# Research Papers — 2026-10 · bundle page 61 of 156 > Full-text markdown of papers #3001–3050 of 7799 published in 2026-10, newest first. AI-readable mirrors of the full 7799-paper 2026-10 archive (50 papers per file). Canonical page: https://aipod.dev/papers/archive/2026-10/page/61 [← Previous bundle](https://aipod.dev/papers/archive/2026-10/page/60.md) [Next bundle →](https://aipod.dev/papers/archive/2026-10/page/62.md) ### Depth-progressive 3D seismic velocity model building via 2D generative diffusion models Authors: Shijun Cheng, Randy Harsuko, Mohammad H. Taufik, Tariq Alkhalifah Published: 2026-10-10 Categories: physics.geo-ph, cs.AI arXiv: https://arxiv.org/abs/2610.14812v1 Three-dimensional (3D) velocity model building (VMB) remains computationally challenging with conventional approaches and native 3D neural networks, largely due to the prohibitive memory and training cost of full 3D formulations. Therefore, we introduce a depth-progressive framework that achieves 3D VMB using 2D generative diffusion models through a depth-as-channel formulation, in which the depth dimension of each 3D velocity patch is mapped to the channel dimension of a 2D U-Net. This allows the model to synthesize overlapping depth patches sequentially from shallow to deep while conditioning the prediction of the deeper patch on the shallower one, as well as on any available geophysical constraints, including well logs and structural attributes. During training, paired shallow-deep patch pairs with randomized overlaps are used to teach the network how deeper velocity structures relate to shallower information, and a classifier-free guidance strategy enables the model to handle missing constraints gracefully. At inference, starting from a shallow velocity prior, the model recursively generates deeper patches in a top-down manner, and a depth-attenuated Gaussian blending scheme merges overlapping predictions into spatially coherent 3D volumes while accounting for the progressive accumulation of uncertainty at greater depths. Numerical experiments on both in-distribution benchmark models, including Overthrust, SEAM Arid, and SEG/EAGE, and the out-of-distribution Marmousi model demonstrate that the method effectively propagates shallow velocity priors downward and produces high-quality 3D velocity volumes across a range of geological settings and conditioning scenarios. Compared to a 3D benchmark of equivalent architecture, the proposed method significantly reduces parameter count, training cost, and inference cost, while maintaining superior VMB accuracy. --- ### KedgeFlow: A Programmable Control Plane for Enterprise AI Agents Authors: Jun He, Deying Yu Published: 2026-10-10 Categories: cs.DC, cs.CR, cs.LG arXiv: https://arxiv.org/abs/2610.14810v1 Enterprise AI agents can execute altered versions of approved actions, concurrently reuse the same permission, or retry operations whose responses were lost. These failures arise when approval and resource execution lack a shared contract. KedgeFlow specifies language-independent contracts for trusted controllers to authorize state-changing operations and privileged gateways to enforce that authority independently of agent reasoning. Inspired by OpenFlow, it connects policy checks supported by evidence, signed permission for an exact action, a shared allowance that can be consumed only once, target-side conditions for applying the change, and retained outcome records. Under trusted issuance of exact-action authority, two conditional properties establish at most one successful claim per allowance and link every governed dispatch to a matching grant, claim, and subsequent dispatch authorization, even when the agent is compromised. A reference implementation exposes this execution path to agent systems through the Model Context Protocol (MCP). In the single-resource evaluation, 6,560 concurrent ledger claim attempts yielded exactly one successful claim for each of 82 allowances. Boundary tests rejected unauthorized and altered requests before consumption. A 32-call gateway race produced one committed mock effect, and a retry after a lost response produced no additional dispatch. Deliberately permissive policy nevertheless authorized an undesirable effect, exposing the limits of execution authorization. Evaluation uses three members of the etcd coordinated key-value store on one host, synthetic authentication and authorization inputs, and a mock target that rejects stale requests. Distributed reservations, enforcement of every production access route, native crash recovery, exactly-once effects at real targets, and interoperability between independent implementations remain untested. --- ### Groups Don't Move as One: A Bregman Clustering View of Linearized Group Influence Authors: Hyunho Choi, Jaeseung Heo, Youngbin Choi, MoonJeong Park, Dongwoo Kim Published: 2026-10-10 Categories: cs.LG arXiv: https://arxiv.org/abs/2610.14809v1 Influence functions provide an efficient alternative to retraining for assessing the effects of removing or reweighting training data. For groups, however, the standard first-order estimate reduces individual contributions to a sum that can conceal substantial differences among members, including opposing effects that cancel. We establish a Bregman centroid formulation of the linearized group influence objective, interpreting this aggregate as a single representative and using Bregman information (BI) to quantify the variation it omits. Extending this formulation to multiple clusters reveals distinct contributions whose sum recovers the original estimate. We further propose an estimator that jointly optimizes the weights of these cluster contributions, rather than simply summing them. Experiments uncover cluster-specific influence patterns and a positive association between target-specific BI and approximation error. The proposed estimator achieves higher $R^2$ than standard group influence in predicting target changes under the nonlinear proximal Bregman response function. Simplified clustering and correction variants reduce computational cost. The centroid view thus explains what additive group influence leaves out and provides a basis for using that information to improve estimation. --- ### ASIWorld: Benchmarking Autonomous Agents in a Persistent Survival Sandbox Authors: Jianfeng Lin, Runze Wang, Haorong Hong, Yuxuan Xie, Zeno Zeng, Zhongyuan Liu Published: 2026-10-10 Categories: cs.AI arXiv: https://arxiv.org/abs/2610.14807v1 Evaluating autonomous agents requires examining whether they can translate goals into coherent action across extended interactions. In persistent environments, this requires coordinating interdependent actions under incomplete observations, accounting for delayed effects, and adapting as earlier decisions shape subsequent opportunities. We introduce ASIWorld, a survival sandbox that combines persistent world and character state with character-scoped observations and a programmable asynchronous interface to study sustained goal execution. Agents pursue interdependent production and exploration tasks through direct actions or agent-authored programs that continue running between model calls. The evaluation assesses task success, the attainment of development achievements, and survival costs, complemented by an analysis of execution behavior. In an exploratory study of nine agent configurations using a common harness, eight complete the production task and six complete both tasks. Analyses of execution trajectories and performance across difficulty levels reveal different preparation, combat, and recovery behaviors among configurations with similar task outcomes. Incomplete preparation is associated with stalled progression, while configurations that complete the tasks differ in survival costs and the fulfillment of additional achievement criteria. These findings illustrate the value of assessing sustained goal execution through both task outcomes and the behaviors and costs associated with achieving them. --- ### YOUDrive: Driving Style as a Composable Direction for End-to-End Autonomous Driving Authors: Guantian Zheng, Lidong Yu, Tianyu Gao, Jiaxing Chen, Jiashu Li, Chen Lv Published: 2026-10-10 Categories: cs.CV, cs.AI, cs.RO arXiv: https://arxiv.org/abs/2610.14802v1 End-to-end driving policies can achieve strong task performance while offering limited control over their driving style. Driving style describes how a vehicle accelerates, brakes, and maintains distance. Policies trained without explicit preference conditioning lack a direct mechanism for adjusting these behaviors at inference. We propose YOUDrive, a vision--language--action planner for continuous and composable driving-style control. Each teacher's behavior is distilled into a task vector, a learned parameter update that induces a style direction in the planner's representation. These directions can be scaled and combined at inference without retraining the backbone. Conditioned on these directions, a flow-matching decoder models a multimodal trajectory distribution and generates kinematically feasible motion alternatives. We introduce the Style Alignment Score (SAS) to describe human-referenced kinematic behavior after safety and task gating. On NAVSIM navtest, a single coefficient monotonically controls assertiveness across six teacher-derived personas, while PDMS remains above 0.90 across the tested individual and blended configurations, reaching 0.954. Further analyses reveal systematic policy-specific kinematics after controlling for scene context, whereas the evaluated discrete style labels provide limited additional explanatory power. Code, checkpoints, and the metric will be released. --- ### CESNET-TimeSeries26: A Multi-Year Network Traffic Time Series Dataset for Annual Seasonality, Concept Drift, and Long-Term Anomaly Detection Authors: Josef Koumar, Karel Hynek, Kamil Jerabek, Pavel Siska Published: 2026-10-10 Categories: cs.AI, cs.LG, cs.NI arXiv: https://arxiv.org/abs/2610.14799v1 Forecast-based models are fundamental to network traffic management and anomaly detection. However, it is essential to evaluate these forecasting models over extended periods of time in order to assess their adaptability to unexpected phenomena and shifts in underlying patterns. This kind of long-term evaluation requires multi-year data to account for yearly seasonality and behavioral drift. Existing public datasets are too short to support such long-horizon forecasting tasks. We introduce CESNET-TimeSeries26, comprising 995 days of continuous traffic time series from the CESNET3 network, together with a single subnet spanning approximately 4.5 years. The dataset captures 278,466 IP addresses, 285 institutions, and 566 institutional subnets at three aggregation intervals. The dataset provides a definitive longitudinal benchmark for evaluating forecasting models in both operational management and security monitoring. We demonstrate that annual seasonality is the dominant periodic component. Moreover, we show that a forecasting model can exploit the annual component only when its training window covers at least one complete annual cycle, which no sub-year dataset provides. --- ### GREAT-ESDD: Graph Edge Attention Network for Effectively Steered Delivery Decisions Authors: Attila Lischka, Ze Zhou, Xuegang, Ban, Balázs Kulcsár Published: 2026-10-10 Categories: cs.AI arXiv: https://arxiv.org/abs/2610.14792v1 Online, same-day delivery requires operators to constantly update their routes, and the routing cost for deliveries increases with the number of customers. Out-of-home (OOH) delivery options can help companies handle increasing customer numbers while keeping delivery costs low. We develop a dynamic pricing-and-routing framework for this online ordering scenario with home and OOH delivery options. First, we develop a neural solver for the capacitated vehicle routing problem (CVRP) based on the Graph-Edge-Attention-Network (GREAT). We train it to solve CVRP with varying instance sizes and starting conditions, allowing it to reoptimize partially executed routes under changing conditions. It is fast while achieving solution quality competitive with traditional baselines. Our solver processes several CVRP instances in parallel in batches and we use it to solve the candidate routing problems that arise from an incoming customer's choice of home or OOH delivery. This provides an estimate of the delivery cost increase per option, which we use to set dynamic prices. Further, it directly provides the updated route once the customer decides. We trade off delivery fees against routing costs and steer customers toward OOH delivery when home delivery would be costly. However, effective steering requires us to know the customer preferences. To learn those online, we use Thompson sampling with a Laplace-approximated posterior. We call the resulting framework GREAT-ESDD (GREAT for Effectively Steered Delivery Decisions). Compared to a traditional routing solver, it increases profit by around 11% under a similar pricing decision time budget. Moreover, we show that we can estimate customer preferences successfully, outperforming a static pricing strategy after only 5 days of simulated operations. --- ### Counterfactual Regret for Reflective Agents with Lean-Verified Guarantees Authors: Zhongnian Li, Shi Xu, Mengting Xu, Peng Ying, Xinzheng Xu Published: 2026-10-10 Categories: cs.AI arXiv: https://arxiv.org/abs/2610.14790v1 Reflective language agents reuse feedback and memory to adapt across tasks without updating model weights. However, existing guarantees for policy selection and action generation do not by themselves establish when local reflection yields long-term performance guarantees for the complete agent under resource constraints. In this paper, we address this gap by deriving sufficient conditions for sublinear counterfactual regret from local generation and deployment guarantees, while retaining the native controller. Specifically, we establish a resource-sensitive transfer that bounds full-state deployment discrepancies along actual execution paths and preserves the own-history laws of both learner and comparator. Besides, local generation reliability yields a cumulative generation opportunity bound, linking policy acquisition to long-term performance through an explicit quality, coverage, and hard-budget tradeoff. To verify these guarantees, we formalize the conditional proof chain for all thirteen numbered theoretical results in Lean, including a native instance with repeated policy growth and imperfect validation. --- ### Training-Free BioNEL-RAG: Advancing Biomedical Entity Linking through Retrieval-Augmented Generation and Large Language Model Authors: Ziyi Chen, Hui Liu, Ziyang Ling, Ronald X. Xu, Mingzhai Sun Published: 2026-10-10 Categories: cs.CL arXiv: https://arxiv.org/abs/2610.14789v1 Biomedical Named Entity Linking (BioNEL) is the task of mapping mentions extracted from biomedical texts to canonical entities or concepts in the standardized knowledge base. Although existing methods have achieved remarkable performance in mapping accuracy, most rely on training with domain-specific datasets, which fundamentally limits their cross-domain adaptability. To address this limitation, we propose BioNEL-RAG, a novel training-free framework that combines retrieval-augmented generation (RAG) with large language models (LLMs). Within this framework, the RAG module focuses on retrieving candidate entities, while the large language model module plays the role of selecting the most appropriate candidate for entity normalization, which is based on the analysis of the mention's context, with the aid of its remarkable text comprehension capabilities. Experimental results demonstrate that BioNEL-RAG achieves performance comparable to existing methods without requiring training on subdomain datasets, whereas these methods necessitate fine-tuning on specific subdomains such as genes, diseases, or chemical compounds. --- ### Formalizing Forgetting: Theory of Forgetting for Online Continual Learning Authors: Renjie Cao, Yifei Cheng, Li Shen Published: 2026-10-10 Categories: cs.LG arXiv: https://arxiv.org/abs/2610.14787v1 Online continual learning (OCL) requires a learner to acquire a sequence of tasks in a single pass over the data, and catastrophic forgetting of earlier tasks remains its central challenge. Despite the many methods proposed to mitigate forgetting, the theoretical understanding of forgetting in OCL is still limited: existing analyses mostly focus on linear, overparameterized, or neural tangent kernel regimes, and it remains unclear which factors govern forgetting and how they affect it. To fill this gap, we analyze single-pass Shuffle SGD under strongly convex and general convex objectives and derive explicit upper bounds on the expected final forgetting. Further analysis of these bounds shows how the per-task data amount, task similarity, and task ordering enter them, and that their effects can differ across objective classes. Moreover, experiments with neural networks on Split-CIFAR-100 and PMNIST show that the trends suggested by our theory also appear in practice, indicating that these insights extend beyond the convex setting. --- ### LYNX: Analytical Program Synthesis via Grounding and Closed-Loop Refinement Authors: Ruichen Xia, Xiao Han, Dingyu Yang, Luwei Yang, Sai Wu, Dongxiang Zhang, Binhui Yao, Gang Chen Published: 2026-10-10 Categories: cs.AI, cs.DB arXiv: https://arxiv.org/abs/2610.14777v1 LLM-powered data analytics aims to translate natural language instructions into executable analytical programs over underlying data. However, existing approaches, whether treating LLMs as intent-to-code translators or as standalone analytical engines, often fail to reliably align domain semantics with complex schemas and heterogeneous metadata, resulting in brittle execution and semantically inconsistent analytical outputs in real-world workloads. We present LYNX, a multi-agent framework that formulates LLM-driven analytics as a schema- and knowledge-constrained analysis problem rather than a single-pass generation task. LYNX organizes analytical synthesis into a closed-loop workflow of planning, generation, verification, and refinement through three role-specialized agents that collaboratively enforce schema grounding, semantic consistency, and execution correctness. To address domain ambiguity and schema misalignment, LYNX integrates Semantic-Structural Distillation and Stage-Aware Allocation (SSD-SAA). A delta-guided iterative revision continuously diagnoses and repairs semantic and execution-level failures, substantially improving synthesis robustness. We evaluate LYNX on two NL-to-SQL benchmarks, BIRD-SQL Mini-Dev and the more challenging Spider 2.0-Lite. Experimental results demonstrate strong performance and efficiency, achieving 10.66% to 59.54% execution accuracy across lightweight and commercial models, while reducing token consumption by 70.53% compared with the prior most efficient method. Our framework displays high generalization over different backbones, from open-source, lightweight models to the latest commercial models. Additional case studies on quantitative finance analytics and complex multi-relational SQL workloads further demonstrate the framework's robustness, scalability, and practical applicability in real-world analytical scenarios. --- ### Training a Compact, Strong, and Realistic Text-to-Speech Model Authors: Runyuan Cai, Yu Lin, Yiming Wang, Xiaodong Zeng Published: 2026-10-10 Categories: cs.SD, cs.AI arXiv: https://arxiv.org/abs/2610.14776v1 Text-to-speech (TTS) systems are conventionally optimized for intelligibility and speaker similarity. However, neither objective captures the prosodic detail that distinguishes human performances from synthetic speech, leaving a realism gap that larger models and higher-quality data can narrow but not close. In the absence of an objective that directly targets this gap, realism remains a persistent challenge in TTS. We present Audio8-TTS, a compact 0.6B-parameter model for zero-shot voice cloning that achieves state-of-the-art performance in intelligibility and speaker similarity, with particularly strong results in perceptual naturalness. To align the training signal with human perception rather than signal-level fidelity, we construct a quality ladder from widely used open-source TTS systems and train a realism critic to assess generated speech against matched human recordings of the same sentences. An identity-erasure objective discourages the critic from relying on generator fingerprints. We train the model in three stages: multilingual supervised fine-tuning; group relative policy optimization (GRPO) with word error rate (WER) and speaker similarity (SIM) rewards; and GRPO guided by the realism critic. The final stage improves perceived prosodic naturalness and speaker similarity while preserving WER performance. --- ### ProgAgent: Learning Tool Orchestration through Programmatic Training Authors: Hao Li, Keqing He, Longtao Zheng, Junlong Li, Weihao Zeng, Qian Liu, Yuzhen Huang Published: 2026-10-10 Categories: cs.AI arXiv: https://arxiv.org/abs/2610.14774v1 Large language models (LLMs) are increasingly expected to orchestrate diverse tools over long-horizon tasks. Existing agents typically use Direct Tool Calling (DTC), where the model invokes tools sequentially, and tool responses accumulate in the model context, which can lead to context rot and orchestration breakdown. In contrast, Programmatic Tool Calling (PTC) invokes tools through executable code and stores intermediate results in runtime variables rather than appending them to the context. Prior work has explored PTC mainly as an inference-time optimization, but its benefits for agentic training remain unclear, and suitable verifiable training tasks are scarce. In this work, we introduce ProgAgent, a full PTC training stack, comprising a pipeline for synthesizing verifiable long-horizon tasks, a lightweight PTC runtime, and a two-stage training recipe. Using only 215 tasks instantiated at four environment scales, ProgAgent yields higher pass rates than the evaluated baselines of comparable size across four benchmarks. On LOCA-bench, ProgAgent-14B achieves a 30.7% pass rate at a 64k-token environment state, where these baselines score zero. Cross-harness experiments show that PTC-trained models outperform DTC-trained models under both evaluation interfaces, suggesting that the benefits of PTC training extend beyond the interface. --- ### How Should a Camera-First AI Agent Work? Authors: Xiaofan Wu, Xi Zeng, Miaoxia Chen, Peishan Chen, Shuyan Li, Chen Wang, Xiaohan Wang, Wanling Xu, Jiyun Yao, Hanyong Zhong, Jiahao Zhu Published: 2026-10-10 Categories: cs.AI, cs.HC, cs.LO arXiv: https://arxiv.org/abs/2610.14772v2 Camera-first agents must act on the visual world as it unfolds. We argue that the central obstacle for future camera-first agents is not recognition ability - multimodal models (vision-language and live/streaming alike) generalize well through language and remain the right tool for semantic reasoning - but time. Many visual targets are ephemeral: they appear, are query-relevant, and leave within a short window. Binding memory to the same end-to-end path as reasoning means that by the time a monolithic pipeline completes, the evidence is gone. We formalize this as Ephemeral-Target Interaction (ETI) and a Perceptual Deadline Contract separating a hard memory deadline from a soft reasoning deadline, grounded in the bounded, passive nature of sensory and working memory. We propose a two-deadline architecture: an always-on, bounded, non-semantic fast path; an active remembering mechanism that commits a moment to durable memory before it is overwritten (event-driven in v1); and a deferred deliberation layer (multimodal model + tools). We instantiate the contract on real MOT17 track windows, a consistency check, and calibrate the single-path baseline from measured frontier-model latency: even a short answer takes 1.2-9.3 s, and an agentic identify-and-search longer. Answers can be late; evidence cannot be missed. --- ### Prior-Fitted and Frozen Scorers for Multimodal Contrastive Learning Authors: Tillmann Rheude, Roland Eils, Benjamin Wild Published: 2026-10-10 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2610.14770v1 Multimodal contrastive learning typically relies on manually specified scoring functions, such as cosine similarity, to quantify the similarity between representations from different modalities. Several such functions have been recently proposed to extend contrastive learning beyond two modalities. In principle, however, any function that assigns high compatibility to matched samples and low compatibility to mismatched ones could serve as a scorer. Motivated by this broader view, we instantiate the scorer with a Transformer and find that even a randomly initialized model, when kept frozen, can outperform strong handcrafted multimodal contrastive baselines, whereas optimizing the scorer jointly with the encoders is less effective. Building on this observation, we introduce the Prior-Fitted Scorer (PFS): a Transformer scorer pre-trained on synthetic multimodal data generated from structural causal model (SCM) priors before transferring it to real-world data. The SCM exposes the scorer to diverse dependency structures, including shared, pairwise, and private information, providing a structured prior over higher-order cross-modal similarity. Across four real-world datasets, frozen Transformer scorers are competitive to contrastive baselines, and prior-fitting further improves the random initialization. More broadly, our results suggest that multimodal contrastive learning need not rely on manual similarity functions: Transformers can provide strong inductive biases, while causal pretraining can further shape these biases. --- ### Who Receives Respect? A Counterfactual Benchmark for Honorific Allocation Bias in Bengali Authors: Ajwad Abrar, Md Kamrul Hasan, Hasan Mahmud Published: 2026-10-10 Categories: cs.CL arXiv: https://arxiv.org/abs/2610.14769v1 Bengali has no grammatical gender, but honorificity is obligatory: every third-person reference must choose ordinary or respectful register. English has no such signal, so the system must choose a level of deference that the source never specified. We use this property to measure how large language models (LLMs) allocate respect in English--Bengali translation. Our evaluation set contains 2{,}016 sentences across 72 occupations. Each appears in seven sentence frames under four pronoun conditions. Masculine and feminine versions differ by one word, forming counterfactual pairs. We report a single Honorific Bias Score (HBS) for each model on a 0--100 scale, covering both the gender and occupation axes. A lower score means that respect is allocated more evenly. For the occupation analysis, we compute each occupation's honorific rate across the seven no-pronoun sentence frames. The standard deviation of these 72 occupation-level rates ranges from 17.2 to 35.5 percentage points across models, showing substantial variation in how honorifics are allocated across occupations. Occupation is associated with 3.2\% to 51.5\% of the variance in register choice, compared with at most 1.2\% for gender. Gender effects are much smaller overall: only two of the nine models show significant directional gender gaps, and they favor opposite genders. HBS ranges from 19.9 to 70.1 across the nine models, showing substantial variation in how evenly the models allocate respect. --- ### Last Secure Code Benchmark: Is the Code You Vibe Secure? Authors: Tianyi Wu, Chao Ge, Peng Chen, Fengji Zhang, Yongxing Wu, Yuchong Xie, Jian Li, Chaoyuan Peng, Kaikai Zhang, Dong Huang, Mingzhe Du, Yukai Huang et al. Published: 2026-10-10 Categories: cs.CR, cs.AI arXiv: https://arxiv.org/abs/2610.14766v1 AI coding agents now write much of the code that ships, yet whether that code is secure remains unclear. We introduce the Last Secure Code Benchmark, a comprehensive evaluation of secure code generation by coding agents. Built from 150 real vulnerabilities, its 450 tasks span seven widely used languages (Java, C, C++, JavaScript, Python, PHP, and Go), 24 CWEs including 21 of the 2025 CWE Top 25 and 8 of the Top 10 KEV weaknesses, more than any benchmark we compare with, and three vibe-coding scenarios: filling in a function, writing a file, and building a repository's source tree from scratch. Every task asks only for functionality, through a specification that never mentions security, and hidden functional and security tests grade the result. Across six frontier agents, 73.6% to 91.3% of tasks pass functionally, yet even the strongest, Claude Fable 5, produces code that is both functional and secure for only 25.1% of tasks, and the others for 16.4% to 24.7%. The best model only passes 12% of the vulnerability scenarios across all vibe coding setting, posing strong concern on the security of the code written by vibe coding. --- ### CUSE: Fine-Grained Multimodal Unlearning with Counterfactual Guidance and On-Policy Self-Distillation Authors: Chenlu Ding, Jiancan Wu, Yanchen Luo, Tianyu Zhang, Yancheng Yuan, Xiang Wang, Xiangnan He Published: 2026-10-10 Categories: cs.AI, cs.CL arXiv: https://arxiv.org/abs/2610.14763v1 Multimodal large language models (MLLMs) can reproduce sensitive, copyrighted, or unsafe information from their training data. Effective unlearning must remove target information while preserving non-target content, even within the same response. Existing methods typically apply uniform forgetting objectives to complete responses and rely on fixed, off-policy supervision, resulting in coarse token control and misalignment with the model's current behavior. We propose CUSE, a Counterfactual-guided Unlearning framework via on-policy Self-distillation for multimodal large language models. It first constructs counterfactual inputs by removing target-related evidence and uses the resulting changes in token probabilities to identify which tokens require stronger forgetting. It then distills token-level forgetting and retention targets on responses sampled from the current policy, with a frozen copy of the original model serving as the teacher model augmented with the corresponding forgetting- and retention-oriented privileged contexts. This design jointly determines what to forget and what to preserve. On MLLMU-Bench, CUSE reduces forget-set ROUGE-L by 33.6% while achieving 1.2% higher retain-set ROUGE-L than the strongest baseline. --- ### CrossMean: Model Merging by Pairing Merged and Expert Activations Authors: Moshe Eliasof Published: 2026-10-10 Categories: cs.LG arXiv: https://arxiv.org/abs/2610.14762v1 Model merging combines fine-tuned experts into one network without retraining, and closed-form methods compute each merged layer from statistics of its input activations. However, these statistics are measured on a single expert, while in the merged model, each layer receives activations from the merged layers that precede it. Recent work re-measures these statistics in the merged model but still forms them from a single network. We formulate closed-form methods as a linear system and show that they use the same matrix on both sides. We then introduce CrossMean, which replaces this matrix on one side with a bilinear statistic that pairs merged and expert activations. From the same statistics, we also derive a label-free criterion that indicates the task sets in which accounting for the merged activations improves the merge more. Empirically, task sets with a lower criterion value show a larger improvement. On the eight-task benchmark of each vision backbone that we tested and on BERT encoders, CrossMean improves on merging with statistics measured on the experts, and with all linear layers merged, it also improves on re-measuring these statistics in the merged model. These results suggest that merging benefits from pairing merged and expert activations. --- ### The Constitution Speaks Nheengatu: An Open MT System and Reproducible Corpus Pipeline for the Amazonian Língua Geral Authors: Fabrício Carraro, Rodolfo Joel Zevallos, Rodrigo Gonçalves de Souza, Caio Henrique Faustino da Silva, John E. Ortega Published: 2026-10-10 Categories: cs.CL arXiv: https://arxiv.org/abs/2610.14760v1 Nheengatu (yrl), the Amazonian Língua Geral, has 6--10 thousand speakers and is co-official in São Gabriel da Cachoeira (Brazil), yet remains poorly served by evaluated Portuguese--Nheengatu MT resources. We present the largest parallel corpus assembled for the language (14.5k Portuguese--Nheengatu pairs), anchored by the 2023 official translation of the Brazilian Federal Constitution; the first MT evaluation set built from transcribed spontaneous speech; and, to our knowledge, the first open-weight MT models dedicated to bidirectional Portuguese--Nheengatu translation. Three findings organize the paper. First, translation into Nheengatu on held-out constitutional articles plateaus on development data regardless of training length or model size, and much of the remaining error sits in conventional legal terms that the written sources do not attest consistently; a terminology success rate over a frozen lexicon shows that glossary mechanisms at training time and lexical constraints at decoding each raise term realization, the latter at a cost in repetitive output that chrF++ barely registers. Second, a corpus expansion that adds transcribed conversation raises speech chrF++ by 24--27 points while leaving written domains unchanged, a register gap that written-only benchmarks cannot see. Third, a larger NLLB model helps translation into Portuguese but not into Nheengatu on legal or speech text. For Portuguese-to-Nheengatu translation, our final systems reach 31.7--32.8 chrF++ on article-disjoint constitutional text, 64.1--64.5 on the near-domain set (62.1 after removing items with target-side content overlap), and 45.2--47.5 on speech. --- ### Rethinking Opinion Measurement Interfaces for Human-in-the-Loop Optimization Authors: Changkun Ou, Sven Mayer, Daniel Buschek, Andreas Butz Published: 2026-10-10 Categories: cs.HC, cs.LG arXiv: https://arxiv.org/abs/2610.14757v1 Human-in-the-loop optimization refers to systems that present choices to a human decision-maker and query their judgment to optimize system parameters iteratively. The user interface typically provides either an absolute rating scale or asks for a preferred option. However, these approaches can potentially suffer from calibration problems, inconsistencies, and completeness violations of user preferences in such a sequential evaluation. This paper revisits the design space of opinion measurement interfaces for human-in-the-loop optimization. We conducted a study (N=360) using text summarization and image color enhancement tasks to evaluate human sequential feedback in six representative opinion measurement interfaces. Based on our analysis, we recommend using a listwise approach to support sequential opinion measurement, which counterweighs individual rating and preferential choice limitations and conveys more information for subsequent use. We further discuss the trade-offs in different user interface designs and provide guidelines to inform future research. --- ### Amharic-English Cross-Language Question Answering Using Statistical Machine Translation, Named Entity Recognition, and Semantic Indexing Authors: Emebet Bekele, Fekade Getahun Published: 2026-10-10 Categories: cs.CL arXiv: https://arxiv.org/abs/2610.14756v1 The predominance of English-language Web content creates an information-access barrier for speakers of low-resource languages such as Amharic. This paper presents the design, imple-mentation, and evaluation of a bidirectional Amharic-English cross-language question answering (CLQA) system that allows a user to formulate a natural-language question in either Amharic or English and retrieve an answer from documents written in the other language. The system integrates a statistical machine translation (SMT) component for query translation, a maximum-entropy named entity recognition (NER) component for Amharic text processing, semantic-based indexing, question analysis, passage retrieval, and answer extraction. A domain-diverse Amharic-English parallel corpus was used to train the SMT component, while monolingual cor-pora supported language modeling. Documents were normalized, tokenized, stemmed, indexed, and semantically enriched before retrieval. The original evaluation used precision and recall as the principal effectiveness measures. For Amharic-to-English retrieval, the system achieved 72% precision and 79% recall; for English-to-Amharic retrieval, it achieved 65% precision and 70% recall. These findings show that a modular translation-and-retrieval architecture can provide useful cross-language access even under low-resource conditions. This revised article preserves the original experimental claims while situating them in the context of subsequent multilingual transformer and Amharic QA research. --- ### GateSight: Recurrent Memory Anticipates Attention Memory for Query-Agnostic KV Eviction in Hybrid LLMs Authors: Junjie Li, Jiong Lou, Jie Li Published: 2026-10-10 Categories: cs.AI arXiv: https://arxiv.org/abs/2610.14754v1 Hybrid language models interleave Gated DeltaNet (GDN) layers, whose recurrent memory has a fixed size, with full-attention layers, whose key-value (KV) cache grows with the context and dominates serving memory. In this paper, we ask whether the recurrent memory already knows what the attention memory will need. We first show that which GDN heads of Qwen3.8-27B keep long memories is fixed by the weights rather than by the document. Next, we define the future utility of a KV entry by how much removing it changes the attention output over the model's own continuations, and find that the write gate, decay and write magnitude that GDN layers emit during prefill anticipate this utility before any query is seen. Because the relation is spread across many GDN heads and layers, and a linear map recovers its trend but less of its high-utility tail than a nonlinear one, we propose GateSight, which learns it with lightweight per-head multilayer perceptrons that read these signals and no attention score. Their predicted utility decides which entries each attention layer keeps, and how much each layer keeps is set by the model's native attention output gate, whose mean differs by more than an order of magnitude across layers. Finally, at serving time, GateSight scores the cache once after prefill and physically releases discarded KV entries. Experimental results on SCBench and RULER-4096 with Qwen3.8-27B and, with the same recipe, the sparse-attention model Qwen3.8-Flash-Next, up to the 262K-token limit of the models, show that with a tenth of the attention KV cache, GateSight is the most accurate of five query-agnostic methods on three of the four model-benchmark pairs and leads KVzip on all four, by 9.5 to 47.2 points. Notably, it scores a full-length context with negligible overhead relative to prefill. --- ### Random Fourier Features in Near-Linear Time Authors: Sepideh Mahabadi, Sandeep Silwal, Tal Wagner Published: 2026-10-10 Categories: cs.LG, cs.DS arXiv: https://arxiv.org/abs/2610.14751v1 Random Fourier features approximate a shift-invariant kernel in $\mathbb{R}^d$, such as the Gaussian kernel $\exp(-\tfrac12\|x-y\|_2^2)$, using independent samples from its spectral measure to construct an unbiased estimator for the kernel value with subgaussian concentration. However, generating $m$ such features requires $O(md)$ time. A long line of work on structured random features has accelerated this computation for the Gaussian kernel and other radial kernels, albeit with weaker concentration guarantees. We close this gap by showing a random feature map that can be computed in near-linear time $O((d+m)\log d)$ and yields near-subgaussian concentration. We obtain it through a new and tighter analysis of the Fastfood feature map of Le, Sarlós and Smola (ICML 2013), using Brascamp-Lieb inequalities and the Schoenberg representation of radial kernels. --- ### Equal-information evaluation of credit assignment in modular financial reinforcement learning Authors: Gennadiy Chistiakov, Yury Yanovich Published: 2026-10-10 Categories: cs.AI arXiv: https://arxiv.org/abs/2610.14750v1 Credit assignment in modular financial reinforcement learning is difficult to evaluate because an attributed signal may improve a component loss without adding decision-relevant information or improving economic outcomes. We propose an evaluation framework combining equal-information reconstruction controls, fixed-utility comparisons and prespecified downstream guardrails. The framework is instantiated in an attribution-guided modular trading architecture and tested in three controlled studies covering update allocation, supervision-target representation and portfolio risk control. We find that residual-based allocation is exactly reproduced by an ordinary loss-budget control under the disclosed interface; direct shortfall supervision improves component prediction but fails decoded-outcome preservation; and component-wise critic supervision worsens risk-condition certainty equivalent and drawdown relative to scalar supervision under fixed utility. These results identify three distinct evaluation failures: redundant attribution signals, representation-induced prediction trade-offs, and misalignment between component supervision and downstream risk outcomes. A preserved historical engineering case further shows why module-level learning claims require run-level provenance. The contribution is a reusable methodology for evaluating credit assignment in expert systems operating under partial observability, delayed feedback and economic risk constraints. --- ### More Domains, No Gains? FLO: Fast-Slow Reciprocal Learning for Multi-Domain Graph Pre-training Authors: Yupeng Tong, Lianze Shan, Jitao Zhao, Di Jin, Dongxiao He Published: 2026-10-10 Categories: cs.LG arXiv: https://arxiv.org/abs/2610.14748v1 Inspired by the great success of foundation models in the development of artificial general intelligence, Graph Foundation Models (GFMs) have recently attracted significant attention and have the potential to advance graph intelligence. Exhibiting such intelligence requires effectively learning knowledge from large-scale, multi-domain graph data during pre-training. Increasing domain diversity during pre-training is expected to yield richer and more generalizable knowledge. Most existing methods pursue this goal by designing domain alignment strategies that help a single graph model learn from multiple domains. However, we observe that existing methods struggle to benefit from increasing domain diversity. Through empirical and theoretical analysis, we reveal two key limitations. Firstly, pre-training in most GFMs is confined to a single message-passing paradigm. Since propagation flattens relations among neighboring nodes, the graph model can hardly perceive the knowledge beyond propagation, namely the attribute-predictable component of these relations. Secondly, domain alignment strategies often rely on randomly initialized parameters that are jointly updated with the graph model using immediate loss feedback, leaving early knowledge acquisition vulnerable to cross-domain interference. To address these issues, we propose FLO, a Fast--sLow reciprOcal approach for multi-domain graph pre-training. Specifically, a propagation-free attribute branch predicts the graph model's relations without message passing and returns these predictions to refine the model. We further stabilize knowledge acquisition by pairing a fast process that adapts to the current domain with a slow process that integrates knowledge across domains. On 12 datasets, FLO benefits from increasing domain diversity and achieves the best average performance in cross-domain and cross-task transfer. --- ### BridgeWAM: Bridging Video and Action Experts via Latent Bridge Queries Authors: Jiaxin Fang, Ruitong Liu, Zeyu Lai, Xingyu Wang, Mingyuan Jia, Pingyue Sheng, Weikang Wan, Mengdi Xu, Yuxiang Gao Published: 2026-10-10 Categories: cs.RO, cs.CV arXiv: https://arxiv.org/abs/2610.14746v1 World Action Models (WAMs) commonly couple video and action experts through dense, layer-wise Mixture-of-Transformers (MoT) attention, exposing action prediction to irrelevant visual features while enforcing matched expert depths. This coupling hinders selective visual-feature use and introduces redundant computation. We introduce Bridge-of-Experts (BoE), instantiated as BridgeWAM, which aggregates visual, language, and proprioceptive priors through 32 Latent Bridge Queries (LBQs) into a shared, limited latent space. Joint video-action supervision lets action gradients guide the Video Expert toward representations better suited for action prediction, while the Action Expert learns to selectively retrieve task-relevant information through cross-attention. Decoupling expert depths and reusing LBQs during action denoising yield a 5.06B model with a 30-layer Video Expert and a two-layer Action Expert, achieving 98.25% success on LIBERO and 72.24% pooled success on LIBERO-Plus, demonstrating effective task execution and strong generalization with an ultralight action head. Ablations support joint representation adaptation and unified LBQ conditioning. Relative to Fast-WAM, BridgeWAM reduces single-sample inference FLOPs by 28.72% in simulation profiling and real-robot latency to 77.1 ms (a 2.47-fold speedup). Qualitative experimental results further indicate preferential attention to target objects and task destinations across perturbations. --- ### SAGE: Pushing 2-Bit Vector-Quantized LLMs toward Real-World Usability Authors: Xianglong Yan, Zenan Wu, Kewei Zhang, Kaicheng Yang, Yulun Zhang Published: 2026-10-10 Categories: cs.AI arXiv: https://arxiv.org/abs/2610.14745v1 Extreme low-bit weight-only quantization is an effective approach to reducing the deployment costs of large language models. Below 3 bits, scalar quantization often suffers severe accuracy loss, while vector quantization (VQ) better preserves accuracy through more expressive joint encoding. However, existing VQ methods still exhibit substantial degradation on real-world tasks, and two design challenges remain underexplored: (1) protecting sensitive weight channels within jointly encoded vectors; and (2) allocating a fixed bit budget among representation components, such as direction and radius, to suit different weight matrices. To address these challenges, we propose SAGE: Split Adaptation and Grouping by Energy for weight-only vector quantization. First, Energy-Stratified Grouping (ESG) couples channel grouping with activation-weighted fitting, grouping channels from different energy tiers within each subvector to prioritize sensitive channels. Second, Adaptive Bit Split (ABS) balances direction-codebook capacity and radius precision per matrix, selecting the split with the lowest activation-weighted reconstruction error under a fixed budget. Across benchmarks covering knowledge, mathematics, code generation, instruction following, and tool use, SAGE achieves state-of-the-art performance in the 2-bit quantization setting. On Qwen3-30B-A3B, it retains 95.3% of the BF16 model's average benchmark score. With our custom-designed CUDA kernel, SAGE delivers up to a 2.55-fold non-streaming end-to-end speedup over FP16. The code is available at https://github.com/Zincwolf/sage. --- ### Japanese Handwriting Generation Based on Human Preference Feedback Using Direct Preference Optimization Authors: Wakana Kuwata, Ryota Mibayashi, Masanori Tani, Hiroaki Ohshima Published: 2026-10-10 Categories: cs.CV arXiv: https://arxiv.org/abs/2610.14743v1 We propose a Japanese handwriting generation framework that integrates human preference feedback using Direct Preference Optimization (DPO) to align generative models with human cognitive and stylistic preferences. While font generation technologies leveraging Generative Adversarial Networks and diffusion models have advanced significantly, applying these approaches to handwriting generation presents unique challenges. Font generation is a deterministic process yielding consistent styles. Handwriting generation, on the other hand, is inherently non-deterministic. In handwriting, glyph shapes never match exactly even when the same writer writes the same character. We define this shape difference as handwriting variability. Although this variability is essential for conveying the natural quality of human writing, it complicates model training and substantially increases data collection costs to capture diverse distribution patterns. To address these limitations, we incorporate human perceptual feedback into the training loop of handwriting generation. While collecting extensive handwritten data is costly, evaluating pairwise preferences between two generated samples remains very easy for humans. Applying DPO, an efficient RL-free preference optimization method, to our generation model enables high-quality handwriting generation without relying on large-scale dataset collection. Experimental results across four distinct Japanese handwriting styles demonstrate that DPO-based training enhances overall structural accuracy. However, in highly complex styles with significant variability, such as brush pen writing, the model fails to capture style-specific features. These findings highlight a critical challenge in extending DPO to complex handwriting styles. --- ### PTGPO: Prefix Tree Group Policy Optimization Authors: Ruolin Chen, Yang Li, Jihang Wang, Zhuoya Zhao, Yuchen Liu, Jiaming Xu, Qian Zhang Published: 2026-10-10 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2610.14737v1 Training language agents on long-horizon interactive tasks is challenging because sparse terminal rewards provide weak credit signals for intermediate actions. Existing group-based methods either assign a shared trajectory-level advantage or estimate step-level advantages from state groups, which may conflate the current action with subsequent decisions. We propose Prefix Tree Group Policy Optimization (PTGPO), a critic-free method that constructs prefix trees over trajectory suffixes within anchor-state groups. PTGPO assigns each tree edge a credit given by the difference between its child and parent empirical mean returns, capturing the conditional return change revealed by each additional action. These prefix-tree credits provide fine-grained learning signals without requiring a critic or additional rollouts. Experiments on interactive benchmarks ALFWorld, WebShop and Sokoban show that PTGPO consistently improves over RL baselines, demonstrating its effectiveness. --- ### Think in Strategies, Act with Tools: Traceable Agentic Optimization for Combinatorial Problems Authors: Zhongnian Li, Shi Xu, Peng Ying, Weikang Gong, Zhixiao Wang, Xinzheng Xu Published: 2026-10-10 Categories: cs.AI arXiv: https://arxiv.org/abs/2610.14736v1 Optimization by PROmpting (OPRO) directly leverages Large Language Models (LLMs) to generate candidate solutions through natural-language prompting. However, OPRO's optimization process is difficult to track, which prevents LLMs from capturing useful information accumulated over previous iterations, thereby limiting solution quality. To address this limitation, we propose a Tool-Augmented Strategic Agent (TASA) which summarizes and exploits trajectory to improve solution quality. Specifically, TASA plans the strategy for each step based on the optimization trajectory, which decomposes the optimization process into traceable strategy selections. Moreover, we leverage LLM to construct problem-specific tools and introduce a strategy-guided adaptive tool planner, which efficiently utilizes the information of trajectory to update solutions. We establish local convergence under persistent exploration and evaluate the complete framework under a shared candidate-evaluation budget. The results support its solution quality and the traceability of strategy-guided search. --- ### Align-VAE: Aligning Attention and Geometry in VecSet VAEs for 3D Reconstruction Authors: Huaizhi Qu, Mingxin Yang, Xintong Han, Fangneng Zhan, Chunchao Guo Published: 2026-10-10 Categories: cs.CV arXiv: https://arxiv.org/abs/2610.14734v1 Set-based variational autoencoders (VecSet VAEs) represent 3D shapes as compact sets of latent tokens. Although each token originates from a query at a sampled surface location, whether these latents truly align with geometry remains unclear. We investigate this alignment and identify four issues: (i) attention-sink-like tokens that attract queries from across the shape; (ii) spatially periodic repetitions in decoder attention responses; (iii) mismatched normal information in encoder and decoder query features; and (iv) missing supervision of the intrinsic relation between signed distance fields and surface normals. To address these issues, we introduce Align-VAE with four corresponding modifications: (i) a shared bank of learnable decoder tokens provides a workspace for global computation; (ii) distance-biased cross-attention guides retrieval toward nearby predicted spatial anchors; (iii) normal-free encoder queries align the two stages' access to normal information while retaining normals in the encoder keys and values; and (iv) a screened Poisson-inspired regularizer supervises surface values and normal-direction derivatives. These changes align latent retrieval and field prediction with geometry without enlarging the per-shape code. On the ABO dataset, Align-VAE reduces Chamfer distance by 26.6% compared to the standard VecSet VAE. --- ### VitaBench: Memory Lifecycles Beyond Retrieval Authors: Muquan Yu, Dihan Hu Published: 2026-10-10 Categories: cs.LG arXiv: https://arxiv.org/abs/2610.14731v1 Long-term memory benchmarks for language agents primarily evaluate whether information from earlier interactions can be retrieved and used when relevant. Persistent memory systems face an additional problem before future use is known: deciding whether a memory should remain exact, be compressed, consolidated, or deleted. Existing benchmarks capture long-range retrieval, temporal reasoning, evidence reuse, and incremental interaction, but do not directly evaluate these earlier representation decisions. We formalize memory lifecycle evaluation and introduce VitaBench, a benchmark that makes lifecycle decisions measurable through repeated memory reviews and full-horizon counterfactual evaluation. VitaBench contains a controlled Diagnostic tier that isolates future-use patterns and a Naturalistic tier that composes them into generated customer-service-style interaction streams. We characterize lifecycle decision structure with three diagnostics: lifecycle headroom, measuring the advantage of adaptive choices over the best fixed strategy; horizon reversal, measuring when short-horizon decisions change under delayed consequences; and optimal-action diversity, measuring whether different memory states favor different lifecycle operations. Across both tiers, VitaBench exhibits substantial lifecycle structure: lifecycle headroom is 1.364 in Diagnostic and 1.405 in Naturalistic, and 60.2% and 56.5% of states, respectively, change their preferred action between short and full horizons. A learned delayed lifecycle policy further improves over the best fixed strategy on VitaBench-Naturalistic, reaching 101.5 versus 83.4 joint utility. These results show that retrieving past information and deciding how it should persist are distinct evaluation problems, and that lifecycle management requires evaluation of the delayed consequences of representation decisions. --- ### Self-improving data synthesis enables scalable adaptation of embedding models Authors: Zucan Lyu, Rui Tang, Jianlyu Chen, Peisheng Jin, Defu Lian Published: 2026-10-10 Categories: cs.IR, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2610.14730v2 AI agents require specialised retrieval across domains and tasks, but adapting embedding models is constrained by costly manual training-data design. Here we introduce SYNTRA, a self-improving agentic framework that turns a few retrieval examples into reusable synthesis instructions. Given an unlabelled corpus and a few relevant query-document pairs, SYNTRA infers implicit information needs and relevance criteria, then uses feedback on trial outputs to refine instructions for query generation and relevance assessment. Experiments across 24 datasets spanning 15 domains demonstrate that the same synthesis workflow produces effective training data, benefiting even strong embedding models with and without existing labelled training sets. Self-improvement strengthens this supervision on all six evaluated tasks, yielding better retrieval than human calibration on four. The training value persists at larger scales: on MS MARCO, retrieval gains continue across a 32-fold range of synthetic training-data sizes. SYNTRA thus enables scalable embedding-model adaptation with minimal human intervention. --- ### MHWA: Multi-timescale Hierarchical World-Action Model Authors: Pengcheng Pan, Guoqing Ma, Yuhan Zhang, Yang Chen, Yichen Liu, Ziheng Li, Shan Yu Published: 2026-10-10 Categories: cs.AI arXiv: https://arxiv.org/abs/2610.14727v1 Autoregressive world models have emerged as a powerful paradigm for sample efficient decision making, yet their open loop rollouts often exhibit increasing drift that leads to error accumulation even within a relatively small planning horizon. Hierarchical temporal abstraction is a natural candidate to mitigate such drift by reducing the effective prediction horizon; however, existing hierarchical designs can introduce added structural and conceptual complexity, sometimes making them difficult to control and limiting gains over flat baselines. To this end, we introduce MHWA (Multi-timescale Hierarchical World-Action model), which employs a strided high level context channel to reduce global context update frequency, effectively curbing the propagation of compounding drift. Crucially, via context aware routing, the gating network dynamically reconfigures a low level mixture of experts by using both global context and current observations, thereby enabling flexible high-low coordination, helping alleviate interference commonly observed in static hierarchies. Empirically, using RLiable's seed-by-game aggregation, MHWA (77M parameters) achieves an IQM human normalized score (HNS) of 0.884 across three training seeds on a 14-game Atari subset using 10% subsampled offline data, outperforming a matched JOWA-70M baseline (0.529) and achieving numerically comparable performance to JOWA-150M while using nearly half the parameters and 70.88% fewer training FLOPs per sequence. --- ### LightVAE: Towards Compact and Efficient Video Autoencoders Authors: Shiqiao Gu, Yang Yong, Fengjun Zhong, Ruihao Gong, Chengtao Lv, Xiangyu Fan, Lei Yang, Xianglong Liu Published: 2026-10-10 Categories: cs.CV arXiv: https://arxiv.org/abs/2610.14726v1 Latent diffusion models enable high-quality video generation, but video VAE encoding and decoding remain computationally expensive. Existing methods use stage-specific lightweight operators, yet either constrain spatial and temporal computation jointly or remove temporal computation altogether, without distinguishing their compression sensitivities in pretrained Conv3D. To address these limitations, we present LightVAE, a compression framework that exploits redundancy in temporal modeling and network depth. LightVAE exactly reparameterizes eligible causal Conv3D layers into an aggregate spatial response and a difference-driven temporal correction. Based on their distinct compression sensitivities, temporal low-rank compression (TLR) keeps the aggregate spatial kernel dense and applies layer-specific low-rank approximation only to the temporal correction. An algebraically equivalent folded implementation converts the arithmetic savings into practical acceleration. LightVAE further uses width-preserving block pruning to reduce depth and probe-guided feature alignment, guided by feedback from a frozen teacher suffix, to recover quality after compression. The framework also extends to video encoders. Across three video VAE backbones and two GPU platforms, LightVAE consistently accelerates decoding. For the Wan VAE family, it achieves up to $2.24\times$ and $1.11\times$ the decoding throughput of state-of-the-art efficient VAE decoders on NVIDIA H100 and RTX~5090, respectively, with higher DAVIS PSNR and lower LPIPS. On MiniMax-H3, it achieves up to $2.97\times$ the original decoding throughput while retaining 99.51\% of the original model's overall VBench-2.0 score. Our code is available at https://github.com/ModelTC/LightVAE. --- ### Integrating AIRL-Based Reward Estimation into a Control Method Using Fuzzy Inference and Policy Gradient Authors: Seiji Ishihara Published: 2026-10-10 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2610.14723v1 A method integrating policy gradient with fuzzy inference has been proposed. This method, hereafter referred to as the base model, has high flexibility with respect to the configuration of fuzzy inference rules and fuzzy representation. It can also automatically adjust the rule weights and the membership functions in the antecedent and consequent parts using the same framework for learning. In methods that integrate reinforcement learning with fuzzy inference, however, the reward function must be designed in advance. It is therefore useful to add learning-based reward estimation to the base model. This paper proposes a fuzzy control method that integrates a reward estimation mechanism based on the adversarial inverse reinforcement learning (AIRL) framework into the base model. The generative adversarial nets (GAN) framework adopted in AIRL, however, tends to suffer from unstable learning. To address this, the proposed method treats the policy of the base model as the generator and also uses the set of membership functions in the policy as the cost function of the discriminator. This configuration allows the discriminator update for reward estimation and the generator update for policy estimation to be performed simultaneously on the parameters shared by both, without estimating a value function. In simulations of an illustrative vehicle speed control task, the proposed method acquired policies with high generalization performance through stable and efficient learning. --- ### Relational Semantic Grounding for Test-Time Discovery Authors: Tianle Liu, Gongbo Cheng, Chenggong Ni, Minming Gu, Tao Zhou, Fuyuan Hu, Fan Lyu Published: 2026-10-10 Categories: cs.AI arXiv: https://arxiv.org/abs/2610.14721v1 Test-Time Discovery (TTD) aims to recognize known classes and detect novel ones from a continuous unlabeled test stream after deployment. Existing TTD methods mainly rely on visual clustering or prototype comparison, and can only assign anonymous labels to discovered classes, leaving them difficult to interpret in open-world applications. In this paper, we propose SEGA, a Semantic Grounding with Attributes framework for TTD. SEGA leverages large language models to generate attribute-rich descriptions for known classes and constructs an attribute-anchored visual-semantic space during training. At test time, we introduce Relational Semantic Grounding (RSG), which grounds unknown samples with similar and dissimilar known-class anchors, prompts the LLM to generate candidate concepts, and reranks them in the learned semantic space. The selected concept is further used for confidence-guided test-time semantic stabilization. By coupling attribute-level semantic alignment with relational concept grounding, SEGA transforms anonymous test-time clusters into interpretable semantic concepts. Experiments show that SEGA improves TTD baselines with the promise of semantic grounding for open-world. --- ### Multi-Token Random-Walk Learning via Meeting-Triggered Information Exchange Authors: Taichun Yeh, Xingran Chen Published: 2026-10-10 Categories: cs.LG arXiv: https://arxiv.org/abs/2610.14716v1 Random walk (RW)-based algorithms are widely used in decentralized learning due to their low communication overhead and scalability. While most existing methods rely on a single token, multiple tokens can explore the network in parallel and exchange information upon meeting. However, the fundamental benefits of such meeting-triggered interactions remain unclear. In this work, we investigate multi-token RW learning under two information exchange policies: \textsc{Average-Gradient} and \textsc{Average-Model}, which average stochastic gradients and updated models, respectively, when tokens meet. We establish iteration complexity lower bounds for non-convex, convex, and strongly convex objectives on complete graphs, which apply to both exchange policies. Compared with single-token RW learning, the scaling factor in these bounds is reduced by a factor of $m^2$ for non-convex and convex objectives, and by an exponentially in $m$ for strongly convex objectives, where $m$ denotes the number of tokens. Numerical experiments reveal distinct behaviors of the two policies under two performance criteria: mean-token loss (the average loss across all tokens) and best-token loss (the minimum loss across all tokens). \textsc{Average-Model} achieves lower mean-token loss and is relatively insensitive to the network-to-token ratio $N/m$, where $N$ is the number of nodes. In contrast, \textsc{Average-Gradient} achieves lower best-token loss but exhibits a trade-off between parallel exploration and gradient interference as $N/m$ decreases. --- ### PixARG: Pixel-Asynchronous Retrieval-Augmented Diffusion for Fast and Realistic Image Generation Authors: Zewei Xin, Qinya Li, Chaoyue Niu, Fan Wu Published: 2026-10-10 Categories: cs.CV arXiv: https://arxiv.org/abs/2610.14713v1 While diffusion models achieve impressive image generation capabilities, they suffer from high computational costs, implausible artifacts, and limited adaptability to dynamic or private data due to static, centralized training. To address these, we propose PixARG, a novel retrieval-augmented generative framework that integrates dynamic real-world image retrieval into the diffusion pipeline to simultaneously enhance efficiency and visual realism. Specifically, PixARG constructs a pixel-level affinity map between the user's text prompt and the retrieved image to guide a heterogeneous noise initialization. By modulating noise intensity across spatial regions, this mechanism preserves semantically relevant structural priors while perturbing less relevant areas, thereby significantly accelerating convergence. To accommodate this spatially non-uniform noise and ensure synchronized generation, we introduce a pixel-wise asynchronous denoising scheduler and a spatiotemporal denoiser conditioned on the affinity map. Furthermore, we derive an affinity-adaptive training objective that elegantly balances deterministic restoration with stochastic regeneration across pixels. Experiments on LAION and COYO datasets across offline, online, and user-specific settings show PixARG enhances realism and semantic alignment while drastically reducing diffusion steps. --- ### DeKoR: Decomposed Knowledge Routing for Efficient GNN--LLM Fusion Authors: Peiyuan Li, Yongqi Huang, Jitao Zhao, Di Jin, Dongxiao He Published: 2026-10-10 Categories: cs.LG arXiv: https://arxiv.org/abs/2610.14710v1 Large Language Models (LLMs) have shown strong potential for text-attributed graph learning by leveraging rich pre-trained knowledge and language understanding capabilities. Combining LLMs with Graph Neural Networks (GNNs) offers a promising paradigm for expanding the modeling capabilities of graph learning. However, LLMs and GNNs follow distinct learning paradigms, leading to a pronounced mismatch between their knowledge spaces. Such discrepancies make effective knowledge fusion nontrivial. Moreover, existing methods often incur substantial additional computation for only limited performance gains. To address these challenges, we propose DeKoR, an efficient framework for GNN and LLM knowledge fusion. To enable efficient knowledge transfer, DeKoR decomposes LLM knowledge into complementary local and global views and uses the LLM as an offline teacher to distill them into lightweight experts. To enable selective knowledge fusion, an adaptive router learns how much each node should rely on each expert. Extensive experiments on multiple benchmarks demonstrate that DeKoR outperforms existing baselines on both node and subgraph classification, achieving superior effectiveness and efficiency. --- ### Vela: User Behavior as Supervision Sparks LLM Intelligence Authors: Yang Xiao, Yusong Sun, Haoyi Wu, Wenyang Hui, Weiqiang Hang, Luqing Wang, Zejian Wang, Zirui Tu, Jingsheng Mao, Shaobo Hu, Yang Luo, Jiale Guo et al. Published: 2026-10-10 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2610.14709v1 Large language models are increasingly deployed in applications that act on users' long-term behavior histories, from personal assistants to recommendation and advertising systems. These applications demand the user dimension of LLM intelligence: inferring a person's preferences, intent, and likely future actions from weeks to years of behavior data. Yet current models are trained on static text or synthetic data, neither of which captures how user behavior evolves over time. Real-world behavior records, already available at scale on content platforms and in personal-assistant agents, offer a natural supervisory signal. We propose user behavior as supervision and introduce Vela, a framework that converts real user behavior data into trainable tasks and trajectories. From the anonymized behavior records of 100,000 users, Vela generates 100,000 tasks across 91 real-world task templates; difficulty-based filtering and trajectory acceptance criteria retain approximately 75,000 tasks with selected training trajectories. Fine-tuning Qwen3.6-35B-A3B on these trajectories improves performance over the base model by up to 11.5 points on user-related tasks and by an average of 4 points on general benchmarks. Training on more user behavior data further improves user-related performance, and reinforcement learning after fine-tuning outperforms reinforcement learning from the base model by up to 7.7 points. These gains suggest that real-world user behavior is a scalable source of supervision for both user-centered capabilities and the broader intelligence of LLMs. --- ### How to Build a Benchmark That Actually Measures Taste Authors: Xiaofan Wu, Xi Zeng, Miaoxia Chen, Peishan Chen, Shuyan Li, Chen Wang, Xiaohan Wang, Wanling Xu, Jiyun Yao, Hanyong Zhong, Jiahao Zhu Published: 2026-10-10 Categories: cs.AI, cs.CL, cs.CV arXiv: https://arxiv.org/abs/2610.14707v2 Expert-aligned "taste" benchmarks should reward craft judgment, not the ability to game the test. We introduce TasteBench, a pairwise good-vs-bad benchmark across visual, brand, book, and tech items under a frozen rubric with expert-only gold (panel n>=3; percent agreement >=0.60 primary gate; weak engagement labels train/analysis-only), together with a control protocol that separates craft judgment from test-gaming. Applying the protocol to a Round-1 pilot (primary_human_expert; three biological human experts; n=250 pairs) surfaces and quantifies three validity threats: two construction artifacts that inflate the trivial-baseline layer, and a difficulty ceiling that inflates every score. (1) Position: the preferred item is slot A with probability 0.956, so an always_item_a policy scores 0.956 (fair chance 0.5). (2) Lexical leakage: a filename token cue scores 0.912. (3) Difficulty ceiling: craft domains show near-ceiling human agreement (book/brand/tech 1.0), and even under A/B randomization (eval_assignment_seed7; gold_side_A ~0.47) all five VLM cards reach 1.0 on craft, with only visual separating (~0.84-0.91; a post-hoc hard subset, n=107, leaves craft at 1.0). The protocol (expert-only gold, a weak-label firewall, side randomization at construction, filename scrubbing, and hard filters) makes each artifact measurable and, for the two construction artifacts, correctable; we release a reference eval harness and a side-randomization tool, and show the tool removes the position skew (0.916 to 0.495/0.505). The result is a calibrated view of what the benchmark does and does not yet measure: position and leakage are fixable, whereas the craft-domain ceiling is a pair-difficulty limit that Round-2 must address. We do not evaluate engagement, era/circle conditioning, or calibration here, and we claim no general taste SOTA. --- ### Lingdan-V2: Large language models for clinically grounded reasoning in traditional Chinese medicine Authors: Rui Hua, Yu Wei, Ziri Jin, Zhuo Sun, Kai Chang, Jianan Xia, Zixin Shu, Xiaodong Li, Dongmei Jia, Fei Dong, Runshun Zhang, Jian Yu et al. Published: 2026-10-10 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2610.14705v1 Large language models have advanced medical reasoning, but clinically grounded reasoning in traditional Chinese medicine (TCM) remains challenging because prescription decisions require integrating domain knowledge, syndrome differentiation, treatment principles, herbal prescription composition, and dosage determination. We develop a TCM language model family comprising models with 4 billion, 8 billion, and 14 billion parameters. The artificial intelligence contribution is a staged clinical alignment framework integrating large scale continued pretraining on TCM and clinical records, complementary reasoning supervision, and prescription-centered reinforcement learning (RL). We further formulate herbal prescription generation as a structured RL problem with prescription-level feedback, enabling direct optimization of clinical decision outputs beyond knowledge reproduction and generic instruction following. The engineering application is TCM clinical decision support, particularly herbal prescription recommendation. We evaluate the models on a public benchmark containing 25,620 expert-verified items across 13 TCM knowledge and clinical reasoning tasks and on a private retrospective dataset of 4,348 spleen and stomach disease prescription cases. The 14 billion parameter reinforcement aligned model achieves a prescription generation F1 of 37.0 on the public benchmark, compared with 32.3 for DeepSeek-R1 with 671 billion total parameters, and a herb-level F1 of 52.8 on the clinical dataset with improved dosage agreement. Reinforcement post-training consistently improves prescription recommendation across all scales, while physician-source analysis shows that expert heterogeneity affects reward stability. These findings demonstrate a scalable pathway from domain grounding and reasoning supervision to clinically oriented prescription decision support. --- ### Linear-Time Non-Local Codec Placement for Machine Learning Model Compression Authors: Matteo Spallanzani Published: 2026-10-10 Categories: cs.LG arXiv: https://arxiv.org/abs/2610.14703v1 On-device inference imposes constraints on model storage and operand traffic and thus benefits from compressed operand representations, which mandate the use of encoding and decoding operators. However, naïve placement of codec operators can cancel the operand traffic gains sought from compression. In this work, I present an algorithm to place codec operators into a source computational graph to make it compatible with target hardware. The edges of the graph are colored so that each induced subgraph is weakly connected and has a maximal node that both postdominates all subgraph inputs and dominates all subgraph outputs; this unique node - dubbed the $\textit{absolute dominator}$ - is the natural location to observe the traffic flowing through the subgraph and learn parametric codec operators. The algorithm's time complexity is linear in the source graph's size. With respect to naïve local codec placement, the algorithm reduces modeled operand traffic by $24\%$-$44\%$ across four models, including $24\%$ on SigLIP and $34\%$ on SmolLM2. --- ### Towards Merging-conditional Backdoor Unalignment against Large Language Models Authors: Yukun Chen, Yidou Chen, Yiming Li, Xinyu Zhang, Baosong Yang, Hongwei Yao, Zhan Qin Published: 2026-10-10 Categories: cs.CR, cs.AI, cs.CL arXiv: https://arxiv.org/abs/2610.14701v1 Model merging has become widely adopted for enhancing the task-specific performance of Large Language Models (LLMs) without additional training. While the specialized model from third-party platforms may undergo safety audits (e.g., backdoor detection), we reveal a novel and critical threat: One-sided Merging-Conditional Backdoor (OMCB). OMCB remains dormant and undetectable within one standalone model but activates upon merging, triggering safety unalignment. We first theoretically analyze the existence and geometric location of OMCB, revealing that the parameter shift crossing safety decision boundaries is the main cause of backdoor activation. Building upon this insight, we propose SMOKE, a meta-learning-based attack framework. SMOKE implants OMCB via a bi-level optimization strategy: the inner loop synthesizes surrogate meta-vectors through random low-rank initialization and finite-step maximization, while the outer loop optimizes the target model by simulating the merging process. Furthermore, preliminary analysis shows that the injected OMCB has bounded generalization loss under unknown merging configurations. Extensive experiments across diverse model architectures and merging techniques validate the effectiveness of SMOKE and its resistance to existing backdoor detections, highlighting a significant yet previously overlooked vulnerability in the model merging paradigm. --- ### DSABench: Benchmarking Source Discovery, Coordination, and Execution for Multi-Database Agents Authors: Jinlong Luo, Haoran Wu, Bo Xu Published: 2026-10-10 Categories: cs.AI arXiv: https://arxiv.org/abs/2610.14700v1 Most text-to-SQL benchmarks provide the target database. In a large data catalog, however, choosing the sources is part of the problem: an agent must find every database needed for a request, determine how their results fit together, and execute the resulting workflow. We present DSABench to evaluate these capabilities jointly. Its 815 human-reviewed tasks draw on Spider, BIRD, and LiveSQLBench; 618 require several databases to contribute, and 105 allow more than one interpretation of the required sources. Executable contracts define source participation, cross-source dependencies, and source-local execution, while RouteEnv lets agents inspect public catalog evidence without revealing the gold annotations. In our routing-only study across seven model backends, the highest RA-Set score is 60.55%, compared with a highest RA-Graph score of 34.66% when coordination requirements are included. The gap becomes more pronounced when execution is required. Across four backends, median non-privileged Route+Exec-Graph success ranges from 6.69% to 7.98% without clarification and from 11.44% to 13.25% with one clarification; the best observed clarified configuration reaches 16.99%. On ambiguous tasks, clarification improves exact branch-source recovery by 39.8-43.2 percentage points. These findings illustrate why retrieving plausible databases or producing correct local SQL is not enough: an agent must carry the right information through the entire multi-database workflow. --- ### Joint Discrete and SE(3) Diffusion for Protein Interface Design Authors: Zhenqiao Song, Jianwen Xie, Lei Li Published: 2026-10-10 Categories: cs.AI arXiv: https://arxiv.org/abs/2610.14699v1 The design of high-affinity, target-specific protein binders is central to the development of modern therapeutics. However, existing protein-generation methods remain limited in this setting, as many generated designs fail to simultaneously achieve geometric and physicochemical complementarity at binding interfaces. To address this challenge, we introduce PIDiff, a physicochemically enhanced multimodal diffusion framework for protein-protein complex design. PIDiff jointly designs target-conditioned protein sequences and structures by coupling diffusion in discrete sequence space with diffusion on the continuous (SE(3)) structural manifold. At its core, PIDiff employs PIFormer, a novel architecture that interleaves geometry-aware invariant point attention with global sequence self-attention. We further curate PIBench, a protein-protein complex dataset with diverse target scaffolds derived from the Protein Data Bank (PDB). We first pretrain PIDiff on PIBench and subsequently fine-tune it for two therapeutically relevant applications: minibinder design and antibody design. PPFrame consistently outperforms baseline methods across sequence design success rate, structure binding energy, and sequence-structure co-designability on all tasks. --- ### PURE: Refining Feed-forward 3D Reconstruction with Unlabeled Post-training Authors: Yaolong Zhu, Peng Li, Jiahao Lu, Chengfeng Zhao, Mengfei Li, Chao Wang, Wenyue Chen, Zekai Gu, Chengwei Ren, Ronggang Wang, Yuan Liu Published: 2026-10-10 Categories: cs.CV arXiv: https://arxiv.org/abs/2610.14698v1 Feed-forward 3D reconstruction models have achieved remarkable progress through large-scale training and strong learned priors. However, they remain fragile when applied to dynamic in-the-wild videos, where non-rigid motion, occlusions, and illumination changes are prevalent. A key limitation lies in their reliance on expensive depth and camera pose annotations, which hinders scaling to diverse real-world dynamic data and restricts generalization. To address this issue, we propose \textbf{PURE}, an unlabeled post-training refinement framework that improves depth and camera estimation without requiring ground-truth supervision. Instead of relying on annotations, PURE elegantly distills robust structural knowledge from state-of-the-art 2D priors into 4D feed-forward architectures. PURE leverages relative depth constraints, affine-invariant geometry distillation, and pose refinement based on feature correspondences to stabilize geometric predictions under challenging dynamic conditions. Extensive experiments on both synthetic and real-world dynamic benchmarks demonstrate that PURE delivers consistent improvements across multiple state-of-the-art architectures. Furthermore, our data-efficient paradigm effectively mitigates domain shifts, successfully unlocking the architectural limits of these base models in challenging, in-the-wild dynamic scenes. --- ### AlignAudio: A Unified Dual-Alignment Framework for Noise-Robust Speech-Conditioned Generation Authors: Yixuan Li, Xuenan Xu, Zeyu Xie, Mengyue Wu Published: 2026-10-10 Categories: cs.SD, cs.AI arXiv: https://arxiv.org/abs/2610.14697v1 Speech-conditioned generation maps speech signals to target audio outputs. Prior end-to-end speech-conditioned generative models achieve high fidelity in clean acoustic conditions, but their performance degrades substantially under environmental noise. While prior noise-robust approaches have achieved significant success in feature-level purification, maintaining consistent generation under clean and noisy conditioning remains a key challenge. To address this, we propose AlignAudio, a noise-robust generative framework based on dual alignment: (i) aligning the representations of clean and noisy speech to preserve semantic cues, and (ii) enforcing generation-level consistency within a flow matching paradigm to improve conditioning fidelity. Experiments on speech-to-audio (STA) and voice conversion (VC) tasks demonstrate that AlignAudio maintains performance comparable to baselines on clean speech while significantly outperforming them across various noisy conditions. Furthermore, the proposed dual-alignment strategy is applicable to both single- and multi-stream speech conditioning and can be integrated into different flow matching-based generation frameworks and post-training settings. Audio demonstrations and qualitative comparisons: \href{https://alignaudio.github.io/AlignAudio}{\textcolor{cyan}{\textit{https://alignaudio.github.io/AlignAudio}}}. ---