# Research Papers — 2026-02 · bundle page 195 of 210 > Full-text markdown of papers #9701–9750 of 10493 published in 2026-02, newest first. AI-readable mirrors of the full 10493-paper 2026-02 archive (50 papers per file). Canonical page: https://aipod.dev/papers/archive/2026-02/page/195 [← Previous bundle](https://aipod.dev/papers/archive/2026-02/page/194.md) [Next bundle →](https://aipod.dev/papers/archive/2026-02/page/196.md) ### Mitigating Heterogeneity-Induced Drift in Hierarchical Sign-Based Federated Learning Authors: Amirreza Kazemi, Seyed Mohammad Azimi-Abarghouyi, Gabor Fodor, Carlo Fischione Published: 2026-02-02 Categories: cs.DC, cs.IT, cs.LG arXiv: https://arxiv.org/abs/2602.02355v2 Hierarchical federated learning (HFL) is well suited for large-scale wireless and Internet of Things systems, where devices communicate with nearby edge servers before reaching the cloud. In these environments, uplink bandwidth and latency impose strict communication constraints, making aggressive gradient compression essential. One-bit sign-based stochastic gradient descent methods provide an attractive solution in flat federated settings, but their behavior in hierarchical edge--cloud architectures remains insufficiently understood, especially under inter-cluster data heterogeneity. To address this gap, we develop a sign-based HFL framework in which devices transmit binary stochastic-gradient signs to edge servers, edge servers apply majority voting, and the cloud periodically aggregates edge models. Our analysis reveals that inter-cluster heterogeneity induces a persistent bias term in the convergence bound, reflecting the drift of edge models toward local objectives. This term cannot be removed by increasing the number of training rounds or by tuning standard hyperparameters alone. We therefore propose \(\mathtt{DC\text{-}HierSignSGD}\), a drift-corrected sign-based HFL algorithm in which devices apply a cloud-assisted gradient correction before taking the sign. We show that this pre-sign correction mitigates the non-vanishing heterogeneity-induced bias while preserving binary device--edge communication during the repeated local sign-update steps. Experiments under severe inter-cluster heterogeneity demonstrate that \(\mathtt{DC\text{-}HierSignSGD}\) improves the stability and accuracy of sign-based HFL and achieves performance comparable to full-precision hierarchical SGD with substantially lower device--edge communication. --- ### Implicit neural representation of textures Authors: Albert Kwok, Zheyuan Hu, Dounia Hammou Published: 2026-02-02 Categories: cs.CV, cs.AI, cs.GR, cs.LG arXiv: https://arxiv.org/abs/2602.02354v1 Implicit neural representation (INR) has proven to be accurate and efficient in various domains. In this work, we explore how different neural networks can be designed as a new texture INR, which operates in a continuous manner rather than a discrete one over the input UV coordinate space. Through thorough experiments, we demonstrate that these INRs perform well in terms of image quality, with considerable memory usage and rendering inference time. We analyze the balance between these objectives. In addition, we investigate various related applications in real-time rendering and down-stream tasks, e.g. mipmap fitting and INR-space generation. --- ### Artificial Intelligence and Symmetries: Learning, Encoding, and Discovering Structure in Physical Data Authors: Veronica Sanz Published: 2026-02-02 Categories: hep-ph, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2602.02351v1 Symmetries play a central role in physics, organizing dynamics, constraining interactions, and determining the effective number of physical degrees of freedom. In parallel, modern artificial intelligence methods have demonstrated a remarkable ability to extract low-dimensional structure from high-dimensional data through representation learning. This review examines the interplay between these two perspectives, focusing on the extent to which symmetry-induced constraints can be identified, encoded, or diagnosed using machine learning techniques. Rather than emphasizing architectures that enforce known symmetries by construction, we concentrate on data-driven approaches and latent representation learning, with particular attention to variational autoencoders. We discuss how symmetries and conservation laws reduce the intrinsic dimensionality of physical datasets, and how this reduction may manifest itself through self-organization of latent spaces in generative models trained to balance reconstruction and compression. We review recent results, including case studies from simple geometric systems and particle physics processes, and analyze the theoretical and practical limitations of inferring symmetry structure without explicit inductive bias. --- ### Context Learning for Multi-Agent Discussion Authors: Xingyuan Hua, Sheng Yue, Xinyi Li, Yizhe Zhao, Jinrui Zhang, Ju Ren Published: 2026-02-02 Categories: cs.AI, cs.LG, cs.MA arXiv: https://arxiv.org/abs/2602.02350v3 Multi-Agent Discussion (MAD) has garnered increasing attention very recently, where multiple LLM instances collaboratively solve problems via structured discussion. However, we find that current MAD methods easily suffer from discussion inconsistency, LLMs fail to reach a coherent solution, due to the misalignment between their individual contexts.In this paper, we introduce a multi-LLM context learning method (M2CL) that learns a context generator for each agent, capable of dynamically generating context instructions per discussion round via automatic information organization and refinement. Specifically, inspired by our theoretical insights on the context instruction, M2CL train the generators to control context coherence and output discrepancies via a carefully crafted self-adaptive mechanism.It enables LLMs to avoid premature convergence on majority noise and progressively reach the correct consensus. We evaluate M2CL on challenging tasks, including academic reasoning, embodied tasks, and mobile control. The results show that the performance of M2CL significantly surpasses existing methods by 20%--50%, while enjoying favorable transferability and computational efficiency. --- ### Why Steering Works: Toward a Unified View of Language Model Parameter Dynamics Authors: Ziwen Xu, Chenyan Wu, Hengyu Sun, Haiwen Hong, Mengru Wang, Yunzhi Yao, Longtao Huang, Hui Xue, Shumin Deng, Zhixuan Chu, Huajun Chen, Ningyu Zhang Published: 2026-02-02 Categories: cs.CL, cs.AI, cs.CV, cs.IR, cs.LG arXiv: https://arxiv.org/abs/2602.02343v3 Methods for controlling large language models (LLMs), including local weight fine-tuning, LoRA-based adaptation, and activation-based interventions, are often studied in isolation, obscuring their connections and making comparison difficult. In this work, we present a unified view that frames these interventions as dynamic weight updates induced by a control signal, placing them within a single conceptual framework. Building on this view, we propose a unified preference-utility analysis that separates control effects into preference, defined as the tendency toward a target concept, and utility, defined as coherent and task-valid generation, and measures both on a shared log-odds scale using polarity-paired contrastive examples. Across methods, we observe a consistent trade-off between preference and utility: stronger control increases preference while predictably reducing utility. We further explain this behavior through an activation manifold perspective, in which control shifts representations along target-concept directions to enhance preference, while utility declines primarily when interventions push representations off the model's valid-generation manifold. Finally, we introduce a new steering approach SPLIT guided by this analysis that improves preference while better preserving utility. Code is available at https://github.com/zjunlp/EasyEdit/blob/main/examples/SPLIT.md. --- ### LongVPO: From Anchored Cues to Self-Reasoning for Long-Form Video Preference Optimization Authors: Zhenpeng Huang, Jiaqi Li, Zihan Jia, Xinhao Li, Desen Meng, Lingxue Song, Xi Chen, Liang Li, Limin Wang Published: 2026-02-02 Categories: cs.CV arXiv: https://arxiv.org/abs/2602.02341v1 We present LongVPO, a novel two-stage Direct Preference Optimization framework that enables short-context vision-language models to robustly understand ultra-long videos without any long-video annotations. In Stage 1, we synthesize preference triples by anchoring questions to individual short clips, interleaving them with distractors, and applying visual-similarity and question-specificity filtering to mitigate positional bias and ensure unambiguous supervision. We also approximate the reference model's scoring over long contexts by evaluating only the anchor clip, reducing computational overhead. In Stage 2, we employ a recursive captioning pipeline on long videos to generate scene-level metadata, then use a large language model to craft multi-segment reasoning queries and dispreferred responses, aligning the model's preferences through multi-segment reasoning tasks. With only 16K synthetic examples and no costly human labels, LongVPO outperforms the state-of-the-art open-source models on multiple long-video benchmarks, while maintaining strong short-video performance (e.g., on MVBench), offering a scalable paradigm for efficient long-form video understanding. --- ### Rethinking Generative Recommender Tokenizer: Recsys-Native Encoding and Semantic Quantization Beyond LLMs Authors: Yu Liang, Zhongjin Zhang, Yuxuan Zhu, Kerui Zhang, Zhiluohan Guo, Wenhang Zhou, Zonqi Yang, Kangle Wu, Yabo Ni, Anxiang Zeng, Cong Fu, Jianxin Wang et al. Published: 2026-02-02 Categories: cs.IR, cs.AI arXiv: https://arxiv.org/abs/2602.02338v1 Semantic ID (SID)-based recommendation is a promising paradigm for scaling sequential recommender systems, but existing methods largely follow a semantic-centric pipeline: item embeddings are learned from foundation models and discretized using generic quantization schemes. This design is misaligned with generative recommendation objectives: semantic embeddings are weakly coupled with collaborative prediction, and generic quantization is inefficient at reducing sequential uncertainty for autoregressive modeling. To address these, we propose ReSID, a recommendation-native, principled SID framework that rethinks representation learning and quantization from the perspective of information preservation and sequential predictability, without relying on LLMs. ReSID consists of two components: (i) Field-Aware Masked Auto-Encoding (FAMAE), which learns predictive-sufficient item representations from structured features, and (ii) Globally Aligned Orthogonal Quantization (GAOQ), which produces compact and predictable SID sequences by jointly reducing semantic ambiguity and prefix-conditional uncertainty. Theoretical analysis and extensive experiments across ten datasets show the effectiveness of ReSID. ReSID consistently outperforms strong sequential and SID-based generative baselines by an average of over 10%, while reducing tokenization cost by up to 122x. Code is available at https://github.com/FuCongResearchSquad/ReSID. --- ### Building a Correct-by-Design Lakehouse. Data Contracts, Versioning, and Transactional Pipelines for Humans and Agents Authors: Weiming Sheng, Jinlang Wang, Manuel Barros, Aldrin Montana, Jacopo Tagliabue, Luca Bigon Published: 2026-02-02 Categories: cs.DC, cs.AI, cs.DB arXiv: https://arxiv.org/abs/2602.02335v3 Lakehouses are now the default substrate for analytics and AI, but they remain fragile under concurrent, untrusted change: schema mismatches often surface only at runtime, development and production easily diverge, and multi-table pipelines can expose partial results after failure. We present Bauplan, a code-first lakehouse that aims to eliminate a broad class of these failures by construction. Bauplan builds on a storage substrate that already provides atomic single-table snapshot evolution, and adds three pipeline-level correctness mechanisms: typed table contracts to make transformation boundaries checkable, Git-like data versioning to support reproducible collaboration and review, and transactional runs that guarantee atomic publication of an entire pipeline execution. We describe the system design, show how these abstractions fit together into a unified programming model for humans and agents, and report early results from a lightweight Alloy model that both validates key intuitions and exposes subtle counterexamples around transactional branch visibility. Our experience suggests that correctness in the lakehouse is best addressed not by patching failures after the fact, but by restricting the programming model so that many illegal states become unrepresentable. --- ### VQ-Style: Disentangling Style and Content in Motion with Residual Quantized Representations Authors: Fatemeh Zargarbashi, Dhruv Agrawal, Jakob Buhmann, Martin Guay, Stelian Coros, Robert W. Sumner Published: 2026-02-02 Categories: cs.CV, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2602.02334v2 Human motion data is inherently rich and complex, containing both semantic content and subtle stylistic features that are challenging to model. We propose a novel method for effective disentanglement of the style and content in human motion data to facilitate style transfer. Our approach is guided by the insight that content corresponds to coarse motion attributes while style captures the finer, expressive details. To model this hierarchy, we employ Residual Vector Quantized Variational Autoencoders (RVQ-VAEs) to learn a coarse-to-fine representation of motion. We further enhance the disentanglement by integrating codebook learning with contrastive learning and a novel information leakage loss to organize the content and the style across different codebooks. We harness this disentangled representation using our simple and effective inference-time technique Quantized Code Swapping, which enables motion style transfer without requiring any fine-tuning for unseen styles. Our framework demonstrates strong versatility across multiple inference applications, including style transfer, style removal, and motion blending. --- ### TTT-Parkour: Rapid Test-Time Training for Perceptive Robot Parkour Authors: Shaoting Zhu, Baijun Ye, Jiaxuan Wang, Jiakang Chen, Ziwen Zhuang, Linzhan Mou, Runhan Huang, Hang Zhao Published: 2026-02-02 Categories: cs.RO, cs.AI arXiv: https://arxiv.org/abs/2602.02331v1 Achieving highly dynamic humanoid parkour on unseen, complex terrains remains a challenge in robotics. Although general locomotion policies demonstrate capabilities across broad terrain distributions, they often struggle with arbitrary and highly challenging environments. To overcome this limitation, we propose a real-to-sim-to-real framework that leverages rapid test-time training (TTT) on novel terrains, significantly enhancing the robot's capability to traverse extremely difficult geometries. We adopt a two-stage end-to-end learning paradigm: a policy is first pre-trained on diverse procedurally generated terrains, followed by rapid fine-tuning on high-fidelity meshes reconstructed from real-world captures. Specifically, we develop a feed-forward, efficient, and high-fidelity geometry reconstruction pipeline using RGB-D inputs, ensuring both speed and quality during test-time training. We demonstrate that TTT-Parkour empowers humanoid robots to master complex obstacles, including wedges, stakes, boxes, trapezoids, and narrow beams. The whole pipeline of capturing, reconstructing, and test-time training requires less than 10 minutes on most tested terrains. Extensive experiments show that the policy after test-time training exhibits robust zero-shot sim-to-real transfer capability. --- ### Language Steering for Multilingual In-Context Learning Authors: Neeraja Kirtane, Kuan-Hao Huang Published: 2026-02-02 Categories: cs.CL arXiv: https://arxiv.org/abs/2602.02326v2 If large language models operate in a universal semantic space, then switching between languages should require only a simple activation offset. To test this, we take multilingual in-context learning as a case study, where few-shot demonstrations are provided in English but the test query is in a target language. We propose language vectors, computed as the mean activation difference between parallel source and target language examples at a particular layer, and added as an offset to hidden states at inference time to shift the model's internal representations toward the target language. We evaluate our method across three multilingual tasks spanning 19 languages and three models. Our results show consistent improvements on multilingual in-context learning over baselines across all tasks and languages tested, demonstrating that a simple activation offset is sufficient to redirect a model's language mode without any parameter updates. Beyond performance, the vectors encode interpretable linguistic structure, with closely related languages forming tight clusters and vectors transferring across tasks, suggesting that language identity occupies separable and structured directions in a model's activation space. --- ### MolLangData: A Large-Scale Dataset for Molecular Structure-Language Description via a Rule-Regularized Method Authors: Feiyang Cai, Guijuan He, Yi Hu, Jingjing Wang, Joshua Luo, Tianyu Zhu, Srikanth Pilla, Gang Li, Ling Liu, Feng Luo Published: 2026-02-02 Categories: cs.CL, cs.AI, q-bio.BM arXiv: https://arxiv.org/abs/2602.02320v5 Molecular function is largely determined by structure. Accurately aligning molecular structure with natural language is therefore essential for enabling large language models (LLMs) to both understand structure for chemical reasoning and generate molecules from natural-language design intent. However, the substantial cost of human annotation makes it infeasible to construct large-scale, high-quality datasets of structure-grounded descriptions. This work proposes a fully automated annotation framework for generating precise molecular descriptions at scale, such that the original molecule can be unambiguously reconstructed from the description alone. Our approach extends a rule-based chemical nomenclature parser to interpret IUPAC names and construct enriched, XML metadata that explicitly encodes molecular structure. This is then used to guide LLMs in producing accurate natural-language descriptions. Using this framework, we curate MolLangData, a dataset of approximately $163$k molecule--description pairs. A rigorous validation protocol combining expert human and LLM-based reconstruction on a subset of $2,000$ molecules demonstrates $98.6\%$ description precision. Using the curated dataset, we train a $4$B-parameter LLM via large-scale reinforcement learning for language-conditional molecule generation. The proposed framework and dataset provide a reliable foundation for molecule--language alignment, readily beneficial to broader chemical tasks. --- ### A Large-Scale Dataset for Molecular Structure-Language Description via a Rule-Regularized Method Authors: Feiyang Cai, Guijuan He, Yi Hu, Jingjing Wang, Joshua Luo, Tianyu Zhu, Srikanth Pilla, Gang Li, Ling Liu, Feng Luo Published: 2026-02-02 Categories: cs.CL, cs.AI, q-bio.BM arXiv: https://arxiv.org/abs/2602.02320v4 Molecular function is largely determined by structure. Accurately aligning molecular structure with natural language is therefore essential for enabling large language models (LLMs) to reason about downstream chemical tasks. However, the substantial cost of human annotation makes it infeasible to construct large-scale, high-quality datasets of structure-grounded descriptions. In this work, we propose a fully automated annotation framework for generating precise molecular descriptions that preserve complete structural details at scale. Our approach builds upon and extends a rule-based chemical nomenclature parser to interpret IUPAC names and construct enriched, structural XML metadata that explicitly encodes molecular structure. This metadata is then used to guide LLMs in producing accurate natural-language descriptions. Using this framework, we curate a large-scale dataset of approximately $163$k molecule--description pairs. A rigorous validation protocol combining LLM-based and expert human evaluation on a subset of $2,000$ molecules demonstrates a high description precision of $98.6$%. The proposed annotation framework is readily beneficial to broader chemical tasks that rely on structural descriptions, with the resulting dataset providing a reliable foundation for molecule--language alignment. The source code and dataset are hosted at https://github.com/TheLuoFengLab/MolLangData and https://huggingface.co/datasets/ChemFM/MolLangData, respectively. --- ### Enhancing Indoor Occupancy Prediction via Sparse Query-Based Multi-Level Consistent Knowledge Distillation Authors: Xiang Li, Yupeng Zheng, Pengfei Li, Yilun Chen, Ya-Qin Zhang, Wenchao Ding Published: 2026-02-02 Categories: cs.CV arXiv: https://arxiv.org/abs/2602.02318v1 Occupancy prediction provides critical geometric and semantic understanding for robotics but faces efficiency-accuracy trade-offs. Current dense methods suffer computational waste on empty voxels, while sparse query-based approaches lack robustness in diverse and complex indoor scenes. In this paper, we propose DiScene, a novel sparse query-based framework that leverages multi-level distillation to achieve efficient and robust occupancy prediction. In particular, our method incorporates two key innovations: (1) a Multi-level Consistent Knowledge Distillation strategy, which transfers hierarchical representations from large teacher models to lightweight students through coordinated alignment across four levels, including encoder-level feature alignment, query-level feature matching, prior-level spatial guidance, and anchor-level high-confidence knowledge transfer and (2) a Teacher-Guided Initialization policy, employing optimized parameter warm-up to accelerate model convergence. Validated on the Occ-Scannet benchmark, DiScene achieves 23.2 FPS without depth priors while outperforming our baseline method, OPUS, by 36.1% and even better than the depth-enhanced version, OPUS†. With depth integration, DiScene† attains new SOTA performance, surpassing EmbodiedOcc by 3.7% with 1.62$\times$ faster inference speed. Furthermore, experiments on the Occ3D-nuScenes benchmark and in-the-wild scenarios demonstrate the versatility of our approach in various environments. Code and models can be accessed at https://github.com/getterupper/DiScene. --- ### The Shape of Beliefs: Geometry, Dynamics, and Interventions along Representation Manifolds of Language Models' Posteriors Authors: Raphaël Sarfati, Eric Bigelow, Daniel Wurgaft, Siddharth Boppana, Jack Merullo, Atticus Geiger, Owen Lewis, Tom McGrath, Ekdeep Singh Lubana Published: 2026-02-02 Categories: cs.CL arXiv: https://arxiv.org/abs/2602.02315v2 Large language models (LLMs) form implicit beliefs (posteriors over latent variables) from prompts, but we lack a mechanistic account of how these beliefs are encoded in representation space, how they update with new evidence, and how interventions reshape them. We study a controlled setting in which Llama-3.2 infers the parameters of a normal distribution from in-context samples. We show that parameter posteriors are encoded as curved manifolds in representation space, and trace how they evolve along the prompt. Standard linear steering moves representations off-manifold, inducing unintended, coupled changes, whereas geometry-aware methods preserve the target belief family. Our work demonstrates an example of linear field probing (LFP) as a principled approach to tile the data manifold and make interventions that respect the underlying geometry. Our results suggest that LLM beliefs are inherently geometric objects, and that globally linear representations are often inadequate abstractions. --- ### Interpreting and Controlling LLM Reasoning through Integrated Policy Gradient Authors: Changming Li, Kaixing Zhang, Haoyun Xu, Yingdong Shi, Zheng Zhang, Kaitao Song, Kan Ren Published: 2026-02-02 Categories: cs.AI, cs.CL, cs.LG arXiv: https://arxiv.org/abs/2602.02313v2 Large language models (LLMs) demonstrate strong reasoning abilities in solving complex real-world problems. Yet, the internal mechanisms driving these complex reasoning behaviors remain opaque. Existing interpretability approaches targeting reasoning either identify components (e.g., neurons) correlated with special textual patterns, or rely on human-annotated contrastive pairs to derive control vectors. Consequently, current methods struggle to precisely localize complex reasoning mechanisms or capture sequential influence from model internal workings to the reasoning outputs. In this paper, built on outcome-oriented and sequential-influence-aware principles, we focus on identifying components that have sequential contribution to reasoning behavior where outcomes are cumulated by long-range effects. We propose Integrated Policy Gradient (IPG), a novel framework that attributes reasoning behaviors to model's inner components by propagating compound outcome-based signals such as post reasoning accuracy backward through model inference trajectories. Empirical evaluations demonstrate that our approach achieves more precise localization and enables reliable modulation of reasoning behaviors (e.g., reasoning capability, reasoning strength) across diverse reasoning models. --- ### Introns and Templates Matter: Rethinking Linkage in GP-GOMEA Authors: Johannes Koch, Tanja Alderliesten, Peter A. N. Bosman Published: 2026-02-02 Categories: cs.NE arXiv: https://arxiv.org/abs/2602.02311v1 GP-GOMEA is among the state-of-the-art for symbolic regression, especially when it comes to finding small and potentially interpretable solutions. A key mechanism employed in any GOMEA variant is the exploitation of linkage, the dependencies between variables, to ensure efficient evolution. In GP-GOMEA, mutual information between node positions in GP trees has so far been used to learn linkage. For this, a fixed expression template is used. This however leads to introns for expressions smaller than the full template. As introns have no impact on fitness, their occurrences are not directly linked to selection. Consequently, introns can adversely affect the extent to which mutual information captures dependencies between tree nodes. To overcome this, we propose two new measures for linkage learning, one that explicitly considers introns in mutual information estimates, and one that revisits linkage learning in GP-GOMEA from a grey-box perspective, yielding a measure that needs not to be learned from the population but is derived directly from the template. Across five standard symbolic regression problems, GP-GOMEA achieves substantial improvements using both measures. We also find that the newly learned linkage structure closely reflects the template linkage structure, and that explicitly using the template structure yields the best performance overall. --- ### FragmentFlow: Scalable Transition State Generation for Large Molecules Authors: Ron Shprints, Peter Holderrieth, Juno Nam, Rafael Gómez-Bombarelli, Tommi Jaakkola Published: 2026-02-02 Categories: physics.chem-ph, cs.AI arXiv: https://arxiv.org/abs/2602.02310v2 Transition states (TSs) are central to understanding and quantitatively predicting chemical reactivity and reaction mechanisms. Although traditional TS generation methods are computationally expensive, recent generative modeling approaches have enabled chemically meaningful TS prediction for relatively small molecules. However, these methods fail to generalize to practically relevant reaction substrates because of distribution shifts induced by increasing molecular sizes. Furthermore, TS geometries for larger molecules are not available at scale, making it infeasible to train generative models from scratch on such molecules. To address these challenges, we introduce FragmentFlow: a divide-and-conquer approach that trains a generative model to predict TS geometries for the reactive core atoms, which define the reaction mechanism. The full TS structure is then reconstructed by re-attaching substituent fragments to the predicted core. By operating on reactive cores, whose size and composition remain relatively invariant across molecular contexts, FragmentFlow mitigates distribution shifts in generative modeling. Evaluated on a new curated dataset of reactions involving reactants with up to 33 heavy atoms, FragmentFlow correctly identifies 90% of TSs while requiring 30% fewer saddle-point optimization steps than classical initialization schemes. These results point toward scalable TS generation for high-throughput reactivity studies. --- ### Spark: Modular Spiking Neural Networks Authors: Mario Franco, Carlos Gershenson Published: 2026-02-02 Categories: cs.NE, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2602.02306v3 Nowadays, neural networks act as a synonym for artificial intelligence. Present neural network models, although remarkably powerful, are inefficient both in terms of data and energy. Several alternative forms of neural networks have been proposed to address some of these problems. Specifically, spiking neural networks are suitable for efficient hardware implementations. However, effective learning algorithms for spiking networks remain elusive, although it is suspected that effective plasticity mechanisms could alleviate the problem of data efficiency. Here, we present a new framework for spiking neural networks - Spark - built upon the idea of modular design, from simple components to entire models. The aim of this framework is to provide an efficient and streamlined pipeline for spiking neural networks. We showcase this framework by solving the sparse-reward cartpole problem with simple plasticity mechanisms. We hope that a framework compatible with traditional ML pipelines may accelerate research in the area, specifically for continuous and unbatched learning, akin to the one animals exhibit. --- ### Comparing Explanations is Not Enough, Explain the Change: New Standards are Needed to Explain Behavioral Shifts in Large Language Models Authors: Martino Ciaperoni, Marzio Di Vece, Roberto Pellungrini, Luca Pappalardo, Fosca Giannotti, Francesco Giannini Published: 2026-02-02 Categories: cs.AI, cs.LG arXiv: https://arxiv.org/abs/2602.02304v3 Large-scale foundation models exhibit behavioral shifts when subjected to interventions such as scaling, fine-tuning, reinforcement learning with human feedback, or in-context learning. Current explainability methods are structurally ill-suited to explain these shifts, because they either treat models as static objects, as traditional eXplainable AI (XAI) approaches do, or merely compare independent explanations across different checkpoints of a model. As a result, these approaches fail to explain the functional transition between two model instances in which a certain behavior has shifted following an intervention. This gap creates significant governance risks across jurisdictions including the EU AI Act, US state legislation, and Chinese AI regulations, which require documenting causal chains for substantial system modifications. This position paper argues that explaining behavioral shifts in large language models requires a principled approach that treats the shift itself as the primary object of explanation: namely, one that explains how and why an intervention transforms a reference model into an updated model with different behavior. To support this claim, we introduce Comparative XAI (XAI$_Δ$), a novel XAI paradigm aimed at explaining the difference between two model checkpoints where a behavior has shifted, together with a set of desiderata specifying what XAI$_Δ$ explainers and explanations must satisfy, including comparability, validity, actionability, and monitoring, with the goal of grounding model auditing in explicit, measurable requirements. Finally, we provide preliminary evidence suggesting the need for XAI$_Δ$ in practice through illustrative experiments, compiling the resulting findings into a transition report directly usable for governance and incident documentation. --- ### Comparing Explanations is Not Enough, Explain the Change: New Standards are Needed to Explain Behavioral Shifts in Large Language Models Authors: Martino Ciaperoni, Marzio Di Vece, Roberto Pellungrini, Luca Pappalardo, Fosca Giannotti, Francesco Giannini Published: 2026-02-02 Categories: cs.AI, cs.LG arXiv: https://arxiv.org/abs/2602.02304v2 Large-scale foundation models exhibit \emph{behavioral shifts} when subjected to interventions such as scaling, fine-tuning, reinforcement learning with human feedback, or in-context learning. Current explainability methods are structurally ill-suited to explain these shifts, because they either treat models as static objects, as traditional eXplainable AI (XAI) approaches do, or merely compare independent explanations across different checkpoints of a model. As a result, these approaches fail to explain the functional transition between two model instances in which a certain behavior has shifted following an intervention. This gap creates significant governance risks across jurisdictions including the EU AI Act, US state legislation, and Chinese AI regulations, which require documenting causal chains for substantial system modifications. This position paper argues that explaining behavioral shifts in large language models requires a principled approach that treats the shift itself as the primary object of explanation: namely, one that explains how and why an intervention transforms a reference model into an updated model with different behavior. To support this claim, we introduce \textit{Comparative} XAI (XAI$_Δ$), a novel XAI paradigm aimed at explaining the difference between two model checkpoints where a behavior has shifted, together with a set of desiderata specifying what XAI$_Δ$ explainers and explanations must satisfy, including comparability, validity, actionability, and monitoring, with the goal of grounding model auditing in explicit, measurable requirements. Finally, we provide preliminary evidence suggesting the need for XAI$_Δ$ in practice through illustrative experiments, compiling the resulting findings into a transition report directly usable for governance and incident documentation. --- ### Advancing General-Purpose Reasoning Models with Modular Gradient Surgery Authors: Min Cai, Yu Liang, Longzheng Wang, Yan Wang, Yueyang Zhang, Long Xia, Zhiyuan Sun, Xi Ye, Daiting Shi Published: 2026-02-02 Categories: cs.CL, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2602.02301v2 Reinforcement learning (RL) has played a central role in recent advances in large reasoning models (LRMs), yielding strong gains in verifiable and open-ended reasoning. However, training a single general-purpose LRM across diverse domains remains challenging due to pronounced domain heterogeneity. Through a systematic study of two widely used strategies, Sequential RL and Mixed RL, we find that both incur substantial cross-domain interference at the behavioral and gradient levels, resulting in limited overall gains. To address these challenges, we introduce **M**odular **G**radient **S**urgery (**MGS**), which resolves gradient conflicts at the module level within the transformer. When applied to Llama and Qwen models, MGS achieves average improvements of 4.3 (16.6\%) and 4.5 (11.1\%) points, respectively, over standard multi-task RL across three representative domains (math, general chat, and instruction following). Further analysis demonstrates that MGS remains effective under prolonged training. Overall, our study clarifies the sources of interference in multi-domain RL and presents an effective solution for training general-purpose LRMs. --- ### Decoupling Generalizability and Membership Privacy Risks in Neural Networks Authors: Xingli Fang, Jung-Eun Kim Published: 2026-02-02 Categories: cs.LG, cs.AI, cs.CR arXiv: https://arxiv.org/abs/2602.02296v2 A deep learning model usually has to sacrifice some utilities when it acquires some other abilities or characteristics. Privacy preservation has such trade-off relationships with utilities. The loss disparity between various defense approaches implies the potential to decouple generalizability and privacy risks to maximize privacy gain. In this paper, we identify that the model's generalization and privacy risks exist in different regions in deep neural network architectures. Based on the observations that we investigate, we propose Privacy-Preserving Training Principle (PPTP) to protect model components from privacy risks while minimizing the loss in generalizability. Through extensive evaluations, our approach shows significantly better maintenance in model generalizability while enhancing privacy preservation. --- ### EvalQReason: A Framework for Step-Level Reasoning Evaluation in Large Language Models Authors: Shaima Ahmad Freja, Ferhat Ozgur Catak, Betul Yurdem, Chunming Rong Published: 2026-02-02 Categories: cs.LG arXiv: https://arxiv.org/abs/2602.02295v1 Large Language Models (LLMs) are increasingly deployed in critical applications requiring reliable reasoning, yet their internal reasoning processes remain difficult to evaluate systematically. Existing methods focus on final-answer correctness, providing limited insight into how reasoning unfolds across intermediate steps. We present EvalQReason, a framework that quantifies LLM reasoning quality through step-level probability distribution analysis without requiring human annotation. The framework introduces two complementary algorithms: Consecutive Step Divergence (CSD), which measures local coherence between adjacent reasoning steps, and Step-to-Final Convergence (SFC), which assesses global alignment with final answers. Each algorithm employs five statistical metrics to capture reasoning dynamics. Experiments across mathematical and medical datasets with open-source 7B-parameter models demonstrate that CSD-based features achieve strong predictive performance for correctness classification, with classical machine learning models reaching F1=0.78 and ROC-AUC=0.82, and sequential neural models substantially improving performance (F1=0.88, ROC-AUC=0.97). CSD consistently outperforms SFC, and sequential architectures outperform classical machine learning approaches. Critically, reasoning dynamics prove domain-specific: mathematical reasoning exhibits clear divergence-based discrimination patterns between correct and incorrect solutions, while medical reasoning shows minimal discriminative signals, revealing fundamental differences in how LLMs process different reasoning types. EvalQReason enables scalable, process-aware evaluation of reasoning reliability, establishing probability-based divergence analysis as a principled approach for trustworthy AI deployment. --- ### Hallucination or Creativity: How to Evaluate AI-Generated Scientific Stories? Authors: Alex Argese, Pasquale Lisena, Raphaël Troncy Published: 2026-02-02 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2602.02290v2 Generative AI can turn scientific articles into narratives for diverse audiences, but evaluating these stories remains challenging. Storytelling demands abstraction, simplification, and pedagogical creativity-qualities that are not often well-captured by standard summarization metrics. Meanwhile, factual hallucinations are critical in scientific contexts, yet, detectors often misclassify legitimate narrative reformulations or prove unstable when creativity is involved. In this work, we propose StoryScore, a composite metric for evaluating AI-generated scientific stories. StoryScore integrates semantic alignment, lexical grounding, narrative control, structural fidelity, redundancy avoidance, and entity-level hallucination detection into a unified framework. Our analysis also reveals why many hallucination detection methods fail to distinguish pedagogical creativity from factual errors, highlighting a key limitation: while automatic metrics can effectively assess semantic similarity with original content, they struggle to evaluate how it is narrated and controlled. --- ### AROpt: An Optimization Method for Autoregressive Time Series Forecasting Authors: Zheng Li, Jerry Cheng, Huanying Gu Published: 2026-02-02 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2602.02288v3 Current time-series forecasting models are primarily based on transformer-style neural networks. These models achieve long-term forecasting mainly by scaling up the model size rather than through genuinely autoregressive (AR) rollout. From the perspective of large language model training, traditional time-series forecasting model training ignores the monotonic error-growth heuristic. In this paper, we propose a novel training method for time-series forecasting that enforces two key properties: (1) AR prediction errors should increase with the forecasting horizon. Violations of this trend are interpreted as rollout inconsistency and are softly penalized during training, and (2) the method enables models to be able to concatenate short-term AR predictions to form flexible long-term forecasts. Empirical results demonstrate that our method establishes a new state-of-the-art across multiple benchmarks, achieving an MSE reduction of more than $10\%$ compared to iTransformer and other recent strong baselines. Furthermore, it enables short-horizon forecasting models to perform reliable long-term predictions at horizons over 7.5 times longer. Code is available at https://github.com/LizhengMathAi/AROpt --- ### Cross-Lingual Stability of LLM Judges Under Controlled Generation: Evidence from Finno-Ugric Languages Authors: Isaac Chung, Linda Freienthal Published: 2026-02-02 Categories: cs.CL arXiv: https://arxiv.org/abs/2602.02287v1 Cross-lingual evaluation of large language models (LLMs) typically conflates two sources of variance: genuine model performance differences and measurement instability. We investigate evaluation reliability by holding generation conditions constant while varying target language. Using synthetic customer-support dialogues generated with identical parameters across Estonian, Finnish, and Hungarian, we test whether automatic metrics and LLM-as-a-judge scoring produce stable model rankings across these morphologically rich, related Finno-Ugric languages. With a small set of Estonian native speaker annotations as a reference point, we find systematic ranking instabilities: surface-level metrics (lexical diversity, surface and semantic similarity) maintain cross-language stability, but pragmatic judgments (coherence, instruction-following) exhibit rank inversions and near-zero correlations. Because generation is controlled, these inconsistencies reflect how judge scoring behaves differently across languages rather than true model differences. This controlled design provides a diagnostic probe: evaluation methods that fail to maintain stability under identical generation conditions signal transfer failure before deployment. Our findings suggest that zero-shot judge transfer is unreliable for discourse-level assessment in morphologically rich languages, motivating language-specific calibration against targeted human baselines. We release our controlled generation protocol, synthetic data, and evaluation framework to enable replication across language families at https://github.com/isaac-chung/cross-lingual-stability-judges. --- ### DFKI-Speech System for WildSpoof Challenge: A robust framework for SASV In-the-Wild Authors: Arnab Das, Yassine El Kheir, Enes Erdem Erdogan, Feidi Kallel, Tim Polzehl, Sebastian Moeller Published: 2026-02-02 Categories: cs.SD, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2602.02286v1 This paper presents the DFKI-Speech system developed for the WildSpoof Challenge under the Spoofing aware Automatic Speaker Verification (SASV) track. We propose a robust SASV framework in which a spoofing detector and a speaker verification (SV) network operate in tandem. The spoofing detector employs a self-supervised speech embedding extractor as the frontend, combined with a state-of-the-art graph neural network backend. In addition, a top-3 layer based mixture-of-experts (MoE) is used to fuse high-level and low-level features for effective spoofed utterance detection. For speaker verification, we adapt a low-complexity convolutional neural network that fuses 2D and 1D features at multiple scales, trained with the SphereFace loss. Additionally, contrastive circle loss is applied to adaptively weight positive and negative pairs within each training batch, enabling the network to better distinguish between hard and easy sample pairs. Finally, fixed imposter cohort based AS Norm score normalization and model ensembling are used to further enhance the discriminative capability of the speaker verification system. --- ### AI4SLT: Empirical Processes in Lean 4 for Formal Statistical Learning Theory Authors: Yuanhe Zhang, Jason D. Lee, Fanghui Liu Published: 2026-02-02 Categories: cs.LG, cs.CL, math.ST arXiv: https://arxiv.org/abs/2602.02285v2 We present the first comprehensive Lean 4 formalization of statistical learning theory (SLT) grounded in empirical process theory. Our en-to-end formal infrastructure implement the missing contents in latest Lean library, including a complete development of Gaussian Lipschitz concentration, Dudley's entropy integral theorem for sub-Gaussian processes, and an application to least-squares (sparse) regression with a sharp rate. The project was carried out using a human-AI collaborative workflow, in which humans design proof strategies and AI agents execute tactical proof construction, leading to the human-verified Lean 4 toolbox for SLT. Beyond implementation, the formalization process exposes and resolves implicit assumptions and missing details in standard SLT textbooks, enforcing a granular, line-by-line understanding of the theory. This work establishes a reusable formal foundation and opens the door for future developments in machine learning theory. The code is provided in https://github.com/YuanheZ/lean-stat-learning-theory. --- ### Choice-Model-Assisted Q-learning for Delayed-Feedback Revenue Management Authors: Owen Shen, Patrick Jaillet Published: 2026-02-02 Categories: cs.LG, stat.ML arXiv: https://arxiv.org/abs/2602.02283v1 We study reinforcement learning for revenue management with delayed feedback, where a substantial fraction of value is determined by customer cancellations and modifications observed days after booking. We propose \emph{choice-model-assisted RL}: a calibrated discrete choice model is used as a fixed partial world model to impute the delayed component of the learning target at decision time. In the fixed-model deployment regime, we prove that tabular Q-learning with model-imputed targets converges to an $O(\varepsilon/(1-γ))$ neighborhood of the optimal Q-function, where $\varepsilon$ summarizes partial-model error, with an additional $O(t^{-1/2})$ sampling term. Experiments in a simulator calibrated from 61{,}619 hotel bookings (1{,}088 independent runs) show: (i) no statistically detectable difference from a maturity-buffer DQN baseline in stationary settings; (ii) positive effects under in-family parameter shifts, with significant gains in 5 of 10 shift scenarios after Holm--Bonferroni correction (up to 12.4\%); and (iii) consistent degradation under structural misspecification, where the choice model assumptions are violated (1.4--2.6\% lower revenue). These results characterize when partial behavioral models improve robustness under shift and when they introduce harmful bias. --- ### MoLF: Mixture-of-Latent-Flow for Pan-Cancer Spatial Gene Expression Prediction from Histology Authors: Susu Hu, Stefanie Speidel Published: 2026-02-02 Categories: cs.LG arXiv: https://arxiv.org/abs/2602.02282v3 Inferring spatial transcriptomics (ST) from histology enables scalable histogenomic profiling, yet current methods are largely restricted to single-tissue models. This fragmentation fails to leverage biological principles shared across cancer types and hinders application to data-scarce scenarios. While pan-cancer training offers a solution, the resulting heterogeneity challenges monolithic architectures. To bridge this gap, we introduce MoLF (Mixture-of-Latent-Flow), a generative model for pan-cancer histogenomic prediction. MoLF leverages a conditional Flow Matching objective to map noise to the gene latent manifold, parameterized by a Mixture-of-Experts (MoE) velocity field. By dynamically routing inputs to specialized sub-networks, this architecture effectively decouples the optimization of diverse tissue patterns. Our experiments demonstrate that MoLF establishes a new state-of-the-art, consistently outperforming both specialized and foundation model baselines on pan-cancer benchmarks. Furthermore, MoLF exhibits zero-shot generalization to cross-species data, suggesting it captures fundamental, conserved histo-molecular mechanisms. --- ### A Physical Theory of Backpropagation: Exact Gradients from the Least-Action Principle Authors: Antonino Emanuele Scurria Published: 2026-02-02 Categories: cs.LG, cs.AI, cs.NE, physics.class-ph, physics.comp-ph arXiv: https://arxiv.org/abs/2602.02281v2 Backpropagation is typically presented as a symbolic procedure: a backward pass topologically distinct from inference, with non-local error signals and synchronous global clocking, features with no clear analog in physical reality. Existing physics-inspired alternatives recover gradients only approximately, in vanishing-perturbation limits, or under weight-symmetry constraints incompatible with feedforward architectures. In this paper, we address this gap by deriving exact backpropagation from Hamilton's least-action principle. By recasting the forward dynamics in continuous time and adapting a Lagrangian formalism for non-conservative systems to the resulting flow, we unify inference and gradient computation within a single variational framework on a doubled phase space, whose two conjugate fields jointly encode activations and sensitivities. A single global Lagrangian governs the dynamics: the task loss enters as a symmetry-breaking perturbation of the forward manifold, and credit assignment emerges as the tension that develops between the conjugate states. Inference and gradient computation thus unfold simultaneously through local interactions, requiring no separate backward circuit. Ultimately, standard backpropagation is recovered exactly as the discrete-time projection of this continuous flow. This perspective unifies the formalism of physics with backpropagation, opening a principled pathway for applying tools from classical mechanics - symplectic geometry, Noether's theorem, path-integral methods - to the analysis of learning dynamics. As a downstream consequence, it also points toward analog and neuromorphic substrates in which learning is embodied in the hardware itself. --- ### RACC: Representation-Aware Coverage Criteria for LLM Safety Testing Authors: Zeming Wei, Zhixin Zhang, Chengcan Wu, Yihao Zhang, Xiaokun Luan, Meng Sun Published: 2026-02-02 Categories: cs.SE, cs.AI, cs.CL, cs.CR, cs.LG arXiv: https://arxiv.org/abs/2602.02280v2 Large Language Models (LLMs) face severe safety risks from jailbreak attacks, yet current safety testing largely relies on static datasets and lacks systematic criteria to evaluate test suite quality and adequacy. While coverage criteria have proven effective for smaller neural networks, they are impractical for LLMs due to computational overhead and the entanglement of safety-critical signals with irrelevant neuron activations. To address these issues, we propose RACC (Representation-Aware Coverage Criteria), a set of coverage criteria specialized for LLM safety testing. RACC first extracts safety representations from the LLM's hidden states using a small calibration set of harmful prompts, then measures test prompts' concept activations against these directions, and finally computes coverage through six criteria assessing both individual and compositional safety concept coverage. Experiments on multiple LLMs and safety benchmarks show that RACC reliably rewards high-quality jailbreak test suites while remaining insensitive to redundant or invalid inputs, which is a key distinction that neuron-level criteria fail to make. We further demonstrate RACC's practical value in two applications, including test suite prioritization and attack prompt sampling, and validate its generalization across diverse settings and configurations. Overall, RACC provides a scalable and principled foundation for coverage-guided LLM safety testing. --- ### Kimi K2.5: Visual Agentic Intelligence Authors: Kimi Team, Tongtong Bai, Yifan Bai, Yiping Bao, S. H. Cai, Yuan Cao, Ziwei Chai, Y. Charles, H. S. Che, Cheng Chen, Guanduo Chen, Huarong Chen et al. Published: 2026-02-02 Categories: cs.CL, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2602.02276v2 We introduce Kimi K2.5, an open-source multimodal agentic model designed to advance general agentic intelligence. K2.5 emphasizes the joint optimization of text and vision so that two modalities enhance each other. This includes a series of techniques such as joint text-vision pre-training, zero-vision SFT, and joint text-vision reinforcement learning. Building on this multimodal foundation, K2.5 introduces Agent Swarm, a self-directed parallel agent orchestration framework that dynamically decomposes complex tasks into heterogeneous sub-problems and executes them concurrently. Extensive evaluations show that Kimi K2.5 achieves state-of-the-art results across various domains including coding, vision, reasoning, and agentic tasks. Agent Swarm also reduces latency by up to $4.5\times$ over single-agent baselines. We release the post-trained Kimi K2.5 model checkpoint to facilitate future research and real-world applications of agentic intelligence. --- ### Kimi K2.5: Visual Agentic Intelligence Authors: Kimi Team, Tongtong Bai, Yifan Bai, Yiping Bao, S. H. Cai, Yuan Cao, Y. Charles, H. S. Che, Cheng Chen, Guanduo Chen, Huarong Chen, Jia Chen et al. Published: 2026-02-02 Categories: cs.CL, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2602.02276v1 We introduce Kimi K2.5, an open-source multimodal agentic model designed to advance general agentic intelligence. K2.5 emphasizes the joint optimization of text and vision so that two modalities enhance each other. This includes a series of techniques such as joint text-vision pre-training, zero-vision SFT, and joint text-vision reinforcement learning. Building on this multimodal foundation, K2.5 introduces Agent Swarm, a self-directed parallel agent orchestration framework that dynamically decomposes complex tasks into heterogeneous sub-problems and executes them concurrently. Extensive evaluations show that Kimi K2.5 achieves state-of-the-art results across various domains including coding, vision, reasoning, and agentic tasks. Agent Swarm also reduces latency by up to $4.5\times$ over single-agent baselines. We release the post-trained Kimi K2.5 model checkpoint to facilitate future research and real-world applications of agentic intelligence. --- ### dziribot: rag based intelligent conversational agent for algerian arabic dialect Authors: El Batoul Bechiri, Dihia Lanasri Published: 2026-02-02 Categories: cs.CL arXiv: https://arxiv.org/abs/2602.02270v1 The rapid digitalization of customer service has intensified the demand for conversational agents capable of providing accurate and natural interactions. In the Algerian context, this is complicated by the linguistic complexity of Darja, a dialect characterized by non-standardized orthography, extensive code-switching with French, and the simultaneous use of Arabic and Latin (Arabizi) scripts. This paper introduces DziriBOT, a hybrid intelligent conversational agent specifically engineered to overcome these challenges. We propose a multi-layered architecture that integrates specialized Natural Language Understanding (NLU) with Retrieval-Augmented Generation (RAG), allowing for both structured service flows and dynamic, knowledge-intensive responses grounded in curated enterprise documentation. To address the low-resource nature of Darja, we systematically evaluate three distinct approaches: a sparse-feature Rasa pipeline, classical machine learning baselines, and transformer-based fine-tuning. Our experimental results demonstrate that the fine-tuned DziriBERT model achieves state-of-the-art performance. These results significantly outperform traditional baselines, particularly in handling orthographic noise and rare intents. Ultimately, DziriBOT provides a robust, scalable solution that bridges the gap between formal language models and the linguistic realities of Algerian users, offering a blueprint for dialect-aware automation in the regional market. --- ### Bridging the Sim-to-Real Gap with multipanda ros2: A Real-Time ROS2 Framework for Multimanual Systems Authors: Jon Škerlj, Seongjin Bien, Abdeldjallil Naceri, Sami Haddadin Published: 2026-02-02 Categories: cs.RO, cs.AI, cs.SE, eess.SY arXiv: https://arxiv.org/abs/2602.02269v1 We present $multipanda\_ros2$, a novel open-source ROS2 architecture for multi-robot control of Franka Robotics robots. Leveraging ros2 control, this framework provides native ROS2 interfaces for controlling any number of robots from a single process. Our core contributions address key challenges in real-time torque control, including interaction control and robot-environment modeling. A central focus of this work is sustaining a 1kHz control frequency, a necessity for real-time control and a minimum frequency required by safety standards. Moreover, we introduce a controllet-feature design pattern that enables controller-switching delays of $\le 2$ ms, facilitating reproducible benchmarking and complex multi-robot interaction scenarios. To bridge the simulation-to-reality (sim2real) gap, we integrate a high-fidelity MuJoCo simulation with quantitative metrics for both kinematic accuracy and dynamic consistency (torques, forces, and control errors). Furthermore, we demonstrate that real-world inertial parameter identification can significantly improve force and torque accuracy, providing a methodology for iterative physics refinement. Our work extends approaches from soft robotics to rigid dual-arm, contact-rich tasks, showcasing a promising method to reduce the sim2real gap and providing a robust, reproducible platform for advanced robotics research. --- ### HopFormer: Sparse Graph Transformers with Explicit Receptive Field Control Authors: Sanggeon Yun, Raheeb Hassan, Ryozo Masukawa, Sungheon Jeong, Mohsen Imani Published: 2026-02-02 Categories: cs.LG arXiv: https://arxiv.org/abs/2602.02268v1 Graph Transformers typically rely on explicit positional or structural encodings and dense global attention to incorporate graph topology. In this work, we show that neither is essential. We introduce HopFormer, a graph Transformer that injects structure exclusively through head-specific n-hop masked sparse attention, without the use of positional encodings or architectural modifications. This design provides explicit and interpretable control over receptive fields while enabling genuinely sparse attention whose computational cost scales linearly with mask sparsity. Through extensive experiments on both node-level and graph-level benchmarks, we demonstrate that our approach achieves competitive or superior performance across diverse graph structures. Our results further reveal that dense global attention is often unnecessary: on graphs with strong small-world properties, localized attention yields more stable and consistently high performance, while on graphs with weaker small-world effects, global attention offers diminishing returns. Together, these findings challenge prevailing assumptions in graph Transformer design and highlight sparsity-controlled attention as a principled and efficient alternative. --- ### OpenSeal: Good, Fast, and Cheap Construction of an Open-Source Southeast Asian LLM via Parallel Data Authors: Tan Sang Nguyen, Muhammad Reza Qorib, Hwee Tou Ng Published: 2026-02-02 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2602.02266v1 Large language models (LLMs) have proven to be effective tools for a wide range of natural language processing (NLP) applications. Although many LLMs are multilingual, most remain English-centric and perform poorly on low-resource languages. Recently, several Southeast Asia-focused LLMs have been developed, but none are truly open source, as they do not publicly disclose their training data. Truly open-source models are important for transparency and for enabling a deeper and more precise understanding of LLM internals and development, including biases, generalization, and multilinguality. Motivated by recent advances demonstrating the effectiveness of parallel data in improving multilingual performance, we conduct controlled and comprehensive experiments to study the effectiveness of parallel data in continual pretraining of LLMs. Our findings show that using only parallel data is the most effective way to extend an LLM to new languages. Using just 34.7B tokens of parallel data and 180 hours on 8x NVIDIA H200 GPUs, we built OpenSeal, the first truly open Southeast Asian LLM that rivals the performance of existing models of similar size. --- ### Unsupervised Physics-Informed Operator Learning through Multi-Stage Curriculum Training Authors: Paolo Marcandelli, Natansh Mathur, Stefano Markidis, Martina Siena, Stefano Mariani Published: 2026-02-02 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2602.02264v1 Solving partial differential equations remains a central challenge in scientific machine learning. Neural operators offer a promising route by learning mappings between function spaces and enabling resolution-independent inference, yet they typically require supervised data. Physics-informed neural networks address this limitation through unsupervised training with physical constraints but often suffer from unstable convergence and limited generalization capability. To overcome these issues, we introduce a multi-stage physics-informed training strategy that achieves convergence by progressively enforcing boundary conditions in the loss landscape and subsequently incorporating interior residuals. At each stage the optimizer is re-initialized, acting as a continuation mechanism that restores stability and prevents gradient stagnation. We further propose the Physics-Informed Spline Fourier Neural Operator (PhIS-FNO), combining Fourier layers with Hermite spline kernels for smooth residual evaluation. Across canonical benchmarks, PhIS-FNO attains a level of accuracy comparable to that of supervised learning, using labeled information only along a narrow boundary region, establishing staged, spline-based optimization as a robust paradigm for physics-informed operator learning. --- ### OmniCode: A Benchmark for Evaluating Software Engineering Agents Authors: Atharv Sonwane, Eng-Shen Tu, Wei-Chung Lu, Claas Beger, Carter Larsen, Debjit Dhar, Simon Alford, Rachel Chen, Ronit Pattanayak, Tuan Anh Dang, Guohao Chen, Gloria Geng et al. Published: 2026-02-02 Categories: cs.SE, cs.AI, cs.CL arXiv: https://arxiv.org/abs/2602.02262v3 LLM-powered coding agents are redefining how real-world software is developed. To drive the research towards better coding agents, we require challenging benchmarks that can rigorously evaluate the ability of such agents to perform various software engineering tasks. However, popular coding benchmarks such as HumanEval and SWE-Bench focus on narrowly scoped tasks such as competition programming and patch generation. In reality, software engineers have to handle a broader set of tasks for real-world software development. To address this gap, we propose OmniCode, a novel software engineering benchmark that contains a broader and more diverse set of task categories beyond code or patch generation. Overall, OmniCode contains 1794 tasks spanning three programming languages - Python, Java, and C++ - and four key categories: bug fixing, test generation, code review fixing, and style fixing. In contrast to prior software engineering benchmarks, the tasks in OmniCode are (1) manually validated to eliminate ill-defined problems, and (2) synthetically crafted or recently curated to avoid data leakage issues, presenting a new framework for synthetically generating diverse software tasks from limited real-world data. We evaluate OmniCode with popular agent frameworks such as SWE-Agent and show that while they may perform well on bug fixing for Python, they fall short on tasks such as Test Generation and in languages such as C++ and Java. For instance, SWE-Agent achieves a maximum of 25.0% with DeepSeek-V3.1 on C++ Test Generation. OmniCode aims to serve as a robust benchmark and spur the development of agents that can perform well across different aspects of software development. Code and data are available at https://github.com/seal-research/OmniCode. --- ### Unlocking the Duality between Flow and Field Matching Authors: Daniil Shlenskii, Alexander Varlamov, Nazar Buzun, Alexander Korotin Published: 2026-02-02 Categories: cs.LG arXiv: https://arxiv.org/abs/2602.02261v1 Conditional Flow Matching (CFM) unifies conventional generative paradigms such as diffusion models and flow matching. Interaction Field Matching (IFM) is a newer framework that generalizes Electrostatic Field Matching (EFM) rooted in Poisson Flow Generative Models (PFGM). While both frameworks define generative dynamics, they start from different objects: CFM specifies a conditional probability path in data space, whereas IFM specifies a physics-inspired interaction field in an augmented data space. This raises a basic question: are CFM and IFM genuinely different, or are they two descriptions of the same underlying dynamics? We show that they coincide for a natural subclass of IFM that we call forward-only IFM. Specifically, we construct a bijection between CFM and forward-only IFM. We further show that general IFM is strictly more expressive: it includes EFM and other interaction fields that cannot be realized within the standard CFM formulation. Finally, we highlight how this duality can benefit both frameworks: it provides a probabilistic interpretation of forward-only IFM and yields novel, IFM-driven techniques for CFM. --- ### Learning Markov Decision Processes under Fully Bandit Feedback Authors: Zhengjia Zhuo, Anupam Gupta, Viswanath Nagarajan Published: 2026-02-02 Categories: cs.LG arXiv: https://arxiv.org/abs/2602.02260v1 A standard assumption in Reinforcement Learning is that the agent observes every visited state-action pair in the associated Markov Decision Process (MDP), along with the per-step rewards. Strong theoretical results are known in this setting, achieving nearly-tight $Θ(\sqrt{T})$-regret bounds. However, such detailed feedback can be unrealistic, and recent research has investigated more restricted settings such as trajectory feedback, where the agent observes all the visited state-action pairs, but only a single \emph{aggregate} reward. In this paper, we consider a far more restrictive ``fully bandit'' feedback model for episodic MDPs, where the agent does not even observe the visited state-action pairs -- it only learns the aggregate reward. We provide the first efficient bandit learning algorithm for episodic MDPs with $\widetilde{O}(\sqrt{T})$ regret. Our regret has an exponential dependence on the horizon length $\H$, which we show is necessary. We also obtain improved nearly-tight regret bounds for ``ordered'' MDPs; these can be used to model classical stochastic optimization problems such as $k$-item prophet inequality and sequential posted pricing. Finally, we evaluate the empirical performance of our algorithm for the setting of $k$-item prophet inequalities; despite the highly restricted feedback, our algorithm's performance is comparable to that of a state-of-art learning algorithm (UCB-VI) with detailed state-action feedback. --- ### Segment to Focus: Guiding Latent Action Models in the Presence of Distractors Authors: Marcus Fechner, Hamza Adnan, Constantin C. Lüth, Matthew T. Jackson, Alexey Zakharov, J. Marius Zöllner Published: 2026-02-02 Categories: cs.LG, cs.CV arXiv: https://arxiv.org/abs/2602.02259v2 Latent action models (LAMs) offer a promising path to pre-training embodied agents on large amounts of action-free video. They infer latent actions between consecutive observations that can later be decoded to ground-truth actions using a small number of labels. However, recent work has shown that this recipe fails in the presence of action-correlated visual distractors common in real-world video, such as dynamic backgrounds, camera shake, or other moving objects. In these scenarios, the standard reconstruction objective drives latent actions to encode exogenous motion instead of agent-controlled dynamics, resulting in policies that underperform when fine-tuned. We observe, however, that endogenous and exogenous factors are typically spatially separated in pixel space: control-relevant change is concentrated on the agent, while distractor motion occurs elsewhere. We exploit this observation by restricting the reconstruction objective to agent pixels, forcing latent actions to explain agent-controlled dynamics rather than exogenous ones. We call this method MaskLAM; it obtains the agent mask zero-shot from off-the-shelf segmentation foundation models (e.g., SAM) and requires no architectural changes, auxiliary losses, or action labels during pre-training. Across two continuous-control benchmarks (Distracting Control Suite, Distracting Meta-World), MaskLAM reduces normalized linear-probe MSE by up to $3.51\times$ and improves normalized return by up to $4.97\times$ over LAPO, while narrowing the gap to LAOM-Labels, which relies on ground-truth action supervision. --- ### Alignment-Aware Model Adaptation via Feedback-Guided Optimization Authors: Gaurav Bhatt, Aditya Chinchure, Jiawei Zhou, Leonid Sigal Published: 2026-02-02 Categories: cs.LG arXiv: https://arxiv.org/abs/2602.02258v2 Fine-tuning is the primary mechanism for adapting foundation models to downstream tasks; however, standard approaches largely optimize task objectives in isolation and do not account for secondary yet critical alignment objectives (e.g., safety and hallucination avoidance). As a result, downstream fine-tuning can degrade alignment and fail to correct pre-existing misaligned behavior. We propose an alignment-aware fine-tuning framework that integrates feedback from an external alignment signal through policy-gradient-based regularization. Our method introduces an adaptive gating mechanism that dynamically balances supervised and alignment-driven gradients on a per-sample basis, prioritizing uncertain or misaligned cases while allowing well-aligned examples to follow standard supervised updates. The framework further learns abstention behavior for fully misaligned inputs, incorporating conservative responses directly into the fine-tuned model. Experiments on general and domain-specific instruction-tuning benchmarks demonstrate consistent reductions in harmful and hallucinated outputs without sacrificing downstream task performance. Additional analyses show robustness to adversarial fine-tuning, prompt-based attacks, and unsafe initializations, establishing adaptively gated alignment optimization as an effective approach for alignment-preserving and alignment-recovering model adaptation. --- ### Well-Posed KL-Regularized Control via Wasserstein and Kalman-Wasserstein KL Divergences Authors: Viktor Stein, Adwait Datar, Nihat Ay Published: 2026-02-02 Categories: math.OC, cs.LG arXiv: https://arxiv.org/abs/2602.02250v2 Kullback-Leibler (KL) divergence regularization is widely used in reinforcement learning, but it becomes infinite under support mismatch and can degenerate in low-noise regimes. Using a unified information-geometric framework, we introduce KL analogs by replacing the Fisher-Rao geometry in the dynamical formulation of the KL with transport-based geometries, and derive closed-form expressions for common distribution families. Between elliptic distributions, these divergences remain finite for degenerating equal covariances and yield a geometric interpretation of regularization heuristics used in Kalman ensemble methods. We demonstrate the utility of these divergences in KL-regularized optimal control. In the fully tractable setting of linear time-invariant systems with Gaussian process noise, the classical KL reduces to a quadratic control penalty that becomes singular as process noise vanishes. Our variants remove this singularity and yield well-posed problems. In both the double integrator and cart-pole examples, the resulting controls preserve nontrivial feedback and achieve better closed-loop performance. --- ### Entropy-Preserving Supervised Fine-Tuning via Adaptive Self-Distillation for Large Reasoning Models Authors: Hao Wang, Hao Gu, Hongming Piao, Kaixiong Gong, Yuxiao Ye, Xiangyu Yue, Sirui Han, Yike Guo, Dapeng Wu Published: 2026-02-02 Categories: cs.LG, cs.CL arXiv: https://arxiv.org/abs/2602.02244v3 The standard post-training recipe for large reasoning models, supervised fine-tuning followed by reinforcement learning (SFT-then-RL), may limit the benefits of the RL stage: while SFT imitates expert demonstrations, it often causes overconfidence and reduces generation diversity, leaving RL with a narrowed solution space to explore. Adding entropy regularization during SFT is not a cure-all; it tends to flatten token distributions toward uniformity, increasing entropy without improving meaningful exploration capability. In this paper, we propose CurioSFT, an entropy-preserving SFT method designed to enhance exploration capabilities through intrinsic curiosity. It consists of (a) Self-Exploratory Distillation, which distills the model toward a self-generated, temperature-scaled teacher to encourage exploration within its capability; and (b) Entropy-Guided Temperature Selection, which adaptively adjusts distillation strength to mitigate knowledge forgetting by amplifying exploration at reasoning tokens while stabilizing factual tokens. Extensive experiments on mathematical reasoning tasks demonstrate that, in SFT stage, CurioSFT outperforms the vanilla SFT by 2.5 points on in-distribution tasks and 2.9 points on out-of-distribution tasks. We also verify that exploration capabilities preserved during SFT successfully translate into concrete gains in RL stage, yielding an average improvement of 5.0 points. --- ### Variational Entropic Optimal Transport Authors: Roman Dyachenko, Nikita Gushchin, Kirill Sokolov, Petr Mokrov, Evgeny Burnaev, Alexander Korotin Published: 2026-02-02 Categories: cs.LG arXiv: https://arxiv.org/abs/2602.02241v2 Entropic optimal transport (EOT) in continuous spaces with quadratic cost is a classical tool for solving the domain translation problem. In practice, recent approaches optimize a weak dual EOT objective depending on a single potential, but doing so is computationally not efficient due to the intractable log-partition term. Existing methods typically resolve this obstacle in one of two ways: by significantly restricting the transport family to obtain closed-form normalization (via Gaussian-mixture parameterizations), or by using general neural parameterizations that require simulation-based training procedures. We propose Variational Entropic Optimal Transport (VarEOT), based on an exact variational reformulation of the log-partition $\log \mathbb{E}(\exp(\cdot))$ as a tractable minimization over an auxiliary log-normalizer. This yields a differentiable learning objective optimized with stochastic gradients and avoids the necessity of MCMC simulations during the training. We provide theoretical guarantees, including finite-sample generalization bounds and approximation results under universal function approximation. Experiments on synthetic data and unpaired image-to-image translation demonstrate competitive or improved translation quality, while comparisons within the solvers that use the same weak dual EOT objective support the benefit of the proposed optimization principle. The code for our solver can be found at https://github.com/DrEternity/VarEOT . --- ### Causal Inference for Preprocessed Outcomes with an Application to Functional Connectivity Authors: Zihang Wang, Razieh Nabi, Benjamin B. Risk Published: 2026-02-02 Categories: stat.ME, stat.ML arXiv: https://arxiv.org/abs/2602.02240v1 In biomedical research, repeated measurements within each subject are often processed to remove artifacts and unwanted sources of variation. The resulting data are used to construct derived outcomes that act as proxies for scientific outcomes that are not directly observable. Although intra-subject processing is widely used, its impact on inter-subject statistical inference has not been systematically studied, and a principled framework for causal analysis in this setting is lacking. In this article, we propose a semiparametric framework for causal inference with derived outcomes obtained after intra-subject processing. This framework applies to settings with a modular structure, where intra-subject analyses are conducted independently across subjects and are followed by inter-subject analyses based on parameters from the intra-subject stage. We develop multiply robust estimators of causal parameters under rate conditions on both intra-subject and inter-subject models, which allows the use of flexible machine learning. We specialize the framework to a mediation setting and focus on the natural direct effect. For high dimensional inference, we employ a step-down procedure that controls the exceedance rate of the false discovery proportion. Simulation studies demonstrate the superior performance of the proposed approach. We apply our method to estimate the impact of stimulant medication on brain connectivity in children with autism spectrum disorder. --- ### Interpretability in Deep Time Series Models Demands Semantic Alignment Authors: Giovanni De Felice, Riccardo D'Elia, Alberto Termine, Pietro Barbiero, Giuseppe Marra, Silvia Santini Published: 2026-02-02 Categories: cs.LG arXiv: https://arxiv.org/abs/2602.02239v3 Deep time series models continue to improve predictive performance, yet their deployment remains limited by their black-box nature. In response, existing interpretability approaches in the field keep focusing on explaining the internal model computations, without addressing whether they align or not with how a human would reason about the studied phenomenon. Instead, we state interpretability in deep time series models should pursue semantic alignment: predictions should be expressed in terms of variables that are meaningful to the end user, mediated by spatial and temporal mechanisms that admit user-dependent constraints. In this paper, we formalize this requirement and state that, once established, semantic alignment must be preserved under temporal evolution: a constraint with no analog in static settings. Provided with this definition, we outline a blueprint for semantically aligned deep time series models, identify properties that support trust, and discuss implications for model design. ---