulation. --- ### Playing to Par: Reinforcement Learning for Provably Optimal Quadrilateral Block Decompositions Authors: Arjun Narayanan, Per-Olof Persson Published: 2026-09-26 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.32146v1 A quadrilateral block decomposition of a planar domain is judged by whether it is complete, whether its elements are well shaped, and how many of its vertices are irregular. The last has a provable floor: the discrete Gauss-Bonnet identity enforces a lower bound on the total vertex irregularity of any all-quadrilateral mesh of a given domain purely based on its topology and corner angles. We train a reinforcement learning agent to build decompositions that reach this bound, which we call par. It acts directly on the mesh's half-edge data structure through local edits, with a policy network whose convolutions follow the mesh's own connectivity, so it applies unchanged to domains larger than any seen in training. The reward targets the floor directly, and it is sparse: random play reaches it on no domain with more than eight sides. We overcome this exploration barrier via behaviour cloning on optimal meshes that are trivial to construct, walked backward into demonstrations, before training it with PPO. On 96 held-out domains the agent produces an all-quadrilateral mesh on every one, a usable one on 95.7 on average, and a provably optimal one on 90; Gmsh's strongest configuration at the same element count completes 51, is usable on 38 and optimal on none, and even at three to fourteen times the elements never produces a more regular mesh. On 64 domains twice the training size the agent completes all, is usable on 62, and keeps a median excess over par below one against Gmsh's 39 at the same element count. --- ### Spectral Reversal: Counteracting Singular Value Bias for Graph Prompting Authors: Hanxu Yang, Yuhuan Zhao, Xiaodong He, Zhao Kang Published: 2026-09-26 Categories: cs.LG, cs.AI, cs.SI arXiv: https://arxiv.org/abs/2609.32143v1 Pre-training Graph Neural Networks (GNNs) via self-supervised learning has become a dominant paradigm, yet efficiently adapting frozen encoders remains a challenge. Graph prompting offers a parameter-efficient alternative to fine-tuning, but existing methods largely treat pre-trained models as opaque feature extractors, ignoring their internal spectral structure. In this work, we identify a systematic phenomenon in pre-trained GNNs, which we term spectral bias: optimization during pre-training disproportionately aligns representations with directions associated with large singular values, leaving low-energy directions under-explored. We show that these underutilized directions can encode complementary information that is beneficial for downstream adaptation, especially under distribution shift. To leverage this insight, we propose Spectral Reverse Prompt (SRP), a prompting framework that rebalances the spectral contributions of frozen GNN encoders. SRP applies a learnable soft-thresholding mask in the spectral domain to down-weight dominant directions while amplifying weaker ones. In addition, SRP incorporates a null-space augmentation module that captures variation in directions with minimal activation under the frozen encoder. Extensive experiments across multiple benchmarks demonstrate that SRP achieves state-of-the-art performance with minimal additional parameters, highlighting that reweighting spectral components is a principled and effective strategy for parameter-efficient graph adaptation. --- ### REALM: Regime-Switching, Explainable, and Activation-Induced Linear Models Authors: Xiaoran Cheng, Sen Na, Jia Li Published: 2026-09-26 Categories: cs.LG, cs.AI, stat.ME, stat.ML arXiv: https://arxiv.org/abs/2609.32141v1 Deep ReLU networks are piecewise-affine mappings that partition the input space into cells, each characterized by a distinct activation pattern. This structure motivates fitting a local linear model within each cell to preserve predictive accuracy while improving interpretability. The challenge is to identify regimes that are stable, data-adaptive, and easy to explain. We propose REALM, a mixture of linear models whose regimes are induced by neural activation patterns. Because the number of activation cells in a deep neural network (DNN) can grow rapidly with depth, we first distill a deep teacher into a wide, shallow student network (WSSN), then binarize and cluster its hidden-layer activations to define the regimes and fit a linear model within each regime. Since the regimes are discovered from internal structure, the router does not carry the predictive burden. To make regime assignment interpretable, we train a multiclass logistic regression, the explanatory gate, to reproduce the regime assignments. The two-level structure is interpretable at both stages in terms of raw tabular or learned convolutional features: the gate identifies features that determine regime assignments, while the linear models identify features that drive predictions within each regime. We analyze an idealized setting that illustrates a trade-off between partition complexity and stability: as the number of regimes grows, finer partitions can improve approximation but may reduce regime-assignment stability. Experiments on tabular and image datasets show that REALM achieves competitive predictive performance relative to other DNN-guided mixture surrogates and inherently interpretable models while producing stable regime-level explanations. --- ### Checking Leakage Witnesses versus Certifying Bounded Non-Leakage Authors: Chao Feng, Burkhard Stiller Published: 2026-09-26 Categories: cs.CR, cs.CL arXiv: https://arxiv.org/abs/2609.32134v1 When a language-model audit finds no leak, what is needed to certify non-leakage? We study guarantees over a declared prompt domain under an executable leakage criterion and decoding rule. For general bounded polynomial-time evaluators, a supplied leaking execution is polynomial-time checkable, while leak existence is \NP-complete and deterministic certification is \coNP-complete. Exact stochastic certification is $\coNP^{\PP}$-complete at every fixed rational cutoff in $(0,1)$. Restricting the computation can change these bounds. For example, certification is in \coNP\ when all randomness is a terminal draw from an efficiently computed finite probability table. Attention models admit polynomial-time certification when local dependency windows of logarithmic length precede one global head, given deterministic decoding, fixed vocabulary, exact rational weighted means, a direct binary affine readout and finite-automaton prompt domains. A construction with two global layers instead makes certification \coNP-complete over template domains, with one head per layer, polynomial width, logarithmic precision and an inverse-polynomial logit margin. Planted-secret experiments measure what finite audits miss relative to complete references. Among 30 secret--model-state pairs that leak under greedy single-prompt execution on their secret's 4,096-prompt domain, uniformly selecting 256 recorded evaluations per pair misses every leak for an expected $41.06\%$ of these pairs. Batched and single-prompt checks disagree on one complete-domain decision among all 48 fine-tuned pairs, while a same-order repeat reproduces every single-prompt output. These results distinguish computational conditions for certification from the coverage and execution conditions needed to interpret a negative audit. --- ### Latency-Aware Client Assignment for Parallel Split Learning With Global Sampling Authors: Mohammad Kohankhaki, Valentin Rentschler, Anke Schmeink Published: 2026-09-26 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.32132v1 In cross-silo split learning, Parallel Split Learning with Global Sampling forms representative pooled batches when class distributions differ across clients, but ignores client delay when several clients can supply the same class. We introduce Latency Budgeted Parallel Split Learning with Global Sampling, which separates each pooled batch's integer class target from the choice of clients that supply its examples. The flow variant formulates this assignment as an integral network-flow problem and minimizes modeled client-side completion time for the current target. The fast variant uses a greedy next-completion rule to reduce schedule-construction cost. Both preserve the target stream and use every local example once per epoch. A planning rule selects between the variants while accounting for the cost of constructing both candidate schedules. On CIFAR-10, the flow variant reduces modeled training time by 6.75%, with a 0.30 percentage-point decrease in final accuracy. On Tiny ImageNet with 20 candidate classes per client, the fast variant reduces modeled time by 16.87% and reaches all four validation targets earlier than the latency-unaware baseline. Across 405 schedule comparisons, the planning rule stays within 2% of the lower realized cost in 96.54% of cases. In our evaluation, latency-aware provider assignment reduces modeled training time without changing the prescribed class targets, while the preferred variant depends on whether assignment savings outweigh schedule-construction overhead. --- ### Gaussian Image Steganography via Parameter-Domain Keyed Embeddings Authors: Tong Wu, Runze Cheng, Xiaoyue Fan, Kaan Akşit Published: 2026-09-26 Categories: cs.CV, cs.GR arXiv: https://arxiv.org/abs/2609.32131v1 2D Gaussian-based image representation is becoming increasingly popular, and our work proposes a new approach to steganography by embedding information within Gaussian parameters rather than image pixels. We first fit the parameters of this representation to the target image and employ a secret key to select a subset of these parameters for fine-tuning, allowing us to embed an 8-bit message while maintaining high visual fidelity in the reconstructed image. Thirty fitting experiments on three synthetic images show that the correct key can recover the message without error, while decoding with incorrect keys yields a Bit Error Rate (BER) of $0.543$, close to random guessing. Compared with random selection using the key, selecting the least-disturbing edits recovers the message more reliably (one-sided $p=0.031$), and the average PSNR cost is only $0.091$ dB in visual quality. The embedding method transfers to 112 natural images at $256\times256$ using 4,096 Gaussians. The correct-key BER is $0.000$, and decoding under a wrong key stays close to random guessing at $0.520$. Our method embeds the payload through three Gaussian parameter types: log-anisotropy, opacity, and color luminance. In a separate nine-fit reduced setting, color luminance is removed, so the payload uses two instead of three parameter types, a $33.3\%$ reduction; all 256 Gaussians remain in the fitted representation. The correct key still recovers the message without error. However, wrong-key BER rises from $0.514$ to $0.571$, moving farther from random guessing ($0.5$). --- ### What Should We Freeze? Guarded Freezing: Connectivity Shapes the Fine-Tuning of Pretrained Models Authors: Leonel Aguilar Published: 2026-09-26 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.32124v1 When adapting pre-trained models through fine-tuning, freezing weights alone might not preserve performance, as updates elsewhere can change the inputs to the frozen core, ultimately affecting overall performance. We first analyse the case where a selected frozen core can be isolated and propose removal-value, a capacity-based score that approximates HOPE's removal cost averaged over removal orders. We show that in VGG-8, cutting paths from trainable neurons into a frozen core makes selection using this score useful: $70\%$ frozen preserves $5.22\pm0.51$ percentage points more old-task accuracy than DEFT at similar new-task accuracy. In transformers, shared residual streams leave paths into frozen neurons open. For this case, we derive drift-value, a forward-only proxy for the output disturbance from updating each weight entry under a local update model. In language models, at 40 epochs, this policy exceeds adapted Wanda and RIA freezing scores in settings with substantial retention loss, while its differences from Fisher remain unresolved. After 160 epochs on Qwen2.5-1.5B, it retains $0.0433\pm0.0102$ more than static Fisher. In DINOv3 vision-transformer adaptation to point clouds, drift-value retains $0.440$ image accuracy versus $0.187$ for a random mask of the same count. These results motivate Guarded Freezing: select by removal-value when incoming paths are cut, and by drift-value when they remain. --- ### READ-Bench: Benchmarking Historical Instance Retrieval for Time-Series Diagnosis Authors: Gerardo Pastrana, Haojun Li, Dhruv Mehta, Anoushka Vyas, Sina Khoshfetrat Pakazad, Henrik Ohlsson, John Paparrizos Published: 2026-09-26 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.32123v1 Time-series diagnostic systems rarely rely on retrieving relevant historical cases, and when they do, retrieval is evaluated only indirectly through downstream prediction. We introduce READ-Bench, a benchmark for historical-case retrieval across 12 diagnostic datasets, centered on multivariate time series, that defines relevance by shared fault or event type rather than signal shape, so visually different traces of the same fault count as relevant while similar-looking traces of different faults do not. Treating retrieval as a base retriever followed by a reranker, we evaluate classical distances, symbolic retrievers, self-supervised and foundation-model embedders, and their fusion, plus label-aware and language-model rerankers, under one protocol that varies supervision, pollution, and corpus scale with significance testing. Under a common channel-independent interface, pretrained representations offer no statistically detectable advantage over strong classical and symbolic baselines for search alone. The decisive factor is a small amount of resolved-case supervision at reranking, namely a Gaussian-process reranker that propagates a few neighbor labels in embedding space, which helps far more than more sophisticated representations or language-model reasoning and holds under pollution and at full corpus scale. Guided by these findings, we fuse a normal-residual-scored embedder with a dynamic time warping leg via reciprocal-rank fusion, then rerank with the Gaussian-process reranker, improving NDCG@10 over its own search stage on all 12 datasets, by +0.11 from reranking and +0.16 over the strongest single base retriever. --- ### Using LMs to Model the Effects of Context and Coreference during Sentence Comprehension Authors: Kohei Kajikawa, Lin Ai, Tatsuki Kuribayashi, Ethan Gotlieb Wilcox Published: 2026-09-26 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.32119v1 Language models (LMs) are often used as a tool to model human language processing. Recent studies suggest that severely restricting LMs' context window improves their fit to human psycholinguistic data by simulating human working memory constraints. However, it is possible that this strict memory-decay approach overlooks humans' reliance on long-range structural representations, such as discourse structre. In this work, we systematically vary the context window size of GPT-2 across four large-scale naturalistic English reading-time datasets and observe a U-shaped relationship: Although restricted contexts (< 20 tokens) successfully capture local memory limitations, expanded contexts (500--1,000 tokens) ultimately yield the highest overall psycholinguistic fit. To investigate the mechanism driving this benefit, we conduct a counterfactual inference-time experiment that disrupts cross-sentential entity chains by pronominalizing repeated discourse entities. Obscuring these structural linkages significantly degrades the predictive power of larger context windows by 20% to 40%. Our experiments demonstrate that tracking long-range coreference relations is one important factor for the alignment between LM surprisal and human reading behavior, and approximate the extent to which human comprehenders use global discourse relations during language processing. --- ### Handwritten Digit Leakage from Smartphone Motion Sensors Across Unseen Users and Phone Models Authors: Constantino Álvarez Casado, Erkka Rantahalvari, Matteo Pedone, Matti Matilainen, Manuel Lage Cañellas, Le Nguyen, Simo Hosio, Olli Silvén, Miguel Bordallo López Published: 2026-09-26 Categories: cs.LG, cs.HC arXiv: https://arxiv.org/abs/2609.32117v1 Smartphone motion sensors support interactive applications, but their readings may also reveal touchscreen input beyond their intended use. Assuming known drawing intervals, we study whether handwritten digits remain predictable across users and devices, as a 10-class problem on 19,628 HuMIdb recordings from 481 participants. We compare handcrafted features with classical machine learning algorithms, MiniRocket kernels, and a compact sensor patch transformer on accelerometer, linear acceleration, gyroscope, and gravity signals. The transformer achieves 57.74\% accuracy and 82.64\% top-3 accuracy on 75 unseen participants, and 58.77$\pm$0.95\% over 3 seeds for unseen participants on 9 unseen phone models. Low motion recordings remain informative, accuracy is not monotonic in motion level, and the tested contrastive pretraining, augmentation, and derived signals give no consistent gains. Digits are thus predictable beyond familiar users and phone models under assumed segmentation, while acquisition-order shortcuts limit conclusions about practical privacy exposure. Code available at: https://github.com/Arritmic/motion-digit-leakage. --- ### Escaping Alignment: A Physical Trap Model of Best-of-N Jailbreaking Authors: Marco Biroli Published: 2026-09-26 Categories: cs.AI, cond-mat.stat-mech arXiv: https://arxiv.org/abs/2609.32116v1 Best-of-$N$ jailbreaking (BoN) bypasses safeguards of aligned models by drawing $N$ independent augmentations of an unsafe prompt and sampling $M$ completions of each. Previous works have shown that the attack success rate (ASR) seems to follow a power-law in $N$, which we challenge. The exponent drifts with $N$, with an exponential crossover which is a finite-size artifact of the adversarial dataset. Little work has been done to explore the entire two-budget ($N, M$) attack surface as well as its dependence on the generation temperature $T$. We introduce a simple barrier model where each prompt has a baseline safety level and each augmentation a random thermally activated barrier. Then four numbers, each backed by an interpretable safety mechanism, determine the entire ($N, M$) attack surface. They extrapolate predictions from $N \leq 100$ to $N = 10^4$, collapse five distinct models on the same scaling function and predict ASR at different temperatures from the one they were fitted at. --- ### Empowering Hybrid Attention Models on NPUs Authors: Yinyuan Zhang, Daliang Xu, Xiaolong Huang, Wangsong Yin, Yun Ma, Mengwei Xu, Gang Huang Published: 2026-09-26 Categories: cs.DC, cs.AI arXiv: https://arxiv.org/abs/2609.32114v1 Hybrid attention models have emerged as a crucial architecture for Large Language Models (LLMs) (e.g., the Qwen3.5 and Kimi series). Their memory and computational efficiency make them highly attractive for on-device inference, forming a promising synergy with edge Neural Processing Units (NPUs). However, naive execution of these hybrid models on edge NPUs fails to deliver these benefits, often bottlenecking the prefill stage due to severe memory-system inefficiencies and architectural mismatches within the linear attention (LA) layers. We present HA-NPU, the first system to enable efficient hybrid attention LLM inference on edge NPUs without modifying the underlying algorithms. HA-NPU enhances execution efficiency by reorganizing the dataflow of the LA components across three levels: (1) At the core level, it partitions workloads by the head dimension and fuses dependent operators, eliminating cross-core global memory accesses; (2) At the operator level, it reorders execution to consume intermediate tensors immediately, drastically minimizing local-buffer pressure; (3) At the tensor level, it employs dataflow-aware layout planning to minimize transformation overhead between matrix and vector processing units. Compared to competitive baselines, HA-NPU achieves up to 35.95$\times$ LA kernel speedup and 36.14$\times$ energy reduction, delivering up to 2.03$\times$ faster end-to-end request latency. The source code will be made publicly available at https://github.com/yinyuanzhang/HA-NPU --- ### How Reusable Are Benchmarks with Richer Feedback? Authors: Youssef Allouah, John Duchi Published: 2026-09-26 Categories: cs.LG, stat.ML arXiv: https://arxiv.org/abs/2609.32109v1 We study whether benchmarks reliably guide model selection as developers adapt to evaluation feedback across multiple criteria. We find that the worst-case test-set size needed to estimate the best score among $k$ adaptively chosen models, under any convex combination of the criteria, grows exponentially with the number of criteria, reaching the $Θ(\sqrt{k})$ cost of answering $k$ adaptive statistical queries with only $O(\log k)$ criteria, at fixed accuracy and confidence. In attacks on multi-task large language model benchmarks with five to ten criteria, feedback restricted to nondominated task profiles produces large reused-to-held-out score gaps and frequent false winners. These results challenge a prominent explanation for prior observed reliable benchmark reuse---that developers mainly respond to convincing improvements over the current best---in rich-feedback settings, while leaving open how often ordinary model development encounters this vulnerability. --- ### SAMBAR: Selective Anchoring via Method of Multipliers for Balanced Knowledge Acquisition and Retention in Vision-Language-Action Models Authors: Aayushi Shrivastava, Xunlan Zhou, Hongrui Zhao, Ziyu Chen, Negar Mehr Published: 2026-09-26 Categories: cs.LG, cs.RO arXiv: https://arxiv.org/abs/2609.32108v1 Vision-Language-Action (VLA) models leverage large-scale pretraining to ultimately achieve generalist manipulation. Deployed VLA policies must support continual learning to acquire new tasks over time. Teaching a VLA a new task generally requires finetuning it on demonstrations of that task. However, naively finetuning on downstream tasks causes the policy to forget earlier tasks and degrades generalist capabilities. This failure is known as catastrophic forgetting. Most continual learning methods counter it by replaying data from earlier tasks. However, the old task demonstrations are not always readily available. In this paper, we introduce SAMBAR, a continual learning algorithm that prevents catastrophic forgetting during VLA finetuning without requiring access to the demonstrations of any previously learned task. We propose to cast continual learning as a constrained optimization problem and solve it with the method of multipliers. In our approach, the method of multipliers drives the policy to learn the new task without the model parameters drifting far away from their previous values. In contrast to a standard regularization penalty, the method of multipliers raises the penalty as the constraint violation accumulates by using a dual variable. We also selectively anchor the parameters critical to previous tasks to preserve past knowledge, leaving other parameters free for new task acquisition. The combination of dual variable and selective anchoring, therefore, balances knowledge acquisition with knowledge retention. We evaluate our method, SAMBAR, on the LIBERO simulation benchmark and on hardware. When sequentially finetuning on a VLA, every replay-free baseline we compare against completely forgets the first task it learned, whereas SAMBAR retains every task it has learned. --- ### An Attention-Driven Heterogeneous GNN Model for Credit Card Fraud Detection Authors: Kathiresan Jayabalan, Sethuraman Radhakrishnan Published: 2026-09-26 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.32106v1 The global transition to a cashless economy has placed credit cards as the key element of digital transactions, acclaimed for their easy use, speed, and acceptance in most places. However, the growing dependence on this payment method has led to an escalation of the risks associated with credit card (CC) fraud. Detecting this type of fraud is a difficult task because the patterns are constantly changing, there is a data imbalance, and it is necessary to identify the legitimate transactions and the fraud ones at the same time. This study addresses this challenge by proposing a credit card fraud detection (CCFD) framework using a data balancing technique and a deep learning (DL) model. The proposed fraud detection model is trained and evaluated by collecting the dataset called Credit Card Fraud Detection from the Kaggle repository. As the dataset is highly imbalanced, we utilized the Synthetic Minority Oversampling Technique (SMOTE)-Tomek technique to balance the dataset. Further, the balanced dataset is classified using the Heterogeneous Graph Neural Network (HGNN) model. The HGNN model represent various transactions using a heterogeneous graph architecture and by using an attention based message passing technique, it managed to consider the complex relationships, time factors, and user behavior. The integration of SMOTE-Tomek in the model further boosted its capacity to identify fraudulent transactions, while lowering the rate of false positives. The HGNN model attained a 99.97% accuracy, a 99.48% F1-score, a 99.15% precision, and a 98.97% recall. The findings indicates that this model is effective and can be applied to real-world CCFD scenarios. --- ### LLM Unlearning Evaluation with TRIAGE Authors: Danial Ataee, Peter Triantafillou Published: 2026-09-26 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.32103v1 Large language models can memorize private or harmful information, motivating machine unlearning methods that remove targeted knowledge while preserving other capabilities. However, existing evaluations rely primarily on behavioral benchmarks, which assess \emph{whether} a model appears to forget but provide limited insight into \emph{how} unlearning changes the model or affects related knowledge. We introduce \textit{TRIAGE} (\textit{Tripartite Representation-internal Introspection for Adjacency Gap Evaluation}), a benchmark-agnostic evaluation framework for characterizing these changes. TRIAGE uses diagonal approximations of the Fisher information and Hessian to measure changes in parameter sensitivity and local curvature, and utilizes a Forget / \emph{Adjacent-Retain} / \emph{Generic-Retain} partition to quantify an \emph{adjacency gap} in semantically related knowledge. Based on the magnitude and distribution of these changes, TRIAGE further classifies each algorithm's update as \emph{no-op}, \emph{partially localized}, \emph{collateral dominant}, or \emph{globally destructive}. Across 12 unlearning methods, four language models, and the WMDP, TOFU, and MUSE benchmarks, we find that methods with similar behavioral forgetting can produce substantially different internal changes and patterns of collateral damage. These signatures also vary across models and benchmarks, indicating that the effects of unlearning are not determined solely by the unlearning algorithm. TRIAGE can be applied alongside existing unlearning benchmarks to complement behavioral evaluation with a model-internal view of how unlearning reshapes the model's parameter space and affects retained knowledge. --- ### Residual Streams Read, Recurrent States Remember: The Global Workspace in Mamba Models Authors: Wenlong Wang, Fergal Reid Published: 2026-09-26 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.32102v1 Can the global-workspace account of transformer representations extend to state-space language models? We fit Jacobian lenses to the residual streams and recurrent states of Mamba-1, Mamba-2 and Mamba-3, using the original 1000-prompt recipe. Joint residual--state readouts improve recovery of known intermediate concepts over the residual lens on at least five of six task families in every tested Mamba checkpoint. On Mamba-2, state alone exceeds residual and logit lenses on all six families; a normalised joint readout improves on both components on five. Temporal maps and word-list experiments show earlier content remaining state-readable as residual visibility changes. We also propose sign-guarded steering, which improves target top-five success over coordinate exchange on matched verbal-report trials in five models. Recurrent state alone supports this verbal access. These gains do not extend consistently to relational answers: guarded edits often output the edited concept itself, and Mamba-3's joint edits can disrupt successful state-only redirection. Recurrent state thus provides a complementary carrier of workspace content, whose recovery, persistence and causal uses require separate measurements. --- ### Emergent One-Third Scaling Law as Attention Tries to Concentrate Authors: Yizhou Liu, Sara Kangaslahti, Jeff Gore Published: 2026-09-26 Categories: cs.LG, cs.AI, stat.ML arXiv: https://arxiv.org/abs/2609.32100v1 The neural scaling law relating longer training to better performance through a power law is central to today's large language models (LLMs), yet its origin remains debated. One recent proposal is that power laws can emerge from the strong non-linearity of a single softmax head learning peaked distributions. What happens with multiple softmax functions, as in LLMs, is unclear. Here, we show through toy models that any softmax learning peaked distributions, regardless of its position in the model, can develop logit magnitudes that grow in a power law with exponent $1/3$, becoming a training bottleneck whose loss contribution decays as a power law with the same exponent $1/3$. The overall loss therefore obeys $1/3$ scaling whenever at least one softmax learns peaked distributions. We confirm that many softmax functions in LLMs learn peaked distributions and that LLM loss scaling matches this $1/3$ prediction. Moreover, logit growth dynamics reveal that attention heads, rather than the language modeling head, are the bottleneck likely driving the $1/3$ loss scaling in LLMs. Attention trying to concentrate on specific information, which is the heart of Transformers, may therefore also be the heart of the neural scaling law of training. --- ### Regime Boundary Alignment for Evidence-Gated Question Answering Authors: Zeyan Li, Qirong Guo, SIyuan Qiu, Hu Xu, Chun Li, Jianfeng Xu Published: 2026-09-25 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2609.37491v1 Retrieval-augmented language models are expected to answer from the retrieved evidence, but in practice they often keep answering when that evidence is missing. We trace this behavior to the training signal: answer-focused fine-tuning assigns no target to unsupported contexts, so it cannot distinguish a reader that abstains from one that guesses, and unsupported answering stays near 100% even as supported accuracy improves. We introduce Regime Boundary Alignment (RBA), which trains a single reader on matched variants of the same question and gold answer. The reader is trained to produce the gold answer when the context supports it, including when conflicting evidence is also present, and to abstain when the correct support is removed; inference is ordinary decoding, with no verifier, threshold, or regime label. On three multi-hop QA datasets across three seeds, RBA reduces the unsupported-answer rate by more than sixty percentage points relative to conflict-focused training while matching its supported accuracy. On a held-out TriviaQA retrieval-miss slice, the same reader reduces unsupported answering from 100% to below 1% while also improving supported accuracy. These results indicate that evidence-gated answering must be learned on both sides of the support boundary. --- ### Label Less, Learn More: Resource-Efficient Active Semi-Supervised Learning for Onboard Satellite Image Annotation Authors: Ahmed Abdelnaby, Mohamed Elmahallawy, Marius Bernahrndt, Tobias Hecking Published: 2026-09-25 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.37481v1 Large-scale pervasive sensing increasingly relies on high-resolution satellite imagery, yet task-specific onboard vision is constrained by costly annotation and limited computation, memory, energy, and communication resources. Existing approaches largely rely on either data-hungry supervised learning or large vision-language foundation models, limiting efficient adaptation and deployment under these constraints. We present SatLabel, a resource-aware learning framework that transforms limited satellite labels into progressively refined onboard models through adaptive sample acquisition and semi-supervised model adaptation. Rather than repeatedly training on uniformly sampled labels, SatLabel closes the loop between model uncertainty, class imbalance, and pseudo-label quality to selectively acquire informative samples while exploiting abundant unlabeled imagery. This enables a compact student to adapt to target sensing domains with reduced annotation and inference costs. We further introduce an optional Mixture-of-Experts (MoE) student with graph-based feature refinement to enhance representation capacity while retaining a lightweight footprint. We evaluate SatLabel on 11 remote-sensing datasets spanning core, extended, and unseen domains against RemoteCLIP zero-shot inference. SatLabel improves Macro-F1 on most core and extended datasets while maintaining strong cross-dataset transfer to unseen domains. More importantly, the Balanced student contains only 11.2 M parameters and occupies approximately 42.8 MB, compared with 151.3M parameters and 577 MB for RemoteCLIP, while requiring 3.65 versus 5.89 GFLOPs. Across four efficiency benchmarks, it achieves approximately 2x higher GPU-forward throughput and reduces energy per image on datasets. ---