# Research Papers — 2026-09 · bundle page 24 of 266

> Full-text markdown of papers #1151–1200 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/24
[← Previous bundle](https://aipod.dev/papers/archive/2026-09/page/23.md)
[Next bundle →](https://aipod.dev/papers/archive/2026-09/page/25.md)

### RBF-GNN: Rational Basis Functions for Pseudo-Coordinate based Graph Convolutions

Authors: Paweł Batorski, Abtin Pourhadi, Paul Swoboda
Published: 2026-09-29
Categories: cs.CV
arXiv: https://arxiv.org/abs/2609.37015v1

We propose RBF-GNN, a new pseudo-coordinate based graph neural network architecture that takes into account Euclidean, spherical or angular coordinates and uses them to induce a powerful spatial inductive bias. Similar in architecture to SplineCNN, we improve upon the latter by replacing the less efficient sparse-activation based B-splines whose number grows exponentially with dimension by rational Padé basis functions. For effective training we propose a spline-subspace initialization and a variance-preserving weight rescaling. Experimentally, we evaluate on a number of popular neural network architectures that use SplineCNNs. We replace only the SplineCNNs with RBF-GNN. We achieve improved results, including on semantic keypoint matching, shape matching, event based camera computer vision tasks. We will make our implementation publicly available upon acceptance of the paper.

---

### Embedded Bi-Temporal Building Damage Assessment for On-Board Data Reduction

Authors: Thomas Goudemant, Benjamin Francesconi, Marjorie Bellizzi, Adrien Dorise
Published: 2026-09-29
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2609.37013v1

Rapid assessment of building damage after natural disasters is essential to support emergency response. Earth Observation satellites can acquire relevant imagery shortly after an event, but exploitation is limited by uplink and downlink capacity and by ground-processing latency. We address this with a bi-temporal building damage assessment pipeline built on a siamese detector derived from YOLOX, designed to compress information at both ends of the ground/space link. On the ground, pre-disaster reference images are encoded into a compact latent space -- compressed by up to a factor of 64 -- and uplinked to the satellite. On board, this reference is compared with a fresh post-disaster acquisition so that the downlink carries only actionable object-level products, bounding boxes and damage classes, instead of full scenes. This cuts the data exchanged in both directions, while on xBD the strongly compressed reference still preserves most of the detection performance. Because on-board acquisitions suffer from residual pre/post co-registration errors, we introduce a latent-space shift estimation and correction module that regresses the global offset from the coarse feature level and realigns the post-disaster features before fusion. It substantially improves robustness to de-registration -- especially under large shifts, where fusion-only variants collapse -- while also raising nominal accuracy and remaining compatible with the strongest compression. We finally port the pipeline to two embedded targets, a Xilinx Versal VCK190 and an NVIDIA Jetson AGX Orin, and report hardware performance (latency, throughput, power efficiency). The core detector and its compression port cleanly to both, but the operators needed for long-range robustness survive only on the Jetson GPU, whereas the Versal DPU does not.

---

### CADOC: Cache-Aware Dynamic Object Context for Long-Horizon Agents

Authors: Junjie Yao, Zhangchen Zhou, Zhi-Qin John Xu
Published: 2026-09-29
Categories: cs.AI
arXiv: https://arxiv.org/abs/2609.37012v1

For a long-horizon agent, context is the bottleneck: the history is resent with every request, the window caps task length, and reasoning degrades as the history grows. Replacing structured objects with compact retrieval Cards shortens the prompt and keeps the exact originals retrievable, but editing the history can break prefix-cache reuse, and prior recoverable methods time their edits by forecasts of future reuse or by preset intervals. We propose CADOC (Cache-Aware Dynamic Object Context), an online algorithm that replaces structured objects with compact Cards while preserving exact, on-demand retrieval of their original contents. CADOC schedules replacements in batches by balancing accumulated waiting cost against shared cache-reconstruction cost. Its scheduling rule follows from an economic order quantity trade-off, recovers the optimal integer batch under stationary assumptions. Across evaluation, CADOC consistently achieves the lowest aggregate input cost among the compared configurations, which reduces input cost by approximately 40\% on average while maintaining task performance close to full context. CADOC thus provides a cost-derived approach to compressible context management, demonstrating that efficient compression depends not only on shortening prompts but also on scheduling edits to preserve cache reuse.

---

### OPFL: Optimistic Verification of Federated Learning via Empirical Boundary

Authors: Hongxu Su, Jianzhu Yao, Xuechao Wang, Pramod Viswanath
Published: 2026-09-29
Categories: cs.CR, cs.AI
arXiv: https://arxiv.org/abs/2609.37011v1

Federated learning enables multiple clients to collaboratively train models without sharing their private data. However, the lack of visibility into local training makes it difficult to verify whether clients follow the prescribed training procedure or submit malicious updates, such as model poisoning. A natural approach is to replay client training for verification. However, privacy-preserving replay produces numerical results that cannot be directly matched with local client execution because the two run in different environments. We present OPFL, an optimistic verification framework for privacy-preserving federated learning. To protect data privacy, OPFL performs replay inside secure multi-party computation (MPC). Although gradients computed on MPC and local GPUs are not bitwise identical, we observe that their absolute differences are stable and bounded. OPFL therefore calibrates an empirical boundary offline and uses it to distinguish benign numerical deviations from malicious manipulation. To reduce the cost of expensive MPC replay, OPFL adopts optimistic verification by post auditing only sampled training steps. Experiments on LeNet, BERT, and Qwen show that the boundary generalizes across datasets, input lengths, and GPUs, while achieving $0$\% ASR against model poisoning and PGD-based attacks. On a LeNet workload, at $p=0.01$, OPFL is approximately $98.6\times$ faster than full MPC-based FL and $625.5\times$ faster than ZK-based approach.

---

### World2Motion: Turning Video World Models into 3D Human Motion Generators

Authors: Tu Fangyuan, Xiangyue Zhang, Yiyi Cai, Yichen Peng, Kunhang Li, Bo Zheng, Zhixiang Wang, Kaipeng Zhang, Erwin Wu, Haoran Xie, Haiyang Liu
Published: 2026-09-29
Categories: cs.CV
arXiv: https://arxiv.org/abs/2609.37004v1

We present World2Motion, a framework that generates scene-aware 3D human motion and corresponding video from a single image and a text prompt. While existing 3D motion generators learn from motion datasets, their generalization is constrained by limited coverage of environments. In contrast, video world models such as Cosmos 3 offer broader environmental priors but are not designed for full-body motion generation; recovering motion from their generated videos requires costly two-stage inference. To address these, we turn Cosmos 3 into a single-stage 3D motion generator. This adaptation has two challenges: the scarcity of paired video--motion data and temporal instability in the generated motion. First, we construct a training dataset combining synthetic video--motion pairs with real videos paired with estimated 3D motion. Second, we propose a shift-decoupled noise schedule that assigns different noise levels to video and motion through shared denoising progress. This design accommodates the different denoising requirements of the two modalities, reducing motion jitter. Experiments on a multi-source interaction benchmark show that World2Motion has better motion--text alignment and scene interaction compared with the evaluated 3D motion generators. It also matches the interaction success rate of the two-stage baseline while achieving approximately 3.3$\times$ faster inference.

---

### VesselBench-800K: A Large-scale Perception Benchmark for Multimodal Vessel Detection, Counting, and Density Estimation

Authors: Danfeng Hong, Chenyu Li, Jocelyn Chanussot
Published: 2026-09-29
Categories: cs.CV
arXiv: https://arxiv.org/abs/2609.37003v1

Vessel perception from space is crucial for a wide range of maritime applications, from traffic monitoring to environmental protection. However, most existing datasets predominantly focus on general object detection tasks in optical remote sensing (RS) images. Relying solely on single-modality optical RS images proves inadequate for effectively perceiving vessel objects in complex maritime scenarios, where ever-changing weather conditions (e.g., clouds and rain), the need for day-and-night coverage, and the inherent limitations of a single imaging modality pose significant challenges. To fill this gap, we introduce VesselBench-800K, the largest-to-date benchmark dataset on a global scale for vessel perception in multimodal RS images. As its name suggests, VesselBench-800K comprises 800,000 images, each at a resolution of 512x512 pixels, specifically curated for vessel perception tasks such as detection, counting, and density estimation. These multimodal image pairs (i.e., optical, SAR) are collected from diverse platforms, sensors, scenes, shooting heights, and synthetic sources, spanning spatial resolutions from 4.5m to 0.1m. Furthermore, we evaluate numerous state-of-the-art detection, counting, and density estimation models on VesselBench-800K through both qualitative and quantitative comparisons. By revealing previously unrecognized cues, this dataset holds immense potential to significantly advance our understanding of marine traffic. Our VesselBench dataset will be publicly available at https://github.com/danfenghong/IEEE_TGRS_VesselBench to support and contribute to community development.

---

### Visual Parallel Search: Learning to Search High-Resolution Images with Parallel Tile Inspection and Adaptive Zoom

Authors: Xijia Tao, Yihua Teng, Xinyu Fu, Cheng Gong, Ziru Liu, Xudong Xie, Rui Liu, Lingpeng Kong
Published: 2026-09-29
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2609.37002v1

High-resolution visual question answering often fails because a multimodal model does not acquire the small, spatially localized evidence needed to answer a question. Sequential zooming can recover detail, but it asks the main model to choose a region before obtaining a reliable overview. We introduce VPS, a visual parallel-search framework in which a main agent first invokes grid_search to inspect image tiles in parallel with question-conditioned sub-agents, and then adaptively invokes zoom_in PSisual Parallel Search improves mean accuracy over dedicated zoom-only search in 14 of 15 same-model comparisons, with gains up to 8.0 points and especially strong improvements for smaller main models. ZoomBench retains an approximately 3.2-point gain at every tested size. We further develop a supervision pipeline with hint-free verification and a paired role-specific GRPO surrogate for learning the controller and tile-reader roles. SFT improves observed accuracy on all five benchmark splits, including a 4.17-point gain on HR-Bench 4K. Role-specific RL further reshapes search behavior: main-only RL reduces mean tool use from 2.65 to 2.11 with similar pass@1 in an internal four-response evaluation, while external accuracy changes are mixed. Joint training reveals an asymmetry between local evidence reading and global search control. Together, these results support VPS as an effective inference-time scaffold and a trainable decomposition for visual evidence acquisition.

---

### Parameterized Stripe Attention for Efficient Video Generation

Authors: Xingyu Jia, Baole Ai, Ang Wang, Kang Zhao, Yong Li
Published: 2026-09-29
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2609.37001v1

Diffusion Transformers (DiTs) enable high-quality video generation but suffer from substantial inference latency, primarily attributable to the computationally expensive full spatio-temporal attention. While sparse attention methods offer potential solutions, existing approaches face an inherent flexibility--efficiency dilemma: predefined masks lack the flexibility to capture diverse attention patterns, while runtime-determined masks introduce overheads and sacrifice hardware efficiency. We identify the lack of a unified structural characterization of DiT attention as a key limitation of existing methods, and establish that video DiT attention exhibits \textbf{periodic diagonal stripe structures} along both temporal and spatial dimensions. To formally encode these structured patterns within a single efficient kernel, we present {\bf PSA}, a parameterized stripe attention that formalizes the observed stripe regularity, unifying diverse attention patterns for efficient mask generation. This unified representation enables a single hardware-efficient CUDA kernel to process all sparse patterns, achieving FlashAttention-3-level Model FLOPs Utilization. To determine optimal sparsity configurations, we propose a training-free offline search algorithm that automatically maximizes sparsity under a specified error tolerance for each attention head. Experiments on HunyuanVideo and Wan~2.1 demonstrate that PSA achieves 1.57$\times$ and 1.37$\times$ end-to-end speedups over FlashAttention-3 baselines, with acceptable visual quality degradation.

---

### Cross-Organizational SysML Model Integration: A Survey of Challenges and AI-Supported Tasks

Authors: Zirui Li, Torsten Brix, Stephan Husung
Published: 2026-09-29
Categories: cs.SE, cs.AI
arXiv: https://arxiv.org/abs/2609.37000v1

Cross-organizational collaboration is widely regarded as a key promise of SysML-based Model-Based Systems Engineering (MBSE), yet practitioners still face persistent challenges when exchanging and integrating system models. In parallel, Large Language Models (LLMs) raise expectations for AI-assisted model understanding and integration, while reliability and required human oversight continue to pose challenges. This paper reports the results of an online questionnaire survey with 29 MBSE stakeholders involved in cross-organizational collaboration. Respondents rated eight predefined integration challenge categories and six AI-supported task types on five-point Likert scales. The results indicate that stakeholders perceive model integration as a multi-dimensional alignment problem across semantics, behavior, traceability, and exchange interoperability. These perceptions vary by organizational role and frequency of integration involvement. AI is rated highly useful for analysis tasks such as semantic structure analysis and inconsistency detection, and respondents predominantly prefer human-in-the-loop use with mandatory verification. These findings motivate AI support that enhances, rather than replaces, engineering responsibility in SysML-based integration.

---

### ImbalancE: Inference-Time Latent Search Against Degree Imbalance in Link Prediction

Authors: Alberto Bernardi, Luca Costabello, Christophe Gueret
Published: 2026-09-29
Categories: cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2609.36996v1

Knowledge Graph Embedding models have been extensively used to learn representations of entities and relations in Knowledge Graphs for predicting missing links. However, the quality of the learned representations varies a lot across different areas of the graph. If previous research has loosely linked the problem to relation types or degree bias, we show that it is more widespread and it correlates with the degree imbalance of the entities in test triples. In particular, the prediction of a target entity that has a degree much smaller than the degree of the anchor entity is extremely problematic. This is critical in recommender systems and other use cases, where these triples represent important corner cases. To address this issue, we propose an inference-time latent search optimization method capable of significantly improving model predictions on the most imbalanced triples. Built on top of a pre-trained model, it explores the embedding space at evaluation time, blending known and out-of-band information to mitigate the degree imbalance bias. We show the value of our approach on imbalanced triples from common benchmark datasets, where we outperform conventional methods, opening the door to the successful adoption of Knowledge Graph Embedding models on these critical corner cases.

---

### Salt++: Context-Aligned Post-Training for Few-Step Streaming Multimodal Generation

Authors: Xingtong Ge, Yutong Wang, Lunjie Zhu, Haitao Lin, Fangyu Lin, Yushi Huang, Xin Zhang, Yi Zhang, Yu Liu, Jun Zhang
Published: 2026-09-29
Categories: cs.CV, cs.AI, eess.IV
arXiv: https://arxiv.org/abs/2609.36995v1

Few-step streaming audio--video generation requires both causal modeling and step distillation, yet standard training recipes face two context-related challenges. Teacher forcing pairs clean history with a noisy target, but supervises predictive contextual representations only indirectly through velocity prediction. Meanwhile, directly reusing bidirectional score models in causal Distribution Matching Distillation (DMD) creates a mismatch between generation and scoring contexts. We address these challenges with Salt++, a two-stage post-training framework comprising Causal Self-Flow (CSF) and context-aligned autoregressive DMD. CSF exploits contextual information asymmetry by varying the history while keeping the noisy target fixed: a noise-mixed-history student aligns its intermediate representations with those of a clean-history exponential-moving-average teacher. This self-supervised signal encourages the student to extract semantic information and improves cross-modal alignment. Context-aligned AR DMD shares the causal mask and prefix across generator sampling, fake-score training, and real-score evaluation to match generated and reference distributions under a block-conditional KL objective. With calibrated teacher guidance, it performs clean-prefix few-step distillation and then adapts to generated histories without switching objectives or requiring separate consistency distillation. At 480p, Salt++ improves visual and motion quality by 57% and 45% over OmniForcing on JavisBench under the same 4-step causal setting. A separate scale-wise post-training stage extends Salt++ to 4-step $1664\times960$ generation, outperforming bidirectional LTX-2 on six of seven reported metrics. Project page: https://xingtongge.github.io/Saltpp

---

### CypherTurn: A Multi-Turn Benchmark for Conversational Text-to-Cypher Evaluation and the Autonomy Divergence

Authors: Yuzhe Zhang, Weijie Zhu, Haolin Yang, Ziyun Zhang, Xianwei Xue, Mengke Chen, Qiutong Pan, Huaqian Cai
Published: 2026-09-29
Categories: cs.CL
arXiv: https://arxiv.org/abs/2609.36987v1

Graph databases are increasingly queried through natural language, yet every existing benchmark evaluates isolated single-turn queries rather than the multi-turn sessions through which analysts actually work. We introduce CypherTurn, the first benchmark for conversational Text-to-Cypher evaluation, comprising 721 sessions and 5,927 turns across 7 knowledge graphs and 13 conversational phenomena. We evaluate 15 models under a guided oracle protocol and a fully autonomous agentic protocol, yielding four findings. First, the best model reaches only 64.7% execution accuracy, and session-level correctness remains below 5%. Second, despite strong overall rank correlation, frontier models exhibit a consequential reordering of the top of the leaderboard under autonomous operation, a phenomenon we term the Autonomy Divergence, which reveals error-management as a partially independent capability from raw generation skill. Third, scaling action budgets from x3 to x10 fails to close the autonomy gap, as the strongest frontier models self-limit to approximately two actions per turn regardless of available budget. Fourth, single-turn Cypher fine-tuning degrades multi-turn instruction following, while architecture-appropriate specialization outperforms several frontier models. These results establish CypherTurn as an open challenge for conversational graph database reasoning. Code and data are available at https://github.com/BarryQ/CypherTurn.

---

### CF-LoRA: Decoupled Factor Aggregation and Adaptation-Aware Client Clustering for Federated LoRA Fine-Tuning

Authors: Mengjun Yi, Langxing Yang, Suhan Guo, Furao Shen, Jian Zhao
Published: 2026-09-29
Categories: cs.AI, cs.DC
arXiv: https://arxiv.org/abs/2609.36986v1

Federated LoRA fine-tuning enables parameter-efficient adaptation of pre-trained models without sharing private data, but suffers from two fundamental mismatches under heterogeneous client data: a structural aggregation mismatch caused by independently averaging LoRA factors, and a statistical collaboration mismatch caused by enforcing a single global adapter across divergent clients. To address these issues, we propose CF-LoRA, a clustered federated LoRA fine-tuning framework that combines decoupled factor aggregation with adaptation-aware client clustering. CF-LoRA first learns a globally shared $A$ factor while retaining personalized $B_i$ factors, then identifies clients with similar adaptation patterns based on the cosine similarity of their learned $B_i$ factors, and finally performs intra-cluster $B$-factor aggregation with a frozen $A$ factor. By decoupling LoRA factor aggregation, CF-LoRA preserves the low-rank structure and mitigates the structural aggregation mismatch, while adaptation-aware clustering promotes collaboration among clients with similar adaptation patterns and reduces negative transfer caused by statistical heterogeneity. Experiments on four language tasks and four vision datasets with RoBERTa and ViT show that CF-LoRA achieves the highest average accuracy in both modalities while communicating only one LoRA factor per optimization round.

---

### Abductive World Modeling via Causal Representation Learning

Authors: Ziqi Liu, Songhan Yang, Linfan Zhou, Jiatong Liu, Lijun Peng, Long Wan, Yinqi Bai
Published: 2026-09-29
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2609.36985v1

The central challenge of world modeling is to learn representations that capture how the world evolves. However, existing world models predominantly represent future states without explicitly capturing the latent causes underlying their evolution, limiting their ability to reason about why and how the world changes. To address this limitation, we propose Abductive World Modeling (AWM), a framework that learns structured causal representations by abductively inferring latent causes from predicted futures. Specifically, we realize AWM through the Hierarchical Abductive State Pyramid (HASP), which organizes the inferred world state into three complementary components - Entity, Dynamic, and Relation - capturing what exists, how it changes, and how entities interact, respectively. By jointly reasoning over the current observation and its predicted future, HASP abductively infers these latent factors and integrates them into a structured state representation for downstream reasoning. To the best of our knowledge, AWM is the first framework to introduce abductive state inference into latent-space world modeling for learning structured representations of world dynamics. Experiments across physical prediction, causal reasoning, and action understanding demonstrate the effectiveness of our approach. Compared with V-JEPA, a state-of-the-art latent-space world model, AWM improves physical prediction AUROC by 10.7%, causal reasoning accuracy by 16.8%, and action Top-1 accuracy by 68.0%.

---

### REALHOP: Rethinking Multi-Hop Reasoning Evaluation via Behavioral Auditing

Authors: Jiawen Tao, Xiaokun Yuan, Yaoming Li, Chenxu Liu, Mengzhou Wu, Tong Yang, Maxm Pan
Published: 2026-09-29
Categories: cs.AI
arXiv: https://arxiv.org/abs/2609.36984v1

Complex questions often require multi-hop reasoning that connects facts distributed across sources or distant regions of a long context through intermediate steps. Benchmarks commonly evaluate this ability with questions built around predefined reasoning chains, treating a correct answer as evidence that the intended composition was used. Yet answer correctness alone leaves open whether success depends on the evidence associated with each intended step: models may instead rely on memorized associations, shorter paths, or partial evidence. We examine this dependence using the Behavioral Necessity Rate (BNR), which measures how often targeted evidence removal prevents answer recovery on initially correct instances. Across five existing benchmarks, panel-mean BNR ranges from 16.6% to 48.9%, exposing a substantial gap between annotated structure and observed dependence. Guided by this diagnosis, we introduce REALHOP, a diagnose-construct-verify framework that rebinds entities, factorizes selected relations, adds complete competing paths, and places evidence at traceable locations. Structural and semantic checks precede freezing; behavioral interventions follow. On 790 paired MuSiQue questions, REALHOP raises panel-mean BNR from 27.4% to 94.4% while retaining high Full accuracy. It also yields high BNR on REALHOP-FRAMES and REALHOP-LONGBENCH. On 216 long-context questions, the matched multiple-choice spread across 16 models grows from 13.9 to 59.2 points and persists under repeated open-ended evaluation. Together, these results show that a conceptually coherent chain and a correct final answer do not by themselves establish multi-hop reasoning. Verifying that success depends on every intended hop is therefore as fundamental to multi-hop evaluation as measuring answer accuracy itself.

---

### SRJudge: Empowering Large Language Models with Selective Reasoning for Fine-Grained Knowledge Concept Tagging

Authors: Zhiwei Yang, Jiahua Yang, Huiru Lin, Xing Chen, Quanlong Guan
Published: 2026-09-29
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2609.36982v1

Knowledge concept tagging aims to assign specific concept or topic labels to educational content, which is essential for both educators and learners in traditional and online teaching practices. Recent work has explored large language models (LLMs) for this task, achieving promising performance. However, LLMs still struggle to select the correct concept from a large-scale candidate set due to the high dimensionality of the decision space. In this paper, we propose a novel three-stage Select-Reason-Judge (SRJudge) framework, which empowers LLMs with selective reasoning capability for fine-grained knowledge concept tagging. Specifically, the Selector in Stage 1 first narrows the candidate concepts to a top-K shortlist by fine-tuning a small language model (SLM), e.g., BERT, since the top-$K$ predictions hit the correct concept in most cases, thereby reducing the decision space of correct candidates. Next, the Stage 2 Reasoner employs a lightweight LLM for refined reasoning over the shortlisted candidates. It further integrates an improved reinforcement learning strategy with a dynamic task-specific reward function and a pruning mechanism to better align with human reasoning preferences. Finally, a larger LLM acts as a judger that evaluates the overall rationality of the reasoning process and its explanations to determine the final output. In addition, we construct two high-quality datasets for further validation, i.e., the biology dataset S_Bio and the physics dataset S_Phy. Experimental results demonstrate that our method consistently outperforms state-of-the-art baselines across benchmark datasets, verifying its effectiveness and superiority. Resources are available at: https://github.com/Nicozwy/SRJudge.

---

### UltraMatch: Transport Path Routing for Ultra-Fast and Memory-Efficient Image Matching

Authors: Jiajun Le, Yifan Lu, Zizhuo Li, Lei Cao, Junjun Jiang, Jiayi Ma
Published: 2026-09-29
Categories: cs.CV
arXiv: https://arxiv.org/abs/2609.36980v1

Despite recent advances in accuracy and efficiency, coarse matching remains an indispensable yet costly stage in existing semi-dense matchers due to dense token-level matching. We present UltraMatch, an ultra-efficient and scalable semi-dense matching framework that bypasses the quadratic computation and memory cost of dense token-level matching by routing only a small fraction of candidate matching paths. At its core, a lightweight Transport Path Router operates on coarse block representations to rank candidate target blocks for each source block and retain only a small set, restricting subsequent token-level matching to the selected paths and avoiding the construction of the full token-to-token matching matrix. We further design a sparse global Dual-Softmax that performs matching only over the routed block candidates while retaining global competition across the sparse matching space. Beyond matching acceleration, UltraMatch employs deployment-oriented structural reparameterization for feature extraction and a tiny fine matching head with shared parameters, further reducing inference cost and memory consumption. UltraMatch achieves competitive accuracy among semi-dense matchers, while running 1.67$\times$ faster than SuperPoint+LightGlue with only 0.44 GiB peak inference memory. Its scalability enables inference at up to 6K resolution on a single RTX 3090, whereas existing semi-dense matchers run out of memory before reaching 2K. Our routing strategy is also transferable, delivering about 2$\times$ end-to-end speedup in EDM and ELoFTR without accuracy loss. The project repository is available at https://github.com/JiajunLe/UltraMatch.

---

### AMU:Admission and Memory Update for Personalized Conversations---Structured Memory with SLM Guided Control

Authors: Tao Hwang, Yishi Diao
Published: 2026-09-29
Categories: cs.CL
arXiv: https://arxiv.org/abs/2609.36976v1

Large language models (LLMs) have become the foundation of personalized assistants, but maintaining persistent user memory across long-term interactions remains challenging. Existing memory systems often focus on storage, retrieval, or consolidation, while memory writing remains less controlled: transient requests, duplicate statements, and outdated user states may enter memory and later be retrieved for personalization. In this paper, we present AMU: Admission and Memory Update for Personalized Conversations, an SLM-guided (Small language model guided) structured framework for writing-time memory control. AMU uses structured memory filtering to decide what should enter memory and SLM-guided storage management to determine whether an admitted record should be stored separately, discarded as a duplicate, or fused as an update. We evaluate AMU in a controlled memory writing and retrieval setting. Experimental results show that AMU maintains cleaner and more retrievable personalized memories.

---

### A Dual-Track Curation-and-Classification Framework for Resolving Ground-Truth Label Noise in Operational Sentinel-2 Wheat Area Estimation

Authors: Kasimali Agharia, Ujjwal Kumar Gupta
Published: 2026-09-29
Categories: cs.CV
arXiv: https://arxiv.org/abs/2609.36975v1

Operational estimation of wheat-cultivated area is persistently constrained by discordance between administrative record-keeping and remotely sensed classification products. We address this administrative reference discordance for the 2022 Rabi season in Patiala district, Punjab, India, using a thirteen-timestep Sentinel-2 NDVI time series. A curated 849-sample reference dataset, developed through an iterative rule-based bootstrapping procedure, underpins both a feature sensitivity analysis and an operational classifier. Feature sensitivity independently assessed via Cohen's d and gradient-boosted information gain converges on the February-to-March grain-fill window as most discriminative. Four classifiers (1D-CNN, LSTM, hybrid CNN-LSTM, and XGBoost) were benchmarked on an identical 679/170 sample split. XGBoost achieved the highest overall accuracy (78.82%) against deep-learning baselines (64-66%), consistent with tree-based ensembles' favourable parameter-to-sample ratio in low-sample regimes. At full-population deployment across 36.25 million valid district pixels, the operational classifier attained 86.31% precision and 71.05% recall. The predicted wheat extent deviated by only +2.99% from the official tabular target, whereas the government's spatial reference mask exhibited a +25.11% positive area bias against the identical target. This asymmetry indicates that a classifier trained on an auditor-curated reference set reconciles more closely with the official tabular area than the spatial product conventionally used to validate it. We present this dual-track curation-and-classification framework as a methodological reference for crop-area reconciliation in label-noisy administrative settings.

---

### Repetition, Not Length: Isolating the Counting Failure in Neural Text-to-Speech

Authors: Kirill Borodin, Vasilii Kudryavtsev, Maxim Maslov, Grach Mkrtchian
Published: 2026-09-29
Categories: cs.CL, cs.SD
arXiv: https://arxiv.org/abs/2609.36974v1

Text-to-speech models loop, truncate and lose count on text that repeats a phrase many times. We show that repetition itself is what breaks them, not the length that comes with it. Every repeated sentence in our test set is paired with a control of matched sentence and word count in which no word ever repeats back-to-back. Six models from three architectures render the controls almost perfectly and fail the repeated twins: 94.3% against 18.2% exactly right at k >= 6. The gap survives greedy decoding, repetition-penalty sweeps, four independent speech recognisers and 420 analysis specifications without once reversing sign; a held-out fourth architecture lands within a point of its predicted gap, and one of two non-autoregressive baselines shows the same failure. Varying the period of the text shows the failure grows smoothly with periodicity, half of it surviving when no word is adjacent to itself.

---

### Structured Visual Target Learning For Cross-Subject eeg-to-image retrieval

Authors: Salini Yadav, Taveena Lotey, Mickaël Coustaty, Pravendra Singh, Partha Pratim Roy
Published: 2026-09-29
Categories: cs.CV
arXiv: https://arxiv.org/abs/2609.36971v1

Cross-subject EEG-to-image retrieval requires a neural represen- tation trained on source subjects to remain aligned with a visual embedding space for an unseen subject. Whereas existing methods primarily focus on the EEG side, we address this problem from the perspective of the visual target. Our approach preserves the spatial information of the Perception Encoder, converts its patch grid into a compact set of learned visual views, and aggregates them for each image with a block-structured, content-dependent router. The target is learned jointly with the EEG encoder through contrastive learning with MMD regularization across source subjects. For deployment, we propose a training-free representation refinement that aligns frozen embeddings without updating either encoder. Under leave- one-subject-out evaluation on THINGS-EEG2, the structured target achieves 35.3%/65.6% Top-1/Top-5 accuracy, the best among com- pared methods. Refinement raises this to 48.1%/77.1%, an 18.5% Top-1 gain over the strongest compared method, improving all ten held-out subjects.

---

### Equally Good, Yet Different: Benchmarking Rashomon sets in AutoML packages

Authors: Katarzyna Woźnica, Katarzyna Rogalska, Zuzanna Sieńko, Mustafa Cavus
Published: 2026-09-29
Categories: cs.LG
arXiv: https://arxiv.org/abs/2609.36970v1

The Rashomon effect describes the existence of multiple near-optimal models that achieve comparable performance while offering fundamentally different explanations. This creates a critical vulnerability in AutoML: x-hacking, the selective post-hoc choice of a model based on its explanation rather than predictive merit. No existing AutoML framework exposes this risk. We introduce ARSA ML, an open-source Python framework that quantifies Rashomon set structure and predictive multiplicity within AutoML pipelines. Using ARSA ML, we benchmark AutoGluon and H2O across 28 binary classification datasets, and conduct a post-hoc x-hacking analysis revealing a consistent structural asymmetry: AutoGluon produces larger, diverse sets with stable explanations, while H2O generates compact sets with markedly higher prediction divergence and explanation instability -- making H2O users considerably more exposed to x-hacking. This gap persists across all evaluated metrics and epsilon thresholds, pointing to a fundamental difference in each framework's model-building strategy. ARSA ML is available at https://pypi.org/project/arsa-ml/ .

---

### Prior-Driven Enhancements in 3D Gaussian Splatting: Normals and Depths Regularization

Authors: Gyeonggwan Lee, Seunghwan Hong, Junghun Suh
Published: 2026-09-29
Categories: cs.CV
arXiv: https://arxiv.org/abs/2609.36969v1

3D Gaussian Splatting (3DGS) is a state-of-the-art technique for 3D scene rendering, offering high efficiency and excellent visual quality. However, because 3DGS relies on an initial sparse point set from Structure-from-Motion (SfM) and view-dependent properties, it can suffer from geometric inaccuracies and visual artifacts, particularly in complex scenes. To address these challenges, we propose an improved 3DGS approach that regularizes the optimization process by integrating geometric priors, including surface normals and dense depth information. Surface normal regularization improves geometric consistency by aligning Gaussian covariance with local surface structures, while dense depth priors combined with an initial points from SfM enhance per-pixel depth estimation, increasing accuracy and reducing ambiguities. These enhancements enable robust handling of diverse and complex real-world scenarios, minimizing visual distortions and improving reconstruction quality across various environments. To validate our method, we evaluate it on challenging datasets, including street-view scenes and highly reflective environments, while testing it across multiple SfM pipelines. Our results demonstrate compatibility across diverse environments and highlight the robustness of our approach. Experimental findings further show that our method enhances geometric accuracy and visual quality, establishing a reliable solution for real-time 3D scene rendering in complex environments.

---

### TaskBridge: Bridging Unsupervised Tabular Anomaly Detection and In-Context Learning via Virtual Tasks

Authors: Doyun Choi, Dooho Lee, Jaemin Yoo
Published: 2026-09-29
Categories: cs.LG
arXiv: https://arxiv.org/abs/2609.36968v1

Unsupervised tabular anomaly detection (TAD) aims to identify anomalous rows in tabular data using normal training samples. While conventional methods rely on dataset-specific training and configuration search, recent tabular foundation models (TFMs) enable zero-shot anomaly detection on unseen datasets via in-context learning. Most TFM-based approaches, however, require anomaly-specific pretraining from scratch, making detection inherently dependent on synthetic TAD-specific priors and costly to update. Some approaches instead repurpose pretrained general-purpose TFMs for TAD to avoid this burden, but rely on computationally expensive formulations with restrictive anomaly inductive biases. In this work, we introduce TaskBridge, a new framework that efficiently repurposes pretrained general-purpose TFMs for unsupervised TAD by constructing virtual supervised tasks that directly recast anomaly detection as supervised in-context inference of TFMs. The resulting virtual tasks induce predictive structures under which normal queries and their target pairs receive high support, whereas anomalies tend to violate the induced structures and receive lower support, providing direct anomaly evidence. Across 790 real-world datasets, TaskBridge consistently outperforms 30 baselines, including state-of-the-art TFM-based approaches, without anomaly-specific TFM pretraining or dataset-specific model optimization.

---

### JudgeCast: Time Series Forecasting with Experience-Informed Covariate Judgements

Authors: Donguk Kwon, Wooseok Jeong, Dongha Lee
Published: 2026-09-29
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2609.36966v1

Covariate effects vary across contexts and shift over time, requiring forecasters to assess how to use them for each forecasting context. As forecasting proceeds, observations for earlier forecasts become available, providing feedback on past covariate use for subsequent forecasts. However, when multiple covariates act together, the forecast error reveals the numerical discrepancy from the observation but not how the covariates should have been used. We introduce JudgeCast, an experience-based framework for time series forecasting with covariates. Following the judgmental adjustment practice, a frozen TSFM provides the base forecast, while a frozen LLM uses the current context and relevant experience to adjust it. Within the adjustment, assessing covariate effects and determining the numerical adjustment serve distinct roles, so JudgeCast first forms explicit covariate-wise judgments and then determines the adjustment. After observation, JudgeCast uses the observed residual of the base forecast to reconstruct alternative judgments and evaluates the original and alternatives through their resulting adjustments. The best-performing decision is selected and retained as validated experience for subsequent forecasts. Across diverse real-world datasets, JudgeCast outperforms strong baselines. Ablations show that explicit covariate-wise judgment can improve forecast-time adjustment, while residual-guided experience construction yields more reliable forecasting gains than retaining raw decisions as experience.

---

### Chinese-Jev: Bringing System One Model to Chinese-Language Tasks

Authors: Zexiao Wang, Zihao Zhang, Xudong Wang, Pan Wang, Ziyi Ye, Haoyu Zhao, Zuxuan Wu, Shuicheng Yan
Published: 2026-09-29
Categories: cs.CL, cs.CV
arXiv: https://arxiv.org/abs/2609.36965v1

System One models such as Jev offer an efficient alternative to generative language models for tasks that require decisions rather than open-ended responses. However, existing Jev models exhibit limited Chinese-language decision accuracy, restricting their utility in both general and specialized settings. In this paper, we introduce Chinese-Jev, a System One model that addresses this gap through a unified data processing and training pipeline. Our data processing protocol converts heterogeneous Chinese-language annotations into probability targets over candidate options, enabling a shared training formulation across domains and question formats. To enable efficient inference, Chinese-Jev adopts a lightweight encoder-only backbone for text encoding and learns to score candidate answers through decision-oriented training. To address the misalignment between the pre-training distribution and downstream Chinese-language scenarios, we first train the model on a general-purpose corpus of 10 million examples, then fine-tune it separately for the medical, legal, and financial domains. To evaluate decision accuracy and calibration in both general and domain-specific Chinese-language settings, we introduce Chinese-Jev Bench (CJ-Bench). After first-stage pre-training, Chinese-Jev exceeds the accuracy of the closed-source Jev model by 1.24% on general-domain tasks while achieving a 20.3x speedup. Subsequent domain-specific fine-tuning yields a 4.0% accuracy improvement over Jev in medicine and achieves 92% of Jev's average accuracy across specialized domains, with a 17x speedup and an average latency of only 15 ms per example. We further demonstrate on-device deployment of an INT8-quantized model on mobile devices, achieving an inference latency of approximately 1.0 second per decision. The project is available at https://gulucaptain.github.io/Chinese-Jev/.

---

### VStress: Correlation-Aware Auditing and Adaptive Budget Allocation for Repeated Verifiers

Authors: Miaobo Hu, Shuhao Hu, Xiaobo Guo, Xin Wang, Bokun Wang, Peng Zhang, Daren Zha, Jun Xiao
Published: 2026-09-29
Categories: cs.LG, cs.AI, cs.CL, cs.CR
arXiv: https://arxiv.org/abs/2609.36958v1

Repeated verifier calls are useful only when they contribute conditional information. We introduce VStress, an auditable replay contract, and VStress-CA, a correlation-aware allocation policy that estimates the conditional marginal information of an unqueried verifier on a sealed calibration split, discounts uncertainty, normalizes by call cost, and stops or abstains when the next call is not informative. The controller freezes its decision and cost ledger before joining the clean oracle; a dependence-shift alarm disables channel preference and falls back to exact-stop. The controlled audit gives the mechanism boundary: at 35% symmetric corruption, majority-5 improves balanced accuracy from 0.6578 to 0.7739, whereas at 65% it loses 0.1226 points. In the matched fixed-budget comparison, breadth, redundancy, and adaptive allocation obtain balanced accuracies 0.6048, 0.6375, and 0.6538, with 3.4216 calls per item and an RLVR score of 0.6417 for VStress-CA. Dependence diagnostics also increase from same-model repeats to cross-family channels, with conditional marginal gains of 0.0126, 0.0462, and 0.0913. These measurements turn correlation from a post-hoc warning into an auditable allocation decision.

---

### Beyond Readability: Evaluating Task Information Recoverability

Authors: Yiwei Liu
Published: 2026-09-29
Categories: cs.CV
arXiv: https://arxiv.org/abs/2609.36957v1

Direct visual readability and task-information recoverability are different quantities. Failure to decode a target from a fixed observation need not eliminate access to that target through another recovery route. We develop an evaluation perspective that makes the observation, query, target, and available knowledge explicit and measures the overlap between routes' success sets. For information available on the original visible surface under suitable imaging conditions, direct optical recovery reads the target from the image, optionally after restoration; entity-linked recovery uses residual visual evidence to identify the depicted entity and accesses its target through an entity--attribute relation in a specified knowledge resource. Such access can draw on stored knowledge or an external source. A controlled book-cover study instantiates external access with a fixed title--author catalog, comparing optical author recovery with visual title resolution and deterministic lookup under resolution degradation. Entity-linked successes persist across the tested vision--language models, revealing information access beyond the tested direct visual frontier despite substantial differences in absolute performance. A substantial optical-only region remains. These complementary outcomes show why visual degradation should be evaluated through the task information accessible along specified routes and knowledge resources, alongside direct readability.

---

### Controlled Decoding Attacks on Black-Box LLMs

Authors: Jesson Wang, Shawn Li, Wei Yang, Franck Dernoncourt, Ryan A. Rossi, Charith Peris, Yue Zhao
Published: 2026-09-29
Categories: cs.CR, cs.AI
arXiv: https://arxiv.org/abs/2609.36956v1

Manipulating next-token probabilities during generation can bypass the safety alignment of large language models. Existing approaches, however, rely on access to model weights or numerical token probabilities and therefore do not apply to interfaces that return only sampled text. Reconstructing probabilities from sampled outputs offers a possible alternative, but finite sampling produces sparse and noisy estimates, while repeating this process at every generation step incurs substantial query costs. Our empirical observations suggest that large distributional changes along successful jailbreak trajectories are concentrated at a small subset of positions, motivating selective control. We introduce \method{}, a framework for jailbreaking through text-only continuation interfaces that permit repeated sampling and assistant-prefix continuation. Sample-Based Distribution Reconstruction combines sampled outputs with a prior over unobserved actions to obtain a usable control signal. Risk-Gated Residual Control uses the evolving response prefix to decide when to reconstruct and modify the distribution, concentrating sampling costs at selected positions. Speculative Multi-Token Execution further amortizes target calls by verifying and accepting draft prefixes that require no intervention. Across four target endpoints and three benchmarks, \method{} achieves the highest mean score most comparisons against baselines.

---

### Purlin: Separating Orchestration from the Datapath of Collectives

Authors: Osayamen Jonathan Aimuyo, Swapnil Gandhi, Christos Kozyrakis
Published: 2026-09-29
Categories: cs.DC, cs.AI
arXiv: https://arxiv.org/abs/2609.36954v1

Distributed inference depends on GPU collective communication that must keep pace with evolving hardware and specialized workloads. However, existing collective implementations often couple semantics, orchestration (where and when data moves), and the datapath (how data moves). This coupling makes it costly to adopt new hardware mechanisms and customize communication for applications. We present Purlin, a scale-up communication framework that separates these concerns. At the top of Purlin, we specify collectives as a naming of an input and output layout and a copy or reduction operation. In the middle, we introduce a shared orchestration protocol, Stage, Notify, And Consume (SNAC), which derives coordination from these specifications. Below SNAC sits a hardware-specific datapath we call Atom, which implements two key data movement primitives for collectives: copy and reduce. This separation lets us customize collectives and adopt new hardware mechanisms while reusing orchestration via SNAC. We evaluate Purlin on A100, H200, and B200 GPUs. Across seven collectives, Purlin achieves latency speedups of up to 5.14x and bandwidth improvements of up to 4.50x over baselines. Integrated into SGLang, Purlin improves offline LLM serving throughput and interactivity by 1.13x on average and up to 1.37x over baselines. For online LLM inference, Purlin improves interactivity by 1.26x on average and up to 2.85x, with the largest gain occurring under overload. For diffusion image generation, Purlin reduces end-to-end latency by up to 1.13x.

---

### Cool the Sampler, Not the Learner: Sampling Temperature Moves the Staleness Cliff of Importance-Corrected GRPO

Authors: Taiheng Pan
Published: 2026-09-29
Categories: cs.LG, cs.CL
arXiv: https://arxiv.org/abs/2609.36953v1

Production RL for language models lets the sampler fall behind the learner and repairs the resulting mismatch with a truncated importance weight. We ask how long the sampler can go without a refresh under that correction, and find a cliff: on Qwen2.5-Math-1.5B and GSM8K, importance-corrected GRPO refreshed every 192 updates learns well for 180 steps and then degrades severely in all three data seeds before the refresh arrives. Published remedies for staleness act on the update; we act on the sampler instead. Decoupled cooling draws samples at temperature 0.8 while the learner, the reference model and the importance weights stay at temperature 1, with the behaviour probability recorded from the tempered distribution, so the learner's objective is unchanged. All corresponding cooled runs are stable, and the longer interval keeps what the short one delivered: at the same update budget, a cooled sampler refreshed every 192 steps matches an uncooled sampler refreshed every 96 at the end of training (0.857 for both) and averaged over it (0.79), whereas lowering the learning rate to a safe value ends 3-7 points lower. On Qwen2.5-Math-7B the degradation points at interval 192 predict that an interval of 144 is fatal without cooling and survivable with it; on two data seeds the uncooled runs degrade before their first refresh and the cooled runs pass it and end at 92-93% against 68-81%, with one cooled run degrading transiently late in the second cycle. The benefit has a window: at three times the safe interval and in a high-mismatch MATH setting cooling delays degradation without preventing it, stronger cooling is not better, and cooling without the correction collapses. Sampling temperature is a control on staleness tolerance, and temperature and refresh interval should be chosen together.

---

### ER-JEPA: Experience Replay Improves Joint-Embedding Predictive Learning in Language Models

Authors: Jingnan Pu, Zi-En Fan, Feng Lian
Published: 2026-09-29
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2609.36952v1

Large language models (LLMs) excel at token-level generation but may learn undesirable abstract semantics and lack comprehensive perception. LLM-JEPA mitigates this by aligning different views of the same underlying knowledge via a joint-embedding predictive architecture (JEPA). However, strong alignment does not necessarily lead to accurate, stable predictions. To address this, we propose ER-JEPA, which adds an episodic replay path to LLM-JEPA. ER-JEPA stores training pairs in a memory. At each step, it stores and retrieves relevant data to provide additional supervision. This enables learning from both the current batch and stored training pairs, providing additional supervision for token prediction and representation alignment. Experiments across multiple datasets (NL-RX, GSM8K, Spider, and NQ-Open) demonstrate that ER-JEPA consistently outperforms LLM-JEPA.

---

### Scalable Diffusion SBI for Compositional Inference under Simulator Misspecification

Authors: Vincent D. Zaballa, Elliot E. Hui
Published: 2026-09-29
Categories: cs.LG, q-bio.QM, stat.ML
arXiv: https://arxiv.org/abs/2609.36950v1

Simulation-based inference is challenging when many heterogeneous observations must be composed, hierarchical latent structure must be preserved, and the simulator is misspecified relative to observed data. We develop sampling and fine-tuning methods for diffusion-based inference in design-conditional settings, where the same simulator is queried across different experimental conditions $ξ$. We extend compositional score-based inference with a continuous-time diffusion coefficient that accounts for the number of observations, avoiding Jacobian and auxiliary-covariance corrections. We introduce Hierarchical Blockwise Diffusion Sampling (HBDS), which infers shared parameters and group-specific latent states using a single pretrained model, with the hierarchy specified only at sampling time. Together, these methods support variable observation sets and groupings without retraining. To address misspecification, we introduce path-regularized fine-tuning that adapts the learned likelihood to observations and transfers corrections to posterior inference. Using Girsanov's theorem, we quantify path divergence between pretrained and fine-tuned models across experimental designs and interpret it alongside predictive errors to distinguish candidate misspecification correction from unnecessary adaptation. We evaluate compositional sampling on exact-score Gaussian and Simple Likelihood, Complex Posterior benchmarks, HBDS with analytic and learned scores on a controlled hierarchical model, and fine-tuning and localization on a separate analytic model with known design-dependent discrepancy. Finally, we apply the framework to 940 measurements across four cell lines in a mechanistic Bone Morphogenetic Protein signaling model, where fine-tuning improves posterior-predictive accuracy relative to the pretrained model and shifts posterior marginals toward the least-squares reference while retaining spread.

---

### Fine-Tuning on Self-Generated and Reward-Weighted Data: Learning Dynamics, Convergence Rates, and Benefits of Off-Policyness

Authors: Zhiwei Wang, Yanxi Chen, Yaliang Li, Bolin Ding
Published: 2026-09-29
Categories: cs.LG, math.OC
arXiv: https://arxiv.org/abs/2609.36945v1

We study the learning dynamics of fine-tuning a policy model on self-generated and reward-weighted data, with particular focus on a generalized version of REINFORCE -- referred to as RE(S) -- that updates the rollout distribution once every $S \ge 1$ gradient steps. Prior work in bandits and reinforcement learning has developed rich theory for policy gradient methods, and on-policy sampling (i.e., a small $S$, ideally $1$) is often viewed as crucial to their success; yet in prominent application like post-training large language models, reward-guided self-training has proved to be effective even when the rollout distribution is updated infrequently, but theoretical understanding remains limited for the convergence properties of these off-policy methods. To bridge these gaps, we develop a unified theory for RE(S) that covers the full spectrum of $S \ge 1$: it can be interpreted as a stage-wise optimization process, where each stage takes $S$ gradient steps for minimizing the Kullback-Leibler distance to a fixed reward-weighted rollout distribution. For multi-arm bandits with softmax policies, our in-depth analysis and numerical experiments reveal three key findings: (1) for any fixed $S$, RE(S) enjoys global convergence to the optimal policy as the number of rollout distribution updates $B = \lfloor T / S \rfloor \rightarrow \infty$, where $T$ denotes the number of gradient steps; (2) we prove tight two-sided bounds showing that the suboptimality gap of RE(S) achieves an asymptotic $Θ(1 / T)$ convergence rate, while $S$ only affects the length of a burn-in phase; (3) when initialized at a weak policy with a small optimal-action probability, RE(1) gets trapped around suboptimal policies for a long period, whereas RE(S) with a suitable $S$ avoids the detour and achieves significantly faster convergence to the global optimum, highlighting the benefits of off-policyness in this case.

---

### Dual-Channel Robust Group-Relative Policy Optimization via Advantage and Sequence-Weight Estimation

Authors: Zhongyi Li, Wan Tian, Xiang Xu, Yutian Xiao, Yikun Ban, Yijie Peng, Fuzhen Zhuang
Published: 2026-09-29
Categories: cs.AI
arXiv: https://arxiv.org/abs/2609.36944v1

Group-relative policy optimization relies on reward-derived advantages and sequence-level likelihood weights, both of which can be sensitive to localized outliers. Extreme rewards can collapse the contrast among clean responses after group normalization, while token-level log-ratio perturbations can alter sequence weights and clipping decisions. We introduce RoVR-GSPO, a dual-channel robust optimizer that addresses these failure modes separately. Its reward channel combines robust reference estimation with bounded residual credit, while its ratio channel uses differentiable SoftRoVR aggregation to construct robust sequence weights. We provide stability and efficiency analyses for both channels. Experiments on mathematical reasoning, long-context summarization, and tool-call annotation show consistent improvements over GSPO, while controlled perturbation studies demonstrate stronger robustness to reward contamination and token-ratio anomalies.

---

### Safe-by-Design Learning via Energy-based Neural Networks

Authors: Simone Betteti, Morteza Lahijanian, Luca Laurenti
Published: 2026-09-29
Categories: cs.LG, cs.AI, eess.SY
arXiv: https://arxiv.org/abs/2609.36942v1

Learning neural-network models of dynamical systems with safety guarantees is a fundamental requirement for their deployment in safety-critical settings. Safety is commonly established by proving the invariance of a desired subset in state-space, ensuring that every trajectory initialized in this subset remains confined to it for all time under admissible inputs. Existing frameworks, however, either rely on computationally expensive post-hoc verification or employ safety-enforcing mechanisms without formal correctness guarantees. In this paper, we introduce a novel neural architecture grounded in energy-based modern Hopfield networks to guarantee safety-by-design while retaining sufficient expressiveness to model complex nonlinear dynamics. Specifically, we integrate modern Hopfield networks with a port-Hamiltonian neural ODE, enabling by design the construction of barrier functions yielding explicit admissible-input sets and quantitative robustness radii. Across several benchmarks, including an 12-dimensional nanodrone model, our framework achieves state-of-the-art performance while producing certified invariant sets that are more robust to external solicitations than comparable existing approaches.

---

### DispFlow-GS: Displacement Flow Supervision with Motion Disentangling for Monocular Deformable 3D Gaussian Splatting

Authors: Thai Duy Nguyen, Haitian Zhang, Addison Lin Wang
Published: 2026-09-29
Categories: cs.CV
arXiv: https://arxiv.org/abs/2609.36940v1

Accurate dynamic scene reconstruction is important for robotic perception, where temporally consistent representations of dynamic environments are essential. Deformable 3D Gaussian Splatting (3DGS) models dynamic scenes through deformation fields, and recent methods incorporate motion supervision by aligning rendered Gaussian flow with optical flow. However, we find that such Gaussian-flow-based supervision provides only limited improvements in motion modeling. We identify a fundamental limitation of this supervision paradigm, namely a domain gap between rendered Gaussian flow and optical flow. To address this limitation, we propose a motion supervision framework built on Displacement Flow, which splats per-Gaussian 3D displacements onto the image plane to provide direct and stable optimization signals. We further disentangle scene motion from camera motion via intermediate-view rendering, enabling more reliable motion priors and targeted constraints on deformation and geometry. We also observe a discrepancy between motion fidelity and image-based evaluation, where improved motion awareness does not necessarily translate into better rendered image quality or higher image-based metric scores. Motivated by this mismatch, we introduce Deformation-Rendering Consistency (DRC), a motion-aware metric that measures the alignment between predicted deformation and rendering improvement. Experiments on dynamic scene benchmarks show substantial improvements in motion localization and motion--rendering consistency, reaching up to 39% and 6%, respectively, while image-based metrics change by only about 0.1%. These results confirm the observed mismatch between motion fidelity and image-based evaluation, demonstrating the significance of DRC for motion-aware evaluation.

---

### SCA: Spatial Credit Assignment for Reinforcement Learning of GUI Agents

Authors: Shengtian Yang, Ziyu Xiong, Kaibing Yang, Guangfeng Cai, Yewen Li, Peng Jiang, Gai Kun, Qingpeng Cai, Lei Feng
Published: 2026-09-29
Categories: cs.AI
arXiv: https://arxiv.org/abs/2609.36939v1

GUI agents automate tasks on digital devices by grounding language instructions in visual interfaces. Existing group-relative reinforcement learning improves GUI action prediction by comparing the rewards of multiple responses sampled from the same GUI state. However, binary evaluation treats spatially different failed clicks as identical and provides no relative signal when all sampled clicks fail. To address these limitations, we propose Spatial Credit Assignment (SCA), which uses the screen coordinates of sampled clicks to refine group-relative credit. Specifically, SCA predicts each held-out response's reward from the other responses in groups containing both successes and failures, then uses the prediction residual to adjust credit. When all sampled clicks fail, SCA instead orders them by distance to the annotated target. These spatial references are used only to construct the training update; the deployed policy remains unchanged. We evaluate whether this correction improves the policy update itself by comparing its error and directional alignment with the exact return gradient in a controlled synthetic study. Across GUI grounding and offline action-prediction benchmarks, SCA improves grounding across professional domains and achieves the strongest results among reinforcement-fine-tuned models on most action-prediction metrics, with consistent gains across the reported GUI suites.

---

### WeLike2Party! In-Context Motion Transfer for Multi-Human Image Animation

Authors: Sangeyl Lee, Seunghyun Shin, Seungho Park, Wooseok Jeon, Hae-Gon Jeon
Published: 2026-09-29
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2609.36937v1

Human image animation aims to transfer motion from a driving video to subjects in a reference image. Despite remarkable progress in video generation, achieving high-fidelity animation of multiple interacting subjects remains a challenge. Many existing approaches rely on explicit motion representations such as 2D skeletons or parametric body meshes and struggle to preserve identity-motion binding under inter-person occlusion. To address this limitation, we propose WeLike2Party, a multi-human animation framework built on direct in-context video conditioning without explicit pose or mesh extraction at inference. We further introduce Reference Asymmetric RoPE Conditioning to preserve fine-grained appearance details, and Identity Binding Supervision to associate each reference identity with its intended motion trajectory. To support cross-identity training, we construct MotionTwin, a large-scale synthetic dataset comprising 14.4K cross-identity video pairs with shared subject and camera motions, totaling 84.3 hours of photorealistic video. We additionally present MotionTwin-Bench, a cross-identity benchmark specifically designed to evaluate subject-level visual fidelity and identity-motion binding. Extensive experiments on MotionTwin-Bench and real-world videos demonstrate that WeLike2Party outperforms recent state-of-the-art methods in subject-level visual fidelity, identity-motion binding, and overall perceptual quality, particularly in multi-person interactions with substantial occlusion.

---

### CoEM: Empowering Long-Context Reasoning with Commit-on-Evidence Memory

Authors: Jingguang Li, Yebo Wu, Zuyi Guo, Kailang Ma, Xianjie Dai, Han Zheng, Benwang Chen, Li Li, Can Rong, Heye Huang
Published: 2026-09-29
Categories: cs.AI, cs.CL
arXiv: https://arxiv.org/abs/2609.36935v1

Long-context reasoning is essential for complex and long-horizon tasks, yet the performance of large language models (LLMs) degrades as context length increases. Recent approaches address this by processing input chunk by chunk while maintaining a bounded textual memory in model context. However, premature information compression can discard critical details essential for subsequent reasoning. In this paper, we introduce Commit-on-Evidence Memory (CoEM), which learns when to convert source evidence into compact memory facts. Specifically, under a fixed context-memory budget, CoEM preserves potentially useful source excerpts verbatim in a pending set, allowing subsequent context to clarify their relevance before irreversible compression. As new context arrives, a learned policy revisits each pending excerpt and decides whether to promote it to the committed memory, retain it for further consideration, or discard it. A frozen verifier ensures proposed facts are accepted only if supported by retained excerpts and current context. To further guide effective memory management, we train this policy using reinforcement learning by combining fine-grained, step-level evidence rewards with final answer rewards. Extensive experiments demonstrate that CoEM consistently improves long-context reasoning. When evaluated on 6,400 documents long-context input, CoEM outperforms the strongest memory baseline by 10.4-11.4 F1 points on Qwen3.5-9B. Code repository: https://github.com/benmagnifico/CoEM.

---

### VLALight: A Vision-Language-Action Model for Traffic Signal Control

Authors: Pan Zhang, Siqi Lai, Kemu Dong, Hao Liu
Published: 2026-09-29
Categories: cs.AI
arXiv: https://arxiv.org/abs/2609.36934v1

Traffic signal control (TSC) is essential for improving urban mobility and reducing congestion. Although roadside cameras are widely deployed at signalized intersections and provide rich visual observations of evolving traffic, existing TSC methods typically rely on manually engineered traffic states or separate perception modules, creating a gap between physical observations and control decisions. We present VLALight, the first vision-language-action (VLA) model for end-to-end traffic signal control from multi-view roadside videos. VLALight directly maps visual observations to coordinated signal actions through multi-target spatiotemporal traffic reasoning and topology-aware cooperative perception across intersections. To establish this capability, we develop a two-stage supervised cold-start training strategy for visual traffic understanding and signal decision-making, followed by cooperative agentic reinforcement learning that jointly optimizes local control and network-wide traffic efficiency. Furthermore, VLALight introduces adaptive fast and slow reasoning modes, enabling the policy to allocate deeper reasoning only when additional deliberation provides sufficient control benefits. Through balanced mode-aware rollouts and relative advantage optimization, VLALight learns to trade off decision quality and inference cost. Extensive experiments on seven real-world traffic-flow datasets across three urban networks demonstrate that VLALight consistently outperforms transportation-based, RL-based, and LLM/VLM-based baselines. Ablation studies validate the effectiveness of cooperative perception, network-level optimization, and adaptive reasoning. These results demonstrate the potential of VLA models for real-world physical traffic control. Our project is available at https://github.com/usail-hkust/VLALight.git.

---

### Learn from the Gap: Differential-Aware Advantage Pruning with Adaptive Rollout Sampling for GRPO

Authors: Jiahua Yang, Zhiwei Yang, Xianpeng Zhang, Dongyu Chen, Xing Chen, Tianhuang Su, Haonan Lu, Quanlong Guan, Kai Tang, Chuangchuang Wang
Published: 2026-09-29
Categories: cs.AI
arXiv: https://arxiv.org/abs/2609.36932v1

Recently, Group Relative Policy Optimization (GRPO) and its variants have been developed for policy optimization and demonstrated notable performance gains. However, these methods usually incur substantial computational overhead due to per-question multi-rollout sampling and repeated per-token probability evaluation across rollouts. Furthermore, low-information or highly homogeneous trajectories can degrade downstream learning signal efficiency, hindering model optimization and limiting final performance. To address these issues, we propose FastRL, a novel plug-and-play reinforcement learning framework that simultaneously improves training efficiency and the effectiveness of policy learning. Specifically, 1) We introduce an advantage-aware pruning strategy to selectively preserve high-advantage trajectories while maximizing inter-trajectory gradient diversity. 2) Then, we design an adaptive rollout sampling mechanism to dynamically adjust the sampling scale across different training stages based on historical pruning distributions, balancing exploration adequacy and computational efficiency. Experiments demonstrate that FastRL can be seamlessly integrated into GRPO, DAPO, and GSPO variants, achieving an average 2.07$\times$ training speedup on Geometry3K and GeoQA8K-R1V, along with an approximately 1.64\% improvement in average accuracy on visual reasoning benchmarks. Source codes will be available at https://github.com/Nicozwy/FastRL.

---

### Dating the Model: Hidden Dates in System Prompts Affect LLM Evaluation

Authors: Mario Sanz-Guerrero, Minh Duc Bui, Manuel Mager, Katharina von der Wense
Published: 2026-09-29
Categories: cs.CL
arXiv: https://arxiv.org/abs/2609.36931v1

Reproducibility is essential for scientific research, yet prior work shows that LLM outputs vary with hardware and batching. We identify an overlooked factor: the hidden injection of the current date into system prompts, which users cannot control and which changes every day. Across 9 recent LLMs and 6 datasets spanning multiple-choice QA (MCQA), math reasoning, code generation, and machine translation, performance varies solely with the current date, with deltas of up to 6% on MCQA, 14% on math reasoning, 7% on code generation, and 2.84 BLEU on machine translation. Model rankings also shift, affecting leaderboards. This date effect exceeds other sources of non-determinism, such as batch size and numerical precision. Standard prompting techniques -- chain-of-thought and few-shot prompting -- do not reduce the sensitivity; chain-of-thought even amplifies it. Our findings underscore the need for careful evaluation protocols to ensure reproducibility and fair comparisons in LLM research.

---

### SFE-VGGT: Source-Free VGGT Distillation for Event-Based Monocular Depth Estimation

Authors: Thai Duy Nguyen, Addison Lin Wang
Published: 2026-09-29
Categories: cs.CV
arXiv: https://arxiv.org/abs/2609.36929v1

Recent event-based depth estimation methods successfully transfer geometric priors from vision foundation models via cross-modal distillation. However, their reliance on synchronized RGB-event pairs or depth annotations during training severely restricts practical deployment. To overcome this bottleneck, we propose SFE-VGGT, a novel source-free framework that distills the geometric priors of VGGT to the event domain without any paired RGB observations. Our core idea is to reconstruct surrogate frames directly from the target event stream to act as a frozen geometric teacher, entirely eliminating the need for genuine source RGB data. Crucially, as these surrogate frames inherently yield imperfect and spatially varying supervision, directly distilling from them propagates artifacts. To resolve this, we introduce a novel reliability-aware distillation strategy. This includes Density-Aware Feature Distillation to emphasize informative event regions, and Confidence-Weighted Depth Distillation to dynamically regulate supervision based on relative teacher-student prediction confidence. Meanwhile, we propose a Cross-Frame Relational Consistency loss that enforces temporal geometric stability using reliable inter-frame correspondences, bypassing the need for temporally consistent teacher's depth. Extensive experiments demonstrate that, despite source-free, our SFE-VGGT closely matches the accuracy of RGB-dependent baselines under standard conditions and significantly surpasses them in challenging nighttime scenarios. Across MVSEC nighttime sequences, SFE-VGGT reduces the average 10 m depth error by 15.3% compared with EventVGGT. Moreover, our method exhibits robust zero-shot generalization across real-world datasets, proving that highly effective geometric priors can be transferred to event cameras using strictly source-free supervision.

---

### Neuro-Symbolic Computer Use: Learning Reusable Policies for Reliable and Efficient Execution

Authors: Hyewon Suh, Thanh Minh Nguyen, Chih-Lun Lee, Darrow Hartman, Lizhao Liu, Xin Eric Wang, Ang Li, Jiachen Yang
Published: 2026-09-29
Categories: cs.AI
arXiv: https://arxiv.org/abs/2609.36927v1

Many computer tasks recur: the same workflow runs many times, with new inputs and from different starting states. Current computer-use agents re-plan every step of every run, which makes them costly and unreliable on such tasks. We introduce neuro-symbolic computer use, in which a recurring workflow is executed by a learned policy rather than re-derived by an agent on each run. The policy fixes the decisions that are stable across runs (ordering, variables, loops, and branches) in executable code, and delegates observation-dependent decisions, such as grounding and state checks, to neural models. We learn these policies with neuro-symbolic policy iteration: starting from one agent trajectory, it executes the policy, diagnoses failures with task-completion and step-level judges, and revises the code with a coding model informed by an agent's continuation from the point of failure, without access to the benchmark evaluator. Iterating on generated parameter and initial-state variants makes the policy reusable, and a pre-action verifier guards each state-mutating step at deployment. On OSWorld-Verified and ScienceBoard, the learned policies achieve the highest Pass^3 of all methods in all four settings, 3.6-15.8 points above the base agent, while cutting per-run cost by 15-217$\times$ and latency by 3.4-5.1$\times$. On OSWorld-Verified, policies built only on variants transfer to the held-out original tasks, exceeding AutoRPA by 8.6-17.5 points in Pass^3.

---

### State Transport Routing for Short-horizon Adaptation in Multi-horizon Photovoltaic Forecasting

Authors: Xu Yuqing, Zhou Liguo, Sun Ze, Yu Lei, Jiang Mingming
Published: 2026-09-29
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2609.36926v1

Recent power measurements provide valuable information for photovoltaic(PV) power forecasting, but directly extrapolating short-term trends can introduce substantial errors over longer forecast horizons. To address this challenge, we propose state transport routing (STR), a lightweight adapter that refines the predictions of a frozen forecasting model. STR combines the original forecast with two complementary trajectories derived from the latest measured power level and its recent trend. A horizon-conditioned router adjusts their contributions over the first 120 min, while leaving subsequent predictions unchanged. Experiments on four public PV datasets show that STR consistently outperforms a parameter-matched residual adapter. On PVDAQ, the same approach improves five neural forecasting backbones, reducing all-horizon normalized mean absolute error by 0.0201-0.2364 percentage points, with paired 95% confidence intervals excluding zero. No reliable improvement is observed for LightGBM. These findings demonstrate the potential of structured state adaptation to improve short-term forecasting across different neural architectures without retraining the underlying models or altering their longer-horizon predictions.

---

### Track-and-Complete: Learning Humanoid Skills from a Single Failed Human Video

Authors: Sarmad Idrees, Jongeun Choi
Published: 2026-09-29
Categories: cs.RO, cs.CV
arXiv: https://arxiv.org/abs/2609.36924v1

Learning humanoid skills from videos typically requires a successful human demonstration, which often demands custom data collection. Although failures have traditionally been treated only as negative examples in robot learning, they can still reveal a usable trajectory prefix before the task fails, as well as the intended outcome. To leverage this information from a failed-attempt video, we propose TRACC, a pipeline that imitates the useful portion of the motion trajectory and then completes the task based on the inferred task outcome. The usable motion prefix serves as prior knowledge until the failure occurs, after which the task-completion reward guides the policy to learn the intended task goal without requiring a successful task trajectory. We evaluate our method on six in-the-wild failed human tasks from the Oops! dataset. Our experimental results demonstrate the effectiveness of the proposed approach for learning from failed attempts when no successful demonstration is available. Thus, these findings establish failed human videos as a viable source of supervision for humanoid skill learning.

---

### PrecogUI: Proactive GUI Agents via Pre-cognitive Simulation and Experience Retrieval

Authors: Bin Kang, Jiarui Ouyang, Li Jiang, Bin Chen, Zhuotao Tian
Published: 2026-09-29
Categories: cs.AI
arXiv: https://arxiv.org/abs/2609.36923v1

Existing reactive Graphical User Interface (GUI) agents often fail in long-horizon, dynamic scenarios, where unexpected disturbances trigger attention-diverting and cascading failures. To address this, we propose PrecogUI, a pre-cognitive architecture that shifts the paradigm from reactive execution to proactive decision-making. Specifically, we design a Proactive Experience Pool (PEP), which caches recurring anomaly and success patterns as "state-action-result" tuples in a dual-memory repository. Furthermore, we introduce a Proactive Simulation Executor (PSE) that learns to forecast the next symbolic UI layout given a candidate action, enabling early anomaly avoidance and ranking candidate actions by predicted reliability. Finally, a Pre-cognitive Execution Controller (PEC) fuses these priors and predictions, prioritizes handling of foreseen anomalies, and ensures execution robustness through a closed-loop error correction mechanism. For robust evaluation, we develop AutoTraj, an automatic data-generation engine, to construct InterfereBench, a benchmark for long-horizon tasks with strong disturbances. Experiments demonstrate that PrecogUI surpasses state-of-the-art methods on InterfereBench while maintaining competitive performance on public benchmarks. The code will be publicly available.

---

### Benchmarking Automatic Speech Recognition Tools for Iberian Languages

Authors: Fernando López, Pablo Gómez, David Solans, Paulo Villegas, Jordi Luque
Published: 2026-09-29
Categories: cs.CL
arXiv: https://arxiv.org/abs/2609.36920v1

Comprehensive evaluations of automatic speech recognition (ASR) for Iberian languages remain limited, and low-resource languages, biases, and efficiency trade-offs are underexplored. We benchmark eleven systems, ten open-weight models and one commercial API, across five Iberian languages (Basque, Catalan, Galician, Portuguese, Spanish), with German and Turkish as controls. Evaluation uses an 85-hour dataset covering read speech, broadcast media, and audiobooks, assessing accuracy and efficiency via word error rate (WER) and real-time factors (RTF/RTFx). Results show no single model dominates: accuracy, efficiency, and language coverage present clear trade-offs. Low-resource languages, especially Basque, degrade significantly, highlighting the role of training coverage. We observe consistent sex disparities across most systems, highlighting fairness challenges in multilingual ASR. Overall, the benchmark provides practical guidance for real-world model selection.

---

### Seg3DParts: Segmentation-Grounded Controllable Part-Level 3D Generation

Authors: Jiantao Lin, Meixi Chen, Yingjie Xu, Chenbo Fu, Leyi Wu, Hao Chen, Yinchuan Li, Ying-Cong Chen
Published: 2026-09-29
Categories: cs.CV
arXiv: https://arxiv.org/abs/2609.36918v1

Part-level 3D assets are essential for editing, reassembly, and interaction, yet recovering such structure from a single image remains challenging due to occlusion, ambiguous boundaries, and the need for coherent multi-part reasoning. Existing approaches struggle to achieve both controllable part-level generation and coherent multi-part structure, as part identity and spatial allocation are typically inferred implicitly. We present Seg3DParts, a segmentation-grounded framework for controllable part-level 3D generation from a single image. By treating segmentation as an explicit grounding signal, our method defines part identity during generation, enabling each component to be anchored to a corresponding image region. To ensure coherent assemblies, we introduce structured cross-part interaction that allows components to exchange global context throughout the generative process. As a result, Seg3DParts directly generates well-aligned part meshes in a shared canonical space without post-hoc alignment, supporting flexible and controllable decomposition. We further introduce PartObjectNet, a large-scale dataset with over 200K objects and 1M annotated parts. Experiments demonstrate that Seg3DParts achieves superior geometry quality, cross-part coherence, and part-level controllability over existing methods.

---