cs.LG · 2026-10-27 · No. 155

Machine Learning, 2026-10-27.

68 new papers in cs.LG. Titles, authors, abstracts. Links to arXiv. Want this in your inbox every morning? Subscribe →

01 — The papers

68 entries
  1. 01

    Auto-Research Arena: Benchmarking and Advancing Auto-Research in Language Modeling

    Xi Chen, Chongjian Ge, Yiran Xu, Yiqun Mei, Henrik I. Christensen, Hao Tan, Ziwen Chen, Hanwen Jiang

    cs.LG

    As autonomous research systems become more capable, benchmarks must both evaluate their performance and support meaningful research progress. We introduce Auto-Research Arena, a multi-target benchmark for evaluating autonomous research systems and discovering ideas in language modeling. Arena supports diverse research objectives within a shared experimental environment by separating admissibility from improvement: model quality and resource...

    arxiv.org/abs/2610.35866 · PDF

  2. 02

    Work Agents: Systems, Capabilities, and Optimization

    Zishan Xu, Enjun Du, Mingqi Gao, Cheng Yang, Junhao Shen, Zirong Chen, Zhiyuan Yao, Yifu Guo, Yuquan Lu, Jinyang...

    cs.LG

    As foundation models evolve from information generation to digital work execution, agents are increasingly capable of completing complex tasks across tools, applications, and workflows. However, existing research lacks a unified perspective on how agents reliably accomplish delegated work. To address this gap, we present a work-centric, system-level perspective that views work agents as systems combining foundation models and runtime...

    arxiv.org/abs/2610.35845 · PDF

  3. 03

    FOCUS: Active Critic Learning for Long-Horizon Language Model Agents

    Zhengping Jiang, William Jurayj, Liaoyaqi Wang, Chunsheng Zuo, Anqi Liu, Benjamin Van Durme

    cs.LG

    Sparse outcome rewards make it difficult for long-horizon language-model agents to assign credit to earlier actions. Learned critics provide value estimates for intermediate states, but these estimates can be inaccurate, while estimating intermediate values with repetitive continuation samples from points within a trajectory incurs substantial costs. To mitigate these issues, we introduce FOCUS, an RL method that uses selective branching to...

    arxiv.org/abs/2610.35801 · PDF

  4. 04

    Selective Tie Learning: Mitigating Spurious Correlations in Preference Optimization

    Christian Moya, Purav Matlia, Elliott Thornley, Guang Lin

    cs.LG

    Preference optimization can learn spurious correlations, and augmenting the preference data with ties (equally preferred responses) can reduce spurious reliance. However, ties differ in their corrective effect, making their selection for augmentation important. For log-linear policies, we derive each tie's influence on spurious reliance and show that taking the largest predicted reductions is first-order optimal among subsets of equal size....

    arxiv.org/abs/2610.35787 · PDF

  5. 05

    A theory of noisy linear recurrent neural network learning dynamics

    John J. Vastola, Kanaka Rajan

    cs.LG · q-bio.NC

    Biological neural networks are noisy: their dynamics involve difficult-to-predict fluctuations which are likely due not just to chaos, but also to a variety of low-level biophysical processes. Despite the omnipresence of intrinsic noise, its theoretical consequences are incompletely understood, especially in the context of learning. For example, does noise meaningfully shape representation learning, or is it better viewed as a nuisance factor...

    arxiv.org/abs/2610.35785 · PDF

  6. 06

    Improving Signals, Declining Encoder: A Known-Answer Test of Label-Free Monitoring in Self-Supervised Learning

    Nilotpal Kakati

    cs.LG · cs.CV · physics.data-an

    Self-supervised training is monitored without labels, through the loss and statistics of the learned features. We report a case where we know the encoders lost part of what they had learned, so we can check whether these signals noticed. We trained LeJEPA encoders on synthetic images with known generating factors. In nine of ten seeds, the readability of those factors fell abruptly and stayed down, nearly always along the same direction in...

    arxiv.org/abs/2610.35766 · PDF

  7. 07

    LUCID: Local Understanding of Covariate Importance with Conditional Density Estimation

    Yixiao Wang, Srikar Katta, Jon Donnelly, Emanuele Borgonovo, Cynthia Rudin

    cs.LG · stat.ML

    Computing local-level variable importance (VI) is critical for understanding machine learning models. Local permutation-based VI metrics offer a simple yet powerful approach, measuring how much a model's output changes when perturbing its inputs. However, existing perturbation strategies often ignore feature dependence or account for it only through restricted approximations. Perturbing a feature without fully respecting this dependence can...

    arxiv.org/abs/2610.35740 · PDF

  8. 08

    Multi-Objective Reinforcement Learning Using Routed Ensembles and Trajectory Attribution

    Zhenshuo Zhang, Ruoxuan Xiong, Hongyang R. Zhang

    cs.LG · cs.AI · cs.MA · cs.RO · stat.ML

    Multi-objective reinforcement learning is a fundamental problem for modern AI, with applications spanning post-training, reasoning models, and robotic control. A central challenge is to learn policies that balance competing objectives, such as helpfulness versus honesty in safety and alignment, or speed versus energy in robotics. The optimal policy can vary sharply with objective-weighting combinations, and the extent of this variation...

    arxiv.org/abs/2610.35733 · PDF

  9. 09

    TRACE: Memory-Efficient Time Series Class-Incremental Learning via Temporal-Spectral Reconstruction

    Junhui Mao, Junru Zhang, Liming Ye, Cheng Peng, Haoran Shi, Xiaoli Tang, Baokun Wang, Bin Dou, Yabo Dong, Han Yu

    cs.LG

    Web-connected sensing applications continuously generate multivariate time series with evolving behavioral categories, making Time Series Class-Incremental Learning (TSCIL) essential for adapting to new classes without forgetting previously acquired knowledge. However, preserving diverse and evolving class distributions under strict privacy and memory constraints remains a fundamental challenge. Existing approaches often face a trade-off...

    arxiv.org/abs/2610.35727 · PDF

  10. 10

    Can JEV Support Clinical Trial Criteria Assessment?

    Mengxian He, Ziyue Gao, Chun Li, Shunyi Zhang, Hao Wei, Minqing Zhang, Wu Yuan

    cs.LG · cs.AI · cs.CL

    Clinical trial prescreening requires distinguishing unmet conditions from insufficient evidence in patient records. Given patient records, criterion questions, and predefined answer options, we retrospectively compared JEV, a non-generative decision model returning native option probabilities, with schema-constrained Qwen Flash across 2,048 test patient-condition pairs from TrialGPT and question-answer datasets derived from MIMIC-III and...

    arxiv.org/abs/2610.35721 · PDF

  11. 11

    Quantifying the failure of symmetry learning in linear diffusion models due to initialization and finite-sample effects

    John J. Vastola, Kanaka Rajan

    cs.LG

    Data distributions often possess one or more symmetries: for example, all orientations of any specific dog are included in the distribution of all possible dog pictures. Despite the omnipresence of such symmetries, modern generative modeling approaches may fail to incorporate or learn them, which can lead to lower sample efficiency and poorer generalization. To what extent do diffusion models fail to learn data distribution symmetries, and...

    arxiv.org/abs/2610.35720 · PDF

  12. 12

    Audit Where Hacks Pay: Budgeted Strong-Verifier Audits Against Reward Hacking in GRPO

    Shurun Li, Michael Hua Wang, Haibo Zeng, Long Wang

    cs.LG · cs.AI

    Reinforcement learning with verifiable rewards trains on cheap checks, such as unit tests, that a policy can learn to game, while a stronger check is too expensive to run on every rollout. We study how to spend a limited strong-verifier budget in GRPO, separating which weakly passing rollouts to audit from what reward a caught hack receives. At a fixed policy state with centered advantages, the expected update depends on a design only through...

    arxiv.org/abs/2610.35716 · PDF

  13. 13

    Quantitative Generalization Bounds for Multi-Layer Transformers: A Few Long Strings can Beat Many Short Ones

    Marius Henry, Michael Rizvi-Martel, Michael Hahn

    cs.LG · cs.AI · cs.CL

    Transformer-based language models exhibit an astonishing capacity to generalize to longer and more difficult problem instances than those seen in training. However, theoretical understanding of this phenomenon still lags behind. Unlike most machine learning models, the generalisation abilities of transformers inherently require considering two axes. On one hand, one must examine in-distribution statistical generalisation, which concerns the...

    arxiv.org/abs/2610.35715 · PDF

  14. 14

    Detecting bus delays and service anomalies in Niterói with autoencoder latent feature representations

    Matheus Germano, Claudio E. C. Campelo

    cs.LG

    Bus riders judge service by two failures: single trips that run late, and days on which the whole system misbehaves. We ask whether one unsupervised representation of raw GPS telemetry can serve both diagnoses on the NetMob 2026 Niterói bus dataset. An autoencoder compresses each trip into a 64-dimensional latent vector; the frozen encoder feeds supervised delay classifiers, and the reconstruction error serves as an anomaly score. Two...

    arxiv.org/abs/2610.35710 · PDF

  15. 15

    Triage: Skip, Skim, or Refine Attention Tiles in Video Diffusion Transformers

    Yeonjae Kim, Omin Kwon, Seong Hoon Seo, Jae W. Lee

    cs.LG · cs.CV

    Attention over tens of thousands of visual tokens dominates the inference of video diffusion transformers (DiTs). Existing methods reduce this cost through sparsity, low precision, or both, but use a single precision for all kept tiles. We observe that the output error introduced by approximation varies widely across tiles. Many tiles can be skipped with little error, others must be computed but gain little beyond low precision, and a few...

    arxiv.org/abs/2610.35706 · PDF

  16. 16

    MiRL: Error-Targeted Reinforcement Learning for Correcting Terminal Agent Misbehaviors

    Runchu Tian, Zaid Khan, Zun Wang, Justin Chih-Yao Chen, Mohit Bansal

    cs.LG · cs.AI · cs.CL

    Error-targeted learning gives a legible way to improve a model: identify specific weaknesses of the model, train the model to fix these weaknesses, and check whether the weaknesses are fixed. We study how to apply error-targeted learning to terminal agents, which solve long-horizon tasks specified in natural language by interacting with a command shell. Existing error-targeted methods are designed for single-turn tasks, and existing analyses...

    arxiv.org/abs/2610.35697 · PDF

  17. 17

    GNN-VIEW: Visual Inspection and Explanation of Graph-based Earth-System Models

    Alessandro Lucatelli, Ferdinand Unterhuber, Katharina Krenzer, Mennatallah El-Assady, Benedikt Soja, Christina...

    cs.LG · cs.HC · physics.ao-ph

    Machine learning Earth-system models are becoming increasingly important, and so is their development. Graph-based models, one of the most prominent approaches, allow developers to constrain the learning process before training via graph connectivity. Yet, developing these models remains challenging: the graph is built algorithmically with little intuition about the final result beyond static plotting, which requires extra work and is often...

    arxiv.org/abs/2610.35693 · PDF

  18. 18

    CounterFlow: Target-Conditioned Flow Matching for Visual Counterfactuals

    Paul Sevellec, Elisa Fromont, Romaric Gaudel, Laurence Roze, Paul Viallard, Laetitia Houeix

    cs.LG · cs.AI · cs.CV

    In eXplainable AI, counterfactual explanations have gained significant traction by offering users direct, intuitive, and actionable insights into model decisions. We introduce CounterFlow, a novel framework based on Conditional Flow Matching that generates visual counterfactual explanations by directly steering a source image toward a user-specified target class. By leveraging target-conditioned vector fields, CounterFlow identifies highly...

    arxiv.org/abs/2610.35679 · PDF

  19. 19

    SCAFFOLD under Hessian Similarity: Sharp Convergence Rates

    Zhirayr Tovmasyan, Slavomír Hanzely, Egor Shulgin, Peter Richtárik, Sebastian U. Stich

    cs.LG · math.OC

    One of the main challenges of federated learning is heterogeneous data split among the clients, leading to biased local updates called client drift. The SCAFFOLD algorithm (Karimireddy et al., 2020) corrects client drift with control variates, but its relation to the similarity of client data is not yet fully understood: the original similarity-based guarantees cover only quadratics with a stronger second-order similarity assumption, and a...

    arxiv.org/abs/2610.35663 · PDF

  20. 20

    A Hybrid Framework for Explainable Aviation Risk: Causal Chain Extraction from Safety Reports with Physics-Based Failure Quantification and Interventional Analysis

    Yutian Pang, Yifang Tian, Jiacheng Xie, Xinyu Zhao, Jueming Hu, John-Paul Clarke, Hans-Arno Jacobsen

    cs.LG

    Aviation authorities hold decades of accident reports whose narratives record how investigators reconstructed each accident. Computational work on these archives has mostly sorted reports into fixed taxonomies. In this work, we propose a hybrid framework in which a text-based corpus identifies the leading factors of a failure mode. Physics-based models then quantify how likely that failure is at a stated operating point. Ordered causal chains...

    arxiv.org/abs/2610.35658 · PDF

  21. 21

    Active Learning of Conditional Generative Models via the Transport Neural Tangent Kernel

    Jayoung Ryu, Kyunghyun Cho, Romain Lopez

    cs.LG

    Many high-throughput scientific experiments produce distributional outputs, requiring conditional generative models to interpret the data, and to predict outcomes at new inputs. Acquiring experimental data at new conditions (e.g., for a new antigen or a new cell type) is often expensive, motivating the need for automated experimental design methods. However, existing active learning frameworks are limited as they assume scalar or...

    arxiv.org/abs/2610.35652 · PDF

  22. 22

    Proofs of strong forms of the Bovdi--Geche universality conjecture and a complex analogue

    Victor Bovdi, Ambrus Pal

    cs.LG · math.NT · math.RA

    We prove two universality theorems for neural networks consisting of a single neuron. First we prove a strong form of the Bovdi--Geche universality conjecture on neurons over finite fields, where the activation function is a certain modification of the discrete logarithm, using the Weil bound on multiplicative character sums over finite fields. Then we show a generalisation of Aizenberg's universality theorem for complex binary neurons with a...

    arxiv.org/abs/2610.35645 · PDF

  23. 23

    Adapt for Planning, Not just Prediction: Optimistic Test-Time World Model Adaptation

    Akshay Mete, Dileep Kalathil, P. R. Kumar

    cs.LG · cs.AI · cs.RO

    Latent world models are typically pre-trained offline and kept frozen at test time, so planning with them can degrade under distribution shift. Test-time adaptation (TTA) can mitigate this degradation by updating the model during deployment. Unlike pre-training on fixed data, TTA for planning is closed-loop: the planner uses model rollouts to select actions, and the resulting transitions drive the next update. This raises the question: which...

    arxiv.org/abs/2610.35636 · PDF

  24. 24

    Distilling neural networks into equations to predict planetary instability

    Simon Alford, Edward Gu, Omar Costilla-Reyes, Dan Tamayo, Miles Cranmer, Kevin Ellis

    cs.LG · astro-ph.EP

    The problem of predicting when a multi-planet system will become dynamically unstable has defied analytical progress over the past three hundred years. Recent machine learning advances achieve high predictive performance of instability dynamics, but as black-box systems cannot be interpreted to connect to existing scientific understanding. Here we show analytic equations that predict when a multi-planet system will undergo instability. We...

    arxiv.org/abs/2610.35617 · PDF

  25. 25

    ExCal: Optimization-Free Multivariate Time Series Anomaly Detection via Excess Calibration

    Kadir-Kaan Özer, René Ebeling, Markus Enzweiler

    cs.LG · cs.AI

    We introduce ExCal, a normal-only multivariate time series anomaly detector that combines raw state retrieval, a signal/derivative subspace, and robust statistics across temporal scales. Each expert map receives a high normality floor from the training prefix. During scoring, positive excess is retained, attenuated for broadly firing maps, and combined with a small rank term. The method fits reference memories and projections without gradient...

    arxiv.org/abs/2610.35568 · PDF

  26. 26

    High-resolution tactile observation makes hidden racket-control mechanisms learnable

    Haoyang Tian, Zhengwei Li, Houcheng Li, Xu Yang, Peigeng Gu, Xinyue Wang, Tairan Zhu, Shuo Jiang, Long Cheng

    cs.LG · cs.HC · cs.RO · eess.SP

    Experts can master fine hand-object manipulation skills, yet limited conscious perception of rapid tactile interactions hinders their ability to articulate the underlying control mechanisms. Badminton racket control requires rapid, fine hand-force coordination, providing a demanding testbed for uncovering these mechanisms. An 864-node tactile-inertial glove with synchronized 100-Hz acquisition recorded over 14,000 strokes from more than 70...

    arxiv.org/abs/2610.35563 · PDF

  27. 27

    Multi-Fidelity Training Reshapes Uncertainty Decomposition in Atomistic Models

    Matteo Tolloso, Davide Bacciu, Luigi Bonati

    cs.LG · cs.AI

    Multi-fidelity training, which combines abundant low-fidelity data with scarce high-fidelity labels, is increasingly common in atomistic modeling, but its effect on predictive uncertainty remains largely unexplored. We investigate this question using deep ensembles of mean-variance MACE models trained on ANI-1x, contrasting high-fidelity training from scratch with low-fidelity pretraining followed by fine-tuning. We characterize aleatoric,...

    arxiv.org/abs/2610.35560 · PDF

  28. 28

    Balanced Sampling in Stochastic Multi-objective Bandits

    Mengfan Xu

    cs.LG · stat.ML

    In stochastic multi-objective bandits, Pareto optimal arms may represent different trade-offs among objectives. Balanced sampling across them is important in practical settings such as online advertising. Yet, there seems to be little consensus on how to measure fairness, to our best knowledge. Achieving fairness is also tough as the set of Pareto optimal arms is unknown, and early sampling decisions may create imbalances that are difficult...

    arxiv.org/abs/2610.35551 · PDF

  29. 29

    Efficient probabilistic weather forecasting at quarter-degree resolution with two-sided transport attention

    Junu Choi, Youngjoon Hong

    cs.LG

    High-resolution probabilistic weather forecasting remains computationally demanding. Here we present Astra, a 0.25$^\circ$ ensemble model trained on eight H200 GPUs. It combines Kantorovich Attention, a two-sided transport attention in which each region chooses where it reads from and writes to, so every region sends and receives at token resolution, with a self-featurised continuous ranked probability score (CRPS) that also scores forecasts...

    arxiv.org/abs/2610.35545 · PDF

  30. 30

    Beyond Modeling Order: Dual-Path Spatiotemporal Representation Learning for Cross-Subject EEG Emotion Recognition

    Zehui Xiao, Yi Zheng, Guixing Fu, Wenkai Lu, Lingfei He, Tao Tan, Yue Liu, Silun Wang, Junhai Xu, Mengting Liu

    cs.LG

    Cross-subject generalization remains a major challenge for EEG-based emotion recognition due to pronounced inter-subject variability in neural activity and emotion-related spatiotemporal patterns. Existing methods mainly alleviate inter subject differences through distribution alignment or contrastive learning, while paying limited attention to representation bias associated with fixed spatiotemporal modeling orders. We hypothesize that...

    arxiv.org/abs/2610.35540 · PDF

  31. 31

    S3AC: Stabilized, Selective, and Scalable Actor-Critic for LLM Reasoning

    Shuo Shen, Guangran Cheng, Chengqi Lyu, Xiangsen Wang, Yanan Sun, Wenwei Zhang

    cs.LG · cs.CL

    Critic-free methods are widely used in reinforcement learning with verifiable rewards (RLVR) for long chain-of-thought (CoT) reasoning, avoiding difficult and costly critic training. Common alternatives lack token-level credit assignment and typically rely on grouped rollouts to estimate prompt-specific baselines. A critic can address both limitations, but long-CoT critic training is hindered by insufficient coverage of trajectories with...

    arxiv.org/abs/2610.35518 · PDF

  32. 32

    Model reduction in time: Learning flow maps for reducing stochastic dynamical systems in time

    Shreya Jha, Benjamin Peherstorfer

    cs.LG · math.NA

    Predictions with dynamical systems can be computationally expensive---even with reduced models---because reaching future prediction times can require many sequential time integration steps. We propose Flow-Map First-Order Trajectory Matching (FM-FTM), a nonintrusive model reduction approach that accelerates ensemble predictions of stochastic dynamical systems by learning flow maps that replace multiple time-integration steps. Our method uses...

    arxiv.org/abs/2610.35504 · PDF

  33. 33

    When Fine-Grained Credit Assignment Helps and Hurts: Multimodal Agentic Reinforcement Learning through Video Editing

    Lei Zhao, Zongheng Cao, Xinyu Hu

    cs.LG

    Reinforcement learning (RL) has shown promise for training multimodal agents to reason and use tools. However, it remains unclear when fine-grained credit assignment improves learning and when it fails to improve task performance. Video editing provides a useful testbed for this question, combining multimodal perception, long-horizon planning, and tool use with delayed outcome feedback. We construct four movie-trailer editing task...

    arxiv.org/abs/2610.35497 · PDF

  34. 34

    Evaluating Bayesian deep learning for robust biomedical time-series classification under distribution shift: A sleep scoring case study

    Mads Birch Sørensen, Ulrike Richter, Francois Gastambide, Michael Riis Andersen

    cs.LG

    In biomedical time-series classification, machine learning models often suffer from deteriorating accuracy and calibration due to distribution shift. We investigate to what degree Bayesian-inspired deep learning improves robustness through Bayesian model averaging (BMA). Using rodent sleep scoring as a case study, we evaluate BMA under distribution shifts in genotype, laboratory, and species. We systematically compare nine approximate...

    arxiv.org/abs/2610.35484 · PDF

  35. 35

    GPDPO: Learning from Reward Conflicts via Group Pareto-Dominance Policy Optimization

    Zixuan Xu, Haoxuan Chen, Jiangxin Sun, Jian-Fang Hu

    cs.LG · cs.CL

    Multi-reward reinforcement learning offers a promising framework for aligning large language models (LLMs) with diverse objectives, but conflicting reward signals can induce ambiguous update directions and hinder stable optimization. Existing approaches address such conflicts through reward reweighting, gradient coordination, or rollout filtering, yet they primarily operate on per-response or group-level optimization signals and underexploit...

    arxiv.org/abs/2610.35480 · PDF

  36. 36

    Amplify, Locate, Estimate (ALE) on Rare-Failure Rates of Fine-Tuned LLMs

    Patrick Reichherzer, Jonathan Evertz, Thorsten Holz

    cs.LG

    Safety fine-tuning can make undesirable model behaviors sufficiently rare that measuring their frequency by direct sampling becomes prohibitively expensive. We introduce Amplify, Locate, Estimate (ALE), a method for estimating such rare failure probabilities efficiently. ALE uses the change in token log-probabilities induced by the safety fine-tuning to define an amplification direction, along which we increase the failure rate for cheap...

    arxiv.org/abs/2610.35472 · PDF

  37. 37

    Multi-Horizon Future Discovery for Physics-Informed Exploration

    Shuwei Wang, Qiang Guan, Brian Qiu, Haoxiang Huang, Haoxi Wang, Chenyu Li, Bo Xu, Sen Cui

    cs.LG

    Exploration in physical control environments requires agents to identify actions whose consequences reveal previously unexplored aspects of the underlying dynamics. However, many intrinsic motivation methods rely on immediate prediction errors, state novelty, or local physical changes, which may overlook informative outcomes that emerge across extended temporal horizons. We investigate whether learning structured representations of reachable...

    arxiv.org/abs/2610.35468 · PDF

  38. 38

    Prior Predictive Matching for Bayesian LoRA

    Tomasz Kuśmierczyk, Eliezer de Souza da Silva, Marcin Sendera

    cs.LG · stat.ML

    Bayesian variants of Low-Rank Adaptation (LoRA) provide principled uncertainty estimation when fine-tuning large foundation models, but require specifying the prior over the low-rank weight updates. We address this hyperparameter selection problem by introducing a closed-form prior-predictive estimator that operates before fine-tuning. We analyze a second-order expansion of the network and quantify how the low-rank prior drives function-space...

    arxiv.org/abs/2610.35457 · PDF

  39. 39

    Intent-Preserving Adversarial User Simulation for Multi-Turn Agent Evaluation

    Atri Vivek Sharma, Brian Formento, Alessio Lomuscio

    cs.LG

    Multi-turn agents are commonly evaluated with LLM-based simulated users, which can be overly cooperative and can therefore overestimate an agent's performance. We here introduce StressSim, an intent-preserving adversarial user simulator framework for stress-testing agents with challenging yet valid user interactions. StressSim optimises the simulator's high-level adversarial plan for each task instance. Candidate planner prompts are evolved...

    arxiv.org/abs/2610.35426 · PDF

  40. 40

    TransECG-SSL: Hybrid Self-Supervised Representation Learning for ECG Arrhythmia Classification

    Ahtisham Ayyub, Fasih Haider, Muhammad Rehan Usman, Saturnino Luz, Christos Politis, Muhammad Arslan Usman

    cs.LG

    Cross-dataset generalisation remains a major challenge for automated electrocardiogram (ECG) arrhythmia classification because of differences in patient populations, acquisition protocols, annotation practices, and class distributions. Learning representations that remain informative across independent datasets could reduce reliance on dataset-specific labelled data. We propose TransECG-SSL, a self-supervised framework for transferable...

    arxiv.org/abs/2610.35378 · PDF

  41. 41

    Benchmarking uncertainty quantification for engineering regression under graded distribution shift

    Ahmad Lawal, Rahim Pasbanigoloojeh, Asmau Lawal Bilbis

    cs.LG

    Machine learning models in engineering are often used beyond their training conditions, where reliable uncertainty quantification (UQ) matters most. Existing comparisons of regression UQ methods rely on random splits, which assess calibration only in distribution. We introduce a three-rung shift ladder (in distribution, soft shift and hard shift) built along physically meaningful shift variables, and evaluate 12 methods from five families on...

    arxiv.org/abs/2610.35356 · PDF

  42. 42

    OneEEG: A Unified Signal-Visual Multimodal Language Model with Task-Aware Mixture of Projectors for Multi-Task EEG Understanding

    Jiacheng Zhang, Xueyi Zhang, Xin Lu, Siqi Cai, Haizhou Li

    cs.LG

    Electroencephalography (EEG) is widely used in clinical assessment, brain-computer interfaces, and cognitive-state analysis, yet EEG modeling remains fragmented across task paradigms. Recent multimodal large language models (MLLMs) provide a promising interface for flexible multimodal reasoning, but their ability to understand EEG inputs remains limited. In this work, we propose OneEEG, a unified EEG-MLLM framework for multi-task EEG...

    arxiv.org/abs/2610.35355 · PDF

  43. 43

    Should Pseudo-Labels Share Mini-Batches with Labeled Examples? Semi-Supervised Continual Active Learning with Limited Training Cost

    Shiqi Zhang, Tuomas Virtanen

    cs.LG · cs.AI · cs.SD · eess.AS

    Active learning (AL) reduces annotation by querying informative examples. Continual active learning (CAL) also cuts the training cost, the total number of gradient steps, by training one model across AL rounds with a fixed number of mini-batch steps per round. Pseudo-labels add supervision without annotation, but they increase training cost. In CAL, where every round trains on a fixed number of mini-batches, whether and how pseudo-labeled...

    arxiv.org/abs/2610.35345 · PDF

  44. 44

    CoPE-CGL: Community-aware Prototype Extraction for Replay-Free Continual Graph Learning

    Bo-Young Lim, Seokhwan Kim, Kisung Lee, Hyuk-Yoon Kwon

    cs.LG · cs.DB

    Continual graph learning (CGL) maintains models as graphs evolve without retraining from scratch, while mitigating catastrophic forgetting. Existing methods typically target either transductive or inductive settings and often store past subgraphs or require task identifiers at inference. We propose Community-aware Prototype Extraction for Continual Graph Learning (CoPE-CGL), a replay-free framework that uses a parameter-free structural prior,...

    arxiv.org/abs/2610.35313 · PDF

  45. 45

    Shared Context, Separate Engines: Streaming Vision--Language Inference on Edge SoCs

    Yilong Li, Chengpo Yan, Yijing Zeng, Hanlei Wu, Mingjun Song, Suman Banerjee

    cs.LG

    Edge SoCs pack CPUs, GPUs, and NPUs, yet vision--language model (VLM) runtimes keep generation on one engine and leave the others idle. Streaming VLMs expose further parallelism: new frames arrive while earlier queries are still being answered. Engines differ in compute and memory access and contend when they co-run, so no single placement of stages wins across SoCs and workloads. Backend-local KV state keeps other engines from continuing the...

    arxiv.org/abs/2610.35309 · PDF

  46. 46

    Inference Latency-Aware Tsetlin Machine Training

    Yefan Zeng, Shengyu Duan, Rishad Shafik, Alex Yakovlev

    cs.LG

    The Tsetlin Machine (TM) is a logic-based machine learning model well suited to low-latency inference on resource-constrained edge devices. Because each clause is an AND over Boolean literals, both the literals and their associated Tsetlin Automaton (TA) actions can be represented at the bit level, enabling bitwise evaluation and an early exit that is triggered as soon as any included literal evaluates to zero, forcing the clause output to...

    arxiv.org/abs/2610.35304 · PDF

  47. 47

    InterQ-RL: Interaction-Aware Correction of Execution Mismatch in Quantized Agentic RL

    Yingbo Hao, Song Liu, Guanghua Yu, Zhaojing Zhou, Jianchen Zhu, Yi Zou

    cs.LG

    Multi-turn agent reinforcement learning (RL) relies on long closed-loop rollouts, making generation and KV-cache costs a major bottleneck. Quantized rollout reduces these costs, but its low-precision executor also serves as the behavior policy and can induce execution mismatch with the learner. Existing quantized-RL methods are designed primarily for single-turn generation. They align the learner forward with quantized rollout or correct the...

    arxiv.org/abs/2610.35303 · PDF

  48. 48

    Conversational Generalized Linear Bandits with Hybrid Feedback

    Yurui Zou, Qirun Zeng, Jinhang Zuo

    cs.LG

    Conversational bandits improve recommendations by actively querying users about key-terms to elicit their item preferences. However, existing methods typically assume homogeneous feedback for items and key-terms, using either absolute rewards or pairwise comparisons. In practice, item feedback often takes the form of clicks or ratings, while comparisons are more natural for key-terms. To address this gap, we study conversational generalized...

    arxiv.org/abs/2610.35262 · PDF

  49. 49

    NOMAD-EEG: Native Omni-Montage Adaptive Diffusion

    Abdulkader Helwan, Khadidja Henni, Lina Abou-Abbas

    cs.LG · eess.SP

    Purpose. Electroencephalography (EEG) montages, channel counts and recording lengths vary widely, but most EEG foundation models are bound to the channel set and input length they were trained on. We aimed to build a generative EEG foundation model independent of the number, identity and order of electrodes and of the input length. Methods. NOMAD-EEG is a 12.0 M-parameter diffusion model that alternates state-space layers along time, without...

    arxiv.org/abs/2610.35246 · PDF

  50. 50

    SkillLifeBench: Benchmarking Security Across the Lifecycle of Agent Skills

    Chenxu Zhang, Zhou Liu, Wu Xin, Wentao Zhang

    cs.LG · cs.AI

    Web agents reuse skills whose stored code, selected version, and loaded implementation can diverge across tasks and updates. SkillLifeBench links these states to task outcomes in 388 paired scenarios spanning implementation, interface, runtime, composition, and version changes. We evaluate 14 models on two paired decision tracks: 89 runtime tasks and 60 autonomous lifecycle tasks spanning 12 templates derived from introduced cases. Lifecycle...

    arxiv.org/abs/2610.35215 · PDF

  51. 51

    One-Shot Federated Model Editing for Device Dynamics at the Edge

    Youming Tao, Weijun Yin, Huanyi Xie, Yanqing Yao, Dongxiao Yu, Xiuzhen Cheng, Di Wang, Falko Dressler

    cs.LG · cs.AI · cs.DC · cs.NI

    Standard federated learning (FL) requires persistent device connectivity across multiple communication rounds, making it unsuitable for dynamic environments such as vehicular networks, where devices frequently join and leave. One-shot FL offers a promising alternative by completing training in a single communication round. However, one-shot FL suffers from model solidification and cannot readily adapt to subsequent client dynamics, including...

    arxiv.org/abs/2610.35212 · PDF

  52. 52

    Efficient Exact Unlearning for Federated Zeroth-Order Optimization

    Weijun Yin, Yanqing Yao, Falko Dressler, Di Wang, Youming Tao

    cs.LG · cs.AI · cs.DC · cs.NI

    Federated zeroth-order optimization (FedZOO) reduces the memory and communication costs of collaborative training for edge intelligence by avoiding backpropagation while communicating only random seeds and scalar coefficients. However, existing methods lack efficient mechanisms for exact unlearning to fully remove the influence of training data when users request deletion. A key insight is that total variation stability of the learning...

    arxiv.org/abs/2610.35211 · PDF

  53. 53

    Understanding Practical Optimization in Machine Learning

    Egor Shulgin

    cs.LG · math.OC

    Modern machine learning relies on practical optimization methods that often differ from the algorithms studied in classical analyses: they train submodels, clip or normalize gradients, use layer-wise non-Euclidean directions, or approximate costly inner computations. This thesis analyzes the objective, assumptions, update, and computation defined by these procedures, rather than those of simplified proxies. For submodel training, we show that...

    arxiv.org/abs/2610.35209 · PDF

  54. 54

    LEAP: a Latent Embedding Anomaly Patch-Transformer for time-series anomaly detection

    Cristhian Moya-Mota, Ignacio Aguilera-Martos, Diego García-Gil, Julián Luengo

    cs.LG

    Time-series anomaly detection is dominated by reconstruction-based neural detectors, whose objective is structurally flawed. Since the target is also the input, an expressive model can learn to reconstruct anomalies too, leaving uninformative residuals. Recent large-scale benchmarks show that, on curated data, elaborate deep detectors rarely beat simple statistical baselines, while current leaders rely on representation learning. We propose...

    arxiv.org/abs/2610.35190 · PDF

  55. 55

    DAC-Loop: Distillation-Aware Compression for Remote Sensing CNN

    Rico Staedeli, Lorena Raichle, Konstantinos Tzevelekakis, Aron Asefaw, Damian Borth

    cs.LG

    While large-scale models achieve strong results in remote sensing scene classification, deploying them on edge devices remains difficult due to limited compute, memory, and energy budgets. Compact CNN backbones are therefore a practical choice in many real-world settings, offering efficient inference with well-optimized implementations on embedded hardware. To push this efficiency further, we investigate how far a knowledge-distilled student...

    arxiv.org/abs/2610.35175 · PDF

  56. 56

    Biology-Guided Context-Aware Residual Propagation for Spatial Gene Expression Prediction from Histology

    Vu Van An, Long D. Nguyen, Ho Viet Duc Luong, Binh P. Nguyen

    cs.LG

    Spatial transcriptomics provides spatially resolved molecular profiling, but its cost, limited throughput and experimental complexity hinder routine large-scale use. Predicting spatial gene expression from hematoxylin-and-eosin-stained histology images offers a scalable alternative, yet remains challenging because histology provides only an indirect morphological view of molecular activity. Existing methods typically decode expression from...

    arxiv.org/abs/2610.35167 · PDF

  57. 57

    Accelerating SO(2)-Equivariant Materials Models through Low-Precision Emulation

    Denghui Lu, Alexander Maeder, Alexandros Nikolaos Ziogas, Mathieu Luisier, Manasa Kaniselvan

    cs.LG · cs.DC

    Rotationally equivariant architectures improve the physical fidelity of machine-learned (ML) universal materials models based on graph neural networks (GNNs). However, their scalability is limited by the use of non-trivial, high-precision tensor operations during both training and inference. Adopting the reduced-precision formats of matrix multiplication units (MMUs), e.g., FP8 instead of FP32, provides massive increases in throughput, up to...

    arxiv.org/abs/2610.35164 · PDF

  58. 58

    Locked In: Multi-Objective RL for the AI Data Center Build-Out

    Satyapragnya Kar, Rajarshi Das, PR Panigrahi

    cs.LG · cs.AI

    AI data center build-out is a multi-trillion-dollar, multi-decade infrastructure program. Yet siting, power procurement, and cooling decisions are still made with static, single-point scoring tools. These tools are evaluated once, on day-zero numbers. This is a poor fit. A site's water, power, and carbon profile drifts substantially over the 10 to 20 year asset life. The governing decisions are also irreversible, and they bind capital and...

    arxiv.org/abs/2610.35147 · PDF

  59. 59

    TUON: A Tensor-Aware Optimizer Utilizing Tensor-SVD and Polarization for Large Language Model Pretraining

    Anda Tang, Xuanlin Yang, Junjie Wang, Zhouchen Lin

    cs.LG

    Large scale language model~(LLM) pretraining is increasingly bottlenecked by optimizer efficiency, driving optimizer design from vector-aware methods such as AdamW to matrix-aware methods. Matrix-aware optimizers such as Muon exploit matrix structure but treat layers independently, neglecting cross-layer correlations. Tensor-aware optimizers are therefore needed, yet existing ones rely on a matricization-based tensor algebra and remain...

    arxiv.org/abs/2610.35130 · PDF

  60. 60

    STeP-TS: Shared Temporal Filters with Pruning for Time Series

    Hanne Raum, Jasper Hoffmann, Joschka Boedecker

    cs.LG

    Multivariate time series forecasting requires capturing temporal dependencies across multiple scales while accounting for both shared and variable-specific dynamics. Compact forecasting models have emerged as an attractive alternative to increasingly large architectures, but preserving strong predictive performance within tight computational and memory budgets remains challenging. We propose a compact one-layer convolutional framework that...

    arxiv.org/abs/2610.35128 · PDF

  61. 61

    UTA: Universal Training Adaptive Learning Rate Decay

    Anda Tang, Junjie Wang, Xuanlin Yang, Cong Fang, Zhouchen Lin

    cs.LG · cs.AI

    Learning rate decay (LRD) is critical for training high-performance models, yet most designs remain empirically driven. Although the Universal Budget Aware (UBA) LRD offers a theoretical framework, its reliance on SGD dynamics limits its applicability to modern Adam-type optimizers and leaves its key `optimization difficulty parameter' opaque. To bridge this gap, we propose the Universal Training Adaptive (UTA) LRD. Unlike UBA's fixed offline...

    arxiv.org/abs/2610.35113 · PDF

  62. 62

    Subspace Crystallization in Grokking: Rank Compression Tracks the Phase Transition

    Shlok Mehendale, Paras Chopra

    cs.LG · cs.AI

    Grokking describes understanding something deeply and intuitively. In neural networks, it is the phenomenon where a network generalizes long after memorizing its training data. This delay has been linked to a fall in the effective rank of the network's weight matrices, but whether this rank compression drives grokking or follows it is still unresolved. We study a mean-field MLP and a one-layer transformer trained on modular arithmetic and...

    arxiv.org/abs/2610.35112 · PDF

  63. 63

    FlowGenesis: Representation-Driven Self-Organization of Neural Architectures

    Maksim A Kazanskii

    cs.LG

    Deep neural networks typically rely on architectures specified before training, despite the continuous evolution of their internal representations. We introduce FlowGenesis, a framework in which neural architectures adapt through neuron birth and pruning guided by Representation Flux, a local statistic capturing changes in minibatch mean activations. Neurons exhibiting persistently high Flux undergo splitting designed to approximately...

    arxiv.org/abs/2610.35108 · PDF

  64. 64

    Spherical Brownian Diffusion for Language Modeling

    Andrea Miele, Ilija Bogunovic

    cs.LG

    Continuous diffusion language models involve two coupled design choices: how to corrupt continuous token representations and how to invert that corruption at generation time. Recent work has explored language generation on the unit sphere through geodesic and von Mises--Fisher (vMF) paths, raising the question of which forward process to use on this geometry and how much that choice matters relative to inference. We introduce Spherical...

    arxiv.org/abs/2610.35101 · PDF

  65. 65

    A JiT-Based Denoising Defense Method Against Adversarial Examples for Radio Signal Automatic Modulation Classification

    Shiwei Liu, Linyuan Wang, Nuolin Sun, Libin Hou, Bin Yan

    cs.LG

    Automatic modulation classification (AMC) of radio signals is a core technology for spectrum monitoring and electromagnetic situational awareness. Deep neural networks (DNNs) play an important role in this field owing to their powerful feature extraction capabilities and end-to-end learning paradigm. However, their vulnerability to adversarial attacks poses serious security challenges to practical deployment. Adversarial training, a common...

    arxiv.org/abs/2610.35084 · PDF

  66. 66

    Reinforcement Learning Done Right for Explicit Confidence Targets

    Zefeng He, Yichen He, Yuan Lin, Yang Yang, Yu Cheng

    cs.LG · cs.CL

    Large language models (LLMs) can produce plausible but incorrect answers when they are uncertain, which can mislead users and undermine model reliability. Training models to abstain with reinforcement learning is a natural approach, but most existing methods use a fixed error penalty and therefore cannot adapt their behavior to different confidence requirements at inference time. We study \emph{explicit confidence targets}, where a user...

    arxiv.org/abs/2610.35059 · PDF

  67. 67

    Risk-Aware Output Control Should be the Focus of Machine Unlearning

    Yahya Alkhatib, Yuyang Xiao, Wee Peng Tay

    cs.LG · stat.ML

    Well-documented failure modes of machine learning (ML) systems, including bias, toxicity, hallucination, and leakage of sensitive information, has driven the rise of machine unlearning (MU). Most existing work, however, targets parameter-space indistinguishability, seeking unlearned models that are as close as possible to models retrained from scratch on retained data. In this paper, we argue that this objective is misaligned with practical...

    arxiv.org/abs/2610.35052 · PDF

  68. 68

    Mapping the Solution-Set Structure of Grokking on Latin Squares

    Zihan Gu, Lulu Zhao, Tianyi Ma, Hua Zhang, Wenbo Su, Yanan Cao, Yue Hu

    cs.LG

    Grokking is a delayed transition from memorisation to generalisation. A network first fits observed examples while remaining near chance on held-out inputs, then generalises abruptly. Prior work has focused mainly on this delay, finding that the network's main loss becomes weight decay after the fitting phase. We name the solution to this optimization problem as the minimum-norm endpoint, and study which tasks make all minimum-norm endpoints...

    arxiv.org/abs/2610.35036 · PDF

This edition is part of The Daily Abstract — cs.LG archive. Subscribe to receive these in your inbox each morning, automatically translated to Spanish, with reply-to-PDF: arxivdaily.ignorelist.com.

Colophon Set in Georgia, with system sans for interface chrome and a monospaced stack for code and paper identifiers. Sole accent: amber #D99C5E. Built and served on an always-free VM. The masthead is set 14% letterspaced because newspapers do that and it works.