cs.LG · 2026-10-20 · No. 149

Machine Learning, 2026-10-20.

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

01 — The papers

65 entries
  1. 01

    Text-to-Image Generative Modeling Without Language Pretraining

    Keshigeyan Chandrasegaran, Yue Zhao, Armin W. Thomas, Ehsan Adeli, Jiajun Wu, Juan Carlos Niebles, Li Fei-Fei, Michael Poli

    cs.LG · cs.CV

    Modern text-to-image (T2I) systems use internet-scale language pretraining for text-conditioned image generation. We ask whether language pretraining is necessary for strong T2I generation. We introduce Halcyon, a family of early-fusion multimodal autoregressive models pretrained entirely from scratch on GPIC, a dataset of 100M image-text pairs. Halcyon-0.52B outperforms LlamaGen-XL/8 (48.0 vs. 63.3 gFD) with 18.2$\times$ less pretraining...

    arxiv.org/abs/2610.26863 · PDF

  2. 02

    WeAct: Dynamic Weight Switching as Agent Actions

    Rui Shi, Xunjian Yin, Jiacheng Sang, Zichuan Liu, Shuyan Zhou

    cs.LG · cs.AI · cs.CL

    Long-horizon agents must learn heterogeneous capabilities from interleaved trajectories. Updating shared parameters to learn new capabilities can degrade previously acquired ones, resulting in catastrophic forgetting. As a trajectory unfolds, different stages may require different capabilities, yet conventional joint training and static adapter merging maintain a fixed parameter configuration throughout execution. We introduce WeAct...

    arxiv.org/abs/2610.26847 · PDF

  3. 03

    Kronecker Delta Rule for Long Context Linear Attention

    Anamaria-Roberta Hartl, Levente Zólyomi, Michael List, Günter Klambauer, Sebastian Böck, Sepp Hochreiter

    cs.LG · cs.AI

    We propose KronLSTM, a linear attention architecture based on the Kronecker Delta Rule, a novel update rule for associative matrix memories. KronLSTM extends the xLSTM family with a memory-revision mechanism that corrects targeted associations while modifying the memory only along a single direction in matrix space. This minimally invasive memory revision reduces interference with previously stored information compared with delta-rule...

    arxiv.org/abs/2610.26834 · PDF

  4. 04

    Next Best Action: Causal Constrained Offline RL for Business Action Targeting

    Surya Ramachandiran, Yiming Li, Juan Mancilla-Caceres, Tao Ye, Jun Liu

    cs.LG · cs.AI · cs.IR · math.OC

    Next Best Action (NBA) systems must select among customer interventions (e.g., discounts, promotions, and lifecycle actions), maximize long-term economic value, satisfy complex operational constraints, and remain causally grounded. Existing approaches lack a rigorous framework for meeting these requirements jointly. We address this gap by formulating NBA as a Constrained Markov Decision Process (CMDP). We then integrate off-policy causal...

    arxiv.org/abs/2610.26817 · PDF

  5. 05

    Scaling Parameter-Efficient Fine-Tuning for Brain Encoding

    Zhuofan Josh Ying, Guanglong Sun, Nolan Ho, Kieran Liu, Rui Yang, Andrew Chen, Nikolaus Kriegeskorte, Pony Ma, Hossein Adeli

    cs.LG

    Brain encoding models aim to predict neural responses to stimuli. The most widely used approach fits readouts from frozen pretrained backbones. Using neural data to train large models from scratch or fully fine-tune them after task-pretraining has proven difficult because neural data is costly and limited. We show that parameter-efficient fine-tuning (PEFT) enables us to effectively adapt very large pretrained backbones for brain encoding...

    arxiv.org/abs/2610.26802 · PDF

  6. 06

    Learning to Refer: Inference-Time Collaboration in Decentralized Federated Learning

    Yuheng Ma, Chendi Wang, Qiong Zhang

    cs.LG · stat.ML

    In decentralized federated learning, heterogeneous clients may develop complementary expertise, yet each query is typically evaluated only by the client where it arrives. To exploit this distributed expertise without global coordination, we formulate inference-time collaboration as budgeted local expert referral, a graph-constrained routing process in which a query can move only between neighboring clients and only visited clients contribute...

    arxiv.org/abs/2610.26800 · PDF

  7. 07

    Two hidden layer suffice: the caterpillar net

    Jakob Paul Zimmermann

    cs.LG

    Every continuous piecewise linear (CPWL) function is a finite signed sum of maxima of affine functions, making maximum representation a central problem in the study of the expressiveness of ReLU networks. A now-disproved conjecture predicted logarithmic depth for maxima. Inspired by the proof by Brunck et al. (2026) that two hidden layers suffice for general CPWL functions, we give the Caterpillar Net, a nonrecursive construction for the...

    arxiv.org/abs/2610.26795 · PDF

  8. 08

    On Data Scaling in Tabular ML: More Data, Better Generalization?

    Ivan Rubachev, Nikita Lazarev, Artem Babenko

    cs.LG

    Recent work in tabular machine learning has emphasized the importance of diverse and realistic evaluation settings. The niche of large-scale (1M+) data, however, remains under-explored. To study training-data scaling, we assemble eight real-world tabular datasets with up to around ten million training examples, retaining the natural distribution shifts of the tasks and using random-split evaluation as a control. We find that GBDTs and...

    arxiv.org/abs/2610.26794 · PDF

  9. 09

    Training with ML-labeled Data via Stochastic First- and Zeroth-Order Optimization

    Maria-Eleni Sfyraki, Jun-Kun Wang

    cs.LG · math.OC

    Training a machine learning (ML) model is an optimization process that typically requires a set of examples with ground-truth labels beforehand. Obtaining these ground truths usually requires laborious human effort and can be costly. The scalability of human labeling becomes a concern when there are a large number of data points that need to be labeled. To address this practical issue, we consider an optimization setting where a subset of...

    arxiv.org/abs/2610.26790 · PDF

  10. 10

    Topological Sheaves of Irreducibility: A Theoretical Framework for Multi-Agent Task Obstructions

    Steven Motta, Davide Mambelli, Alessandro Rozza

    cs.LG · cs.MA

    Multi-agent systems can differ under identical agents and communication budgets when they support different forms of joint interaction. We introduce Topological Sheaves of Irreducibility (TSI), a sheaf-theoretic framework separating task-level compatibility obstructions from system capability. A task is represented by an interaction complex $K$ and a capacity sheaf $\mathcal{F}$; nonzero $H^n(K;\mathcal{F})$ defines a structural signature,...

    arxiv.org/abs/2610.26777 · PDF

  11. 11

    HarnessPro: Recursive Harness Self-Improvement with Pareto Optimization for Performance-Cost Trade-offs

    Zhuoshi Pan, ZhiYuan Feng, Yukuo Cen, Honglin Lin, Junru Lu, Zhongxia Chen, Shuo Yang, H. Vicky Zhao, Jie Tang

    cs.LG · cs.CL

    Harness development plays a crucial role in building capable AI agents. Recent approaches improve harnesses through recursive self-improvement (RSI), but their performance gains often come with higher execution costs. We introduce HarnessPro, a label-free method for evolving harnesses to improve task performance while reducing execution cost. In this setting, candidate harnesses are evaluated using LLM-generated tests without access to...

    arxiv.org/abs/2610.26770 · PDF

  12. 12

    Simulation Priors for Rapid Adaptation in Model-Based RL for Autonomous Racing

    Adrian Remonda, Jiajun Xi, Nicklas Hansen, Marcus Greiff, Thomas Lew, John Talbot, John Subosits, Xiaolong Wang

    cs.LG · cs.RO · eess.SY

    Reinforcement learning enables learning high-performance control policies, but its reliance on extensive interaction limits deployment on physical systems where data collection is costly, and failures are unsafe. We introduce Prior-Informed Adaptation (PIA), an offline-to-online model-based RL framework that uses structured simulation priors for rapid adaptation. PIA pretrains a world model on heterogeneous low- and high-fidelity simulated...

    arxiv.org/abs/2610.26757 · PDF

  13. 13

    Trajectory Replay: Continuous-Time Mixture Transport for Lifelong Generative Models

    Alessandro Palma, Charles Jones, Tristan Cinquin, Fabian J. Theis, Frederik Wenkel

    cs.LG

    Adapting pre-trained continuous-time generative models to sequentially incoming data without forgetting earlier tasks is a central challenge in lifelong learning. A leading strategy, generative replay, uses a generative model trained on historical tasks to augment its own continual update with synthetic samples from experience. However, it requires full trajectory simulations at every training step, which accumulate discretization errors and...

    arxiv.org/abs/2610.26749 · PDF

  14. 14

    recAMPLIFY: Large-Scale Pretraining of Recursive Protein Language Models

    Saurav Jha, Qianli Liao, Robert Vernon, Sarath Chandar, Quentin Fournier

    cs.LG · cs.AI · cs.CL

    Protein language models (pLMs) are powerful tools for protein annotation and mutation-effect prediction, but progress has largely depended on scaling model size and data, making them expensive and prone to memorization. Test-time scaling offers a complementary path, but methods developed for autoregressive LLMs do not transfer directly: masked encoders lack a generation loop, and amino acids do not verbalize in a reasoning language. Instead,...

    arxiv.org/abs/2610.26747 · PDF

  15. 15

    CAVA: Complementary-Label-Aware Data Valuation

    Tan-Ha Mai, Hsuan-Tien Lin, Vu Nguyen

    cs.LG · cs.CV

    Complementary-label learning (CLL) trains a classifier from labels that name a class an instance does not belong to. Such labels are cheap to collect from crowdworkers or vision-language models (VLMs), making them attractive for large-scale annotation. However, both sources can mistakenly assign the true class as the complementary label. Such annotation errors are particularly harmful in CLL because the model is explicitly trained to suppress...

    arxiv.org/abs/2610.26734 · PDF

  16. 16

    Cross-Fitting for Neural Posterior Estimation

    Jeffrey Regier

    cs.LG · stat.ML

    Neural Posterior Estimation (NPE) fits an amortized conditional density estimator to a simulated training set of complete-data realizations. NPE's training objective, log-likelihood, favors posterior approximations that assign arbitrarily high density to the training points, and typical families of amortized conditional density estimators (e.g., normalizing flows, Gaussian encoders) are flexible enough to oblige. The resulting policy of...

    arxiv.org/abs/2610.26733 · PDF

  17. 17

    Copying Under Repetition: From Prefix Matching to Position Hashing

    Anxin Guo, Tianlong Huang, Chenxiao Yang, Zhiyuan Li

    cs.LG · cs.CL

    Frontier LLMs can copy random text almost flawlessly, yet they often overrun or truncate long blocks of repeated characters. In repeated strings, matching local content no longer reliably identifies where copying should continue or when it should stop. Nevertheless, we show that a two-layer transformer with $O(\log\log N)$-bit precision and $O(\log N/\log\log N)$ width can copy every string up to a chosen design horizon $N$ and terminate...

    arxiv.org/abs/2610.26722 · PDF

  18. 18

    Deep Pontryagin Learning for Continuous-Time Stochastic Control with Jumps

    Jean-Loup Dupret, Patrick Cheridito

    cs.LG · eess.SY · math.OC

    We develop a deep Pontryagin learning framework for high-dimensional continuous-time stochastic control problems that naturally accommodates jump-diffusions. In contrast to direct policy optimization, which repeatedly differentiates the objective through the entire time-discretized dynamics, our method separates adjoint learning from policy improvement by explicitly incorporating state-sensitivity information from the Pontryagin stochastic...

    arxiv.org/abs/2610.26718 · PDF

  19. 19

    Doubly Robust Learning of Heterogeneous Survival Treatment Effects under Time-Varying Treatments

    Tomàs Garriga, Munib Mesinovic, Luke De Bretton-Gordon, Gerard Sanz, Axel Brando, Tingting Zhu

    cs.LG

    Estimating heterogeneous survival treatment effects (HSTEs) from time-to-event data under time-varying treatments is important in fields such as medicine (e.g., to decide which sequence of treatments is better in terms of survival probability). However, this task is challenging because treatment decisions, patient characteristics, and censoring probability often evolve together, requiring sequential adjustment for both treatments and...

    arxiv.org/abs/2610.26710 · PDF

  20. 20

    GeoSplit: Selective Edge--Server Map Acquisition for Cross-User Travel-Mode Recognition

    Zepu Wang, Luyao Niu, Ruolin Li, Jeff, Ban

    cs.LG · cs.DC

    Networked mobility sensing couples local trajectory inference with remote geospatial context, making communication decisions part of the prediction system. We present GeoSplit, an edge--server framework for travel-mode recognition that separates the utility of map evidence, its training-time transfer, and its selective acquisition at runtime. On GeoLife with 10% labels, we evaluate these components under user-day and user-disjoint grouping....

    arxiv.org/abs/2610.26707 · PDF

  21. 21

    Parallel Language Models via Variational Decoding

    Felix Draxler, Robin C. Geyer, Daniel Geyfman, Farrin Marouf Sofian, Justus Will, Philipp Hildebrandt, Giosue...

    cs.LG · cs.CL

    Autoregressive language models emit one token per forward pass. This causes high inference latency, in particular as models are scaled up for quality. In this paper, we introduce Parallel Language Models (PLMs), which generate mutually consistent tokens in one forward pass by routing dependencies through latent auxiliary variables. PLMs are trained using a variational bound that reduces to an autoregressive loss plus a parallel decoding loss....

    arxiv.org/abs/2610.26693 · PDF

  22. 22

    Intra-Rubric Reward Divergence: Mitigating Rubric Hacking through Latent Dimension Recovery and Balancing

    Jianqiang Xia, Han Xu, Dianxi Shi

    cs.LG · cs.AI · cs.CL

    Rubric hacking limits both performance and training robustness in rubric-based reinforcement learning. By decomposing rubrics into sub-dimensions, we reveal the Intra-Rubric Reward Divergence: as the total rubric score increases, certain sub-dimension scores systematically decrease, suggesting the model sacrifices specific quality aspects to optimize the aggregated objective. From the sub-dimension perspective, we observe two phenomena: (1)...

    arxiv.org/abs/2610.26685 · PDF

  23. 23

    Which Model for Which Role? Cost-Efficient Pareto Recommendation for LLM Agentic Workflows

    Qian Xie, Wenyue Hua, Yueli He, Sripad Karne, Armaan Agrawal, Nikos Pagonas, Eugene Wu, Kostis Kaffes, Tianyi Peng,...

    cs.LG · cs.CL · stat.ML

    LLM agentic workflows increasingly employ different models for specialized roles. Users may prefer different trade-offs between task quality and deployment cost, sometimes also considering latency. Finding good assignments incurs a separate search cost from evaluating multiple candidates. We introduce Cost-Coupled Gittins (CC-Gittins), a multi-objective extension of single-objective cost-aware Gittins search for Pareto recommendation. It...

    arxiv.org/abs/2610.26676 · PDF

  24. 24

    PACT: Plan-Augmented Coordination via Trace-Graphs for Dynamic Multi-Agent Combinatorial Optimization

    Xiao Zhou, Alex Schutz, Victor-Alexandru Darvariu, Nick Hawes

    cs.LG

    Multi-agent approaches have been investigated recently as a means of improving the scalability of reinforcement learning for neural combinatorial optimization. Existing methods assume that all problem instance information is known upfront. However, real-world problems are often dynamic: relevant information is progressively revealed as the agents take actions (e.g., arriving at an edge in a routing problem only to discover it is...

    arxiv.org/abs/2610.26671 · PDF

  25. 25

    Nested Shared-Operator Recursion on ARC-AGI-1: An Empirical Study of HRM, TRM and URM

    Sara Neri, Eugenio Gaeta, Andrea Ceschini, Andrea Facchinetti, Giuseppe Fico

    cs.LG

    Recursive reasoning models refine persistent internal states by repeatedly applying learned operators. We investigate whether adding a persistent intermediate state that integrates repeated updates of a faster state through the same operator improves predictive performance relative to each model family's baseline at matched parameter counts. We define this organisation operationally and implement it within three architectures: the...

    arxiv.org/abs/2610.26666 · PDF

  26. 26

    The Effectiveness and Durability of Pretraining Data Filtering in Protein Language Models

    Aaron Maiwald, Roderick McNeill, Jaeyoung Lee, Matthew A. Clarke, Lennart Justen, Alejandro Acelas, Ivo Andrews,...

    cs.LG · cs.AI

    Pretraining filtering is one of the most common safeguards for open-weight language models. Here we study its effectiveness, durability and scalability using the ESM-2 family of protein language models. We find that filtering viral proteins from the pretraining data meaningfully degrades masked language modelling, mutation effect prediction and long-range contact prediction on viral proteins, while largely preserving performance on non-viral...

    arxiv.org/abs/2610.26651 · PDF

  27. 27

    VaDar: Implicit Goal Conditioning via Value Disaggregation

    Shashwat Saxena, Mehul Goel, Sreyas Venkataraman, Sarvesh Patil, Max Simchowitz

    cs.LG · cs.AI · cs.RO

    Long-horizon tasks are important for modern AI and robotics, yet pose a challenge for reinforcement learning. Historically, a popular approach has been to decompose a task into subgoals, and learn a subgoal-conditioned policy to execute each subtask in sequence. We introduce an alternative called Value Disaggregation (VaDar), which trains a single subgoal-independent policy to maximize a linear combination of per-subgoal critic values....

    arxiv.org/abs/2610.26629 · PDF

  28. 28

    SleepPairNet: Learning Multimodal Pediatric Sleep Representations with Random-Pair Cross-Attention

    Saurav Raj Pandey, Harlin Lee

    cs.LG

    Analyzing physiological sleep signals is important for assessing sleep quality and detecting sleep-related disorders. Learning inter-modal relationships across physiological signals can provide richer representations for downstream sleep analysis. We introduce \textbf{SleepPairNet}, a multimodal self-supervised learning (SSL) framework that learns pairwise interactions between physiological signals through cross-attention. We propose a random...

    arxiv.org/abs/2610.26628 · PDF

  29. 29

    Global Time-Aware Conditioning (GTC): A Mechanism for Time-Aware Pediatric Sleep Representation Learning

    Saurav Raj Pandey, Harlin Lee

    cs.LG

    Sleep physiology evolves throughout the night, with changes in sleep cycles, respiratory stability, and cardiovascular dynamics. However, self-supervised learning (SSL) methods for sleep typically process short, fixed-duration windows without explicitly modeling where each window occurs within the full sleep session. We introduce Global Time-Aware Conditioning (GTC), a mechanism that complements local positional encodings by conditioning...

    arxiv.org/abs/2610.26618 · PDF

  30. 30

    Degree Bounds for Separating Polynomial and Neural Invariants for Point Clouds

    Snir Hordan, Harm Derksen, Mircea Petrache, Nadav Dym

    cs.LG · math.AC

    Neural networks that respect the symmetries of point clouds commonly compute high-order representations of the rotation group to improve their expressivity. Such models are known to be universal, and thus able to separate any pair of point clouds, when using representations of sufficiently high order, yet no precise bound on this order, as a function of the number of points, was previously known. We establish the first quadratic lower and...

    arxiv.org/abs/2610.26612 · PDF

  31. 31

    ITU-DESC: a missingness-aware machine-learning dementia risk score in the UK Biobank cohort

    Abdullah Kavaklı, Faik Boray Tek

    cs.LG

    Background: Existing dementia risk scores rely on fixed, pre-selected features and linear models, or ignore bias from informative missingness in cohort data. Here we introduce the Istanbul Technical University Dementia Risk Score (ITU-DESC), an explainable machine-learning score for incident all-cause dementia that combines broad feature selection across roughly 8,000 candidate features with explicit control of missingness bias. Results: We...

    arxiv.org/abs/2610.26604 · PDF

  32. 32

    Bombadil: A Foundation Model for Plant Electrophysiology

    James Kirk, Piotr Bednarski, Filip Twardy, Justin Mast, Lee von Kraus

    cs.LG · eess.SP · q-bio.QM

    Living plants, like animals, use electrical signals to carry information about their physiological state across their body, from roots to branches. These signals can be recorded, as they are in animals, from a pair of differential extracellular electrodes inserted into signal-conducting tissue. While machine-learning models have been trained to predict specific plant stress states from these signals, each of these models has been trained for...

    arxiv.org/abs/2610.26597 · PDF

  33. 33

    Concerns in Benchmarking Jailbreaks: An Empirical Study of Attack Success Rates

    Leon Tu Nguyen, Rohan Menon, Rafael Tappe Maestro, Eric Aubinais, Chih-Hong Cheng

    cs.LG · cs.AI

    Jailbreak methods are commonly compared using Attack Success Rate (ASR), but these comparisons depend on how responses are generated and judged. In this paper, we examine eight attacks against three open-weight language models using three evaluation procedures. Low-resource language attacks achieve a mean ASR of 96.0% under keyword matching but 2.7% under a Qwen3 evaluator following StrongREJECT scoring instructions. Changing the number and...

    arxiv.org/abs/2610.26595 · PDF

  34. 34

    A Fisher-Geometric View of Multi-Teacher On-Policy Distillation

    Liu Kaiyuan, Zhuang Ziyuan, Ma Haoxuan, Gu Yannian, Song Mingfei, Weng Rongxiang, Ye Jieping

    cs.LG · cs.CL

    Consolidating specialized foundation models into a unified generalist is a central goal of post-training. Two approaches are in common use: model merging, which is fast but treated as a heuristic, and multi-teacher on-policy distillation, which matches distributions but demands heavy rollout compute and suffers from cross-task gradient conflict. We show that the two are one objective seen at different orders. Within a shared basin of...

    arxiv.org/abs/2610.26525 · PDF

  35. 35

    Strength Controls Change How Interventions Compare Across Sparse Autoencoder Codes

    Jinchen Zhang, Zheyuan Liu, Yihan Zhu, Meng Jiang

    cs.LG · cs.CL

    Changing sparse autoencoder (SAE) codes can alter both intervention direction and magnitude, so comparing intervention effects requires a strength control specifying what is held equal. Norm matching and output calibration provide different controls, but whether interventions yield different effects across codes after strength matching remains unclear. We compare intervention effects under these controls for fixed feature groups in Gemma and...

    arxiv.org/abs/2610.26489 · PDF

  36. 36

    Wet-Biased Data Pipelines Mislead the Training and Evaluation of Generative Precipitation Nowcasting

    Christoph Metzl, Stefan Kapsreiter, Kianusch Vahid Yousefnia, Virginia Poli, Richard Müller, Frank Krüger, Tobias Bölle

    cs.LG · physics.ao-ph

    Generative models for precipitation nowcasting are trained and evaluated predominantly on wet-biased datasets, which over-represent precipitation events to emphasize rare extremes. Selecting test cases based on target observations is known to favor forecasts that match the selected distribution rather than the intended deployment distribution. We show analytically how this target-dependent sampling also shifts the conditional distribution...

    arxiv.org/abs/2610.26486 · PDF

  37. 37

    A Simple and Unified Method for Gradient-Variation Regret

    Hang Yu, Peng Zhao

    cs.LG · math.OC

    Gradient-variation online learning has attracted increasing attention for its connections to game theory and optimization. It has been studied in both non-stationary online learning, which aims to minimize dynamic regret against time-varying comparator sequences, and universal online learning, which seeks to adapt to unknown loss curvature. Prior work obtained gradient-variation bounds using different techniques tailored to each setting. We...

    arxiv.org/abs/2610.26473 · PDF

  38. 38

    On Learning Expand-and-Sparsify Representations

    Alireza Asghari, Parikshit Ram, Kaushik Sinha

    cs.LG

    Expand-and-sparsify ($\texttt{EaS}$) representations lift an input to a much higher dimension through a linear map and keep only the $k$ largest coordinates. Existing theory builds that map at random and needs the expanded dimension to grow exponentially in the input dimension, which puts the construction out of practical reach. We ask what is gained by learning the map by gradient descent instead, and how that gain depends on how wide the...

    arxiv.org/abs/2610.26470 · PDF

  39. 39

    Rethinking LLM Fine-Tuning via Weight Space Reparameterization: Preserving Safety during Downstream Adaptation

    Minseong Kim, Jongbok Won, Jeesup Park, Yonghee Choi, Dongjun Han

    cs.LG · cs.AI

    Fine-tuning large language models (LLMs) on downstream tasks often weakens their previously aligned safety behavior, creating an inherent trade-off between task adaptation and safety preservation. Existing approaches attempt to mitigate this issue by identifying safety-relevant neurons, layers, or update directions in the original parameter space. However, safety-related information is often distributed across many directions in this space,...

    arxiv.org/abs/2610.26443 · PDF

  40. 40

    A Simple and General Approach to top-$k$ Selection

    Motti Goldberger, Nils Rudi

    cs.LG · stat.ML

    This paper considers the Bayesian top-$k$ identification problem, wherein a decision maker is tasked with selecting $k$ items from a pool of items of unknown quality. To assess the items, the decision maker adaptively allocates a fixed sampling budget, where each sample of an item yields a noisy observation of its quality, with the objective of minimizing the expected opportunity cost of the selected set. We propose an approach that allocates...

    arxiv.org/abs/2610.26407 · PDF

  41. 41

    CTLite: Lightweight Irregular Multivariate Time Series Forecasting with Reusable Continuous-Time Summaries

    Shusen Ma, Xudong Zuo, Yangchun Xie, Yun-Bo Zhao

    cs.LG

    Irregular multivariate time series forecasting (IMTSF) requires capturing temporal dynamics and cross-variable dependencies from sparse, asynchronous observations. Existing methods often rely on learned temporal encoders, coupling historical representation construction with model optimisation and repeatedly processing the same observations. We propose CTLite, a lightweight forecasting framework that decouples historical summarisation from...

    arxiv.org/abs/2610.26379 · PDF

  42. 42

    An Empirical Study of Parameter-Efficient Fine-Tuning for Mixture-of-Experts LLMs

    Minglei Li, Peng Ye, Haonan He, Jingqi Ye, Qingyu Yang, Tao Chen

    cs.LG · cs.CL

    Parameter-efficient fine-tuning (PEFT) methods are critical for adapting large models under resource-constrained settings. However, as mixture-of-experts (MoE) architectures, which offer efficient inference at massive parameter scales, gain increasing prevalence in the community, a critical gap arising from the lack of systematic recipes for applying PEFT to MoE models. To narrow this gap, we conduct a large-scale systematic study that...

    arxiv.org/abs/2610.26370 · PDF

  43. 43

    Logit Averaging During Training Sets a New State of the Art for Non-Invasive Speech BCI

    Miran Özdogan, Gereon Elvers, Michael Bronstein, Oiwi Parker Jones

    cs.LG

    Non-invasive brain-computer interfaces are in practice limited by signal-to-noise ratio, and recent state-of-the-art methods address this limitation by aggregating over repeated trials of the same stimulus. Yet multi-trial aggregation has been treated inconsistently across the literature: some work averages brain signals at the model's input, other work averages predictions at the model's output, and neither treats aggregation as an...

    arxiv.org/abs/2610.26347 · PDF

  44. 44

    CIS-Transolver: A Cross-Layer Incremental-State Neural Operator for Fine-Scale Automotive Aerodynamic Flow Prediction

    Jianglong Liu, Zhicheng Zhu, Jia Hao, Liangyue Jia

    cs.LG · cs.AI

    Neural operators offer an efficient data-driven approach to rapid prediction of complex automotive aerodynamic flows. However, spectral bias and other factors limit accurate reconstruction of fine-scale flow variations near complex geometries. Existing studies improve the representation and fitting of fine-scale information mainly through information extraction and training objectives. Further study is needed to determine how extracted local...

    arxiv.org/abs/2610.26337 · PDF

  45. 45

    Typed Decision Models under Prevalence Shift: Decision Training, Calibration, and Implicit Priors

    Abdul Rahman Iqbal, Muhammad Fawad Nasim, Amir Saleem

    cs.LG · cs.CL

    Typed decision models (TDMs) return a probability over caller-defined options, and Jev is described by its vendor as calibrated. But in deployment class frequencies change, so calibration measured on a fixed test set need not hold where thresholds are applied. Existing evaluations leave two gaps: the gains of vendor decision training have not been separated from those of the backbone under an identical readout, and TDMs have been tested only...

    arxiv.org/abs/2610.26325 · PDF

  46. 46

    CAST: Cast-Admissible Generative Design and Shape Optimization with Paired Depth Images

    Jihoon Kim, Yongmin Kwon, Namwoo Kang

    cs.LG

    Deep generative design can produce complex geometries, but unconstrained representations may create undercuts that prevent release from a two-part mold. We present CAST, which reformulates three-dimensional shape generation under a selected mold-opening constraint as a two-dimensional image generation problem using paired top and bottom depth images. Each valid depth pair defines one material interval along a selected mold-opening direction,...

    arxiv.org/abs/2610.26309 · PDF

  47. 47

    Toward Online Robust Zero-Sum Markov Games with Function Approximation

    Chenhao Zhou, Jiyu Wei

    cs.LG · cs.AI

    Online robust zero-sum Markov games pose a coupled operator-learning and game-solving problem: from nominal interaction data, the learner must recover a worst-case minimax Bellman operator while maintaining strategic coverage and controlling stage-game approximation error. Existing robust reinforcement-learning theory does not yet provide such an operator-learning perspective beyond specialized multi-agent settings. We develop an optimistic...

    arxiv.org/abs/2610.26286 · PDF

  48. 48

    Learning SOCP Solver Dynamics with Graph Neural Networks

    Adam Ghalem, Lou Salaün, Davide Rinaldi, Fragkiskos D. Malliaros

    cs.LG

    Second-order cone programs (SOCPs) are a fundamental class of convex optimization problems widely applied across engineering domains. Classical interior-point methods are robust and accurate, but their iterative nature can be costly in real-time or repeated-solve settings. This has motivated learning-based solvers that approximate optimizer solutions with neural networks. Recent work has increasingly studied graph neural networks (GNNs) for...

    arxiv.org/abs/2610.26285 · PDF

  49. 49

    Trust Your Targets: Is Target Propagation Block-Diagonal Gauss-Newton?

    Giuseppe Castiglione, Christopher Buckley, Ivor Simpson

    cs.LG

    Target propagation (TP) is a backpropagation alternative in which local targets carry the global learning signal. Prior works claim that TP performs Gauss-Newton (GN) optimisation on these targets, but the correspondence has only been established under unrealistic assumptions where hitting them exactly is trivial. Practical variants instead descend on a least-squares problem via a single gradient step. In this paper, we rederive TP from a...

    arxiv.org/abs/2610.26283 · PDF

  50. 50

    SEAM: Spectral Embedding Alignment and Merging for Low-Resource TTS

    Walaa Othman, Baher Mohammad, Stamatios Lefkimmiatis, Ammar Ali

    cs.LG

    Expanding the vocabulary of conditional text-to-speech (TTS) models typically requires extensive fine-tuning with paired audio, creating a major bottleneck for low-resource and cross-lingual adaptation. To address this, we introduce SEAM, a zero-audio initialization method that constructs embeddings for unseen tokens from known subtokens. Unlike standard pooling strategies that ignore sequential structure, SEAM applies a spectral filter to...

    arxiv.org/abs/2610.26252 · PDF

  51. 51

    OceanTensorCube and ST-MMT: A Multimodal Data and Transformer System for Seven-Day Forecasting of Pyropia Bleaching

    Gwangjin Kim, Cheolyoung Jang, Yosep Jo

    cs.LG

    Bleaching of cultured Pyropia (laver) causes losses in Korean seaweed farming. We present a system that forecasts bleaching seven days ahead for farm clusters. OceanTensorCube aligns 32 channels from public marine, meteorological, ocean-model and satellite sources on a 30 m grid, and ST-MMT, a 9.2-million-parameter spatiotemporal transformer with hub-style multi-head cross-attention between variable groups, maps a 14-day input window to a...

    arxiv.org/abs/2610.26209 · PDF

  52. 52

    TaQFormer: A Treatment-as-Query Transformer with Data-Fusion for Multidimensional Treatment Modeling in Coupon Allocation

    Jiahong Li, Zeqin Yang, Yu Xiang, Zhifan Jiang, Jingqi Ma, Enzheng Hua, Jixing Xu, Jun Wu, Fei Wang, Enyun Yu, Jiecheng Guo

    cs.LG

    We study coupon allocation on an online marketplace offering multiple service categories. Issuing coupons requires choosing a vector of category-specific discounts, accounting for both their direct effects and cross-category demand shifts. We present TaQFormer, a treatment-as-query Transformer for this multidimensional treatment setting. Each treatment coordinate is represented as a query token, allowing cross-attention to estimate the...

    arxiv.org/abs/2610.26199 · PDF

  53. 53

    Dynamical Batching: How Class-Cycling Training Shapes the Internal Structure of Neural Networks

    Nicolás Ratier-Werbin, Samuel Lozano, Miguel Ruiz-García

    cs.LG · cond-mat.dis-nn · cond-mat.soft

    We investigate whether cyclically varying the class composition of training batches can modify both the performance and the internal organization of neural networks. Motivated by nonequilibrium systems, in which cycling between related sub-goals can make switching between their solutions increasingly efficient, we introduce Dynamical Batching (DB), a sampling protocol that periodically emphasizes each class while leaving the loss and the...

    arxiv.org/abs/2610.26156 · PDF

  54. 54

    Nash-MOPD: Optimizing Task Trade-offs in Multi-Teacher On-Policy Distillation

    Siyuan Gan, Bo Yang, Linjian Meng, Boyan Wang, Xiaolong Qi, Jing Huo, Yang Gao

    cs.LG · cs.AI

    Multi-teacher on-policy distillation (MOPD) aggregates the capabilities of multiple domain-expert teachers into a single student model. However, all domains in MOPD share the student's finite budget of optimization steps and rollouts. Fixed uses constant domain sampling ratios based on corpus sizes; by contrast, D$^3$-MOPD adjusts the ratios using the teacher-student reverse-KL divergence and its recent decrease. However, favoring large...

    arxiv.org/abs/2610.26124 · PDF

  55. 55

    Optimal Multi-Kernel Combination for Online Multi-Label Classification

    Tingting Zhai, Wei Liu, Renjian Yu

    cs.LG · cs.CV

    Multi-label classification is particularly challenging in streaming scenarios, where data arrives sequentially. Online kernel methods offer a promising solution. However, prevailing methods rely on a single preselected kernel, making them prone to suboptimal choices since the global data distribution is unknown in advance in online settings. Although the recent method, Online Kernel Selection for Online Multi-label Classification (OKS-OMC),...

    arxiv.org/abs/2610.26123 · PDF

  56. 56

    Variance-Aware Non-Stationary Bandits:All-Budget Regret Bounds and Adaptation Costs

    Xiaojin Zhang

    cs.LG

    Low reward variance reduces estimation cost, but does not eliminate monitoring of arms that may improve while unobserved. We refine the variance--variation tradeoff and characterize how both costs depend on supplied variance and variation budgets in bounded, non-stationary multi-armed bandits. Let $A$ be the number of arms, $T$ the horizon, $B$ a bound on consecutive mean variation excluding the initial vector, and $Q$ a bound on the sum of...

    arxiv.org/abs/2610.26116 · PDF

  57. 57

    From Retrieval to Recognition:Knowledge-Enhanced Post-training of Time Series Foundation Models for Classification

    Zhen Liu, Yucheng Wang, Yingpeng Du, Tianjun Wei, Qianli Ma, Min Wu, Jie Zhang, Zhu Sun

    cs.LG

    Time series foundation models (TSFMs) have achieved notable progress in classification tasks through large-scale pretraining. However, their post-training practice mainly focuses on implicit parameter fine-tuning, where task-relevant information is entangled with abundant irrelevant knowledge in vast parameter spaces. This may hinder TSFMs from precisely activating useful knowledge for downstream classification. To this end, we propose...

    arxiv.org/abs/2610.26114 · PDF

  58. 58

    Declarative Explanations for Graph Neural Networks: A Demonstration

    Cibo Yu, Haitong Tang, Yinghui Wu, Arijit Khan, Tingting Zhu, Tingyang Chen, Xiangyu Ke

    cs.LG

    Graph Neural Networks (GNNs) have proven effective for graph analytical tasks, yet their "black-box" nature complicates the understanding of their layer-wise behavior and hinders their application to large graphs. We demonstrate SliceGXQ, an end-to-end declarative and interactive system for querying and exploring layer-wise explanations for GNNs over large-scale graphs. In particular, SliceGXQ provides (1) an interactive graphical user...

    arxiv.org/abs/2610.26104 · PDF

  59. 59

    Learning from a Bandit Learner

    Michela Petriconi, Nicolas Menet, Claire Vernade

    cs.LG

    Classical inverse bandit methods typically aim to recover true reward parameters. In contrast, we infer a demonstrator's subjective beliefs, specifically its internal estimates of suboptimality gaps. We consider no-regret demonstrators that follow randomized policies, such as Thompson sampling, and characterize a family of algorithms for which per-step inversion is possible. This enables gap estimates to be recovered directly from the...

    arxiv.org/abs/2610.26103 · PDF

  60. 60

    Provable Adversarial Detection: Prime Quantization meets Gromov--Wasserstein

    Manohar Kaul, Sadbhavana Babar, Aditya Ramesh, Aditya Saibewar, Fumiyoshi Kasahara, Maki Inui, Hiroaki Kurihara

    cs.LG

    Adversarial vulnerability persists across modern vision architectures from CNNs to vision language models (VLMs), yet existing detection methods rely on heuristics without theoretical guarantees. We address the fundamental question of when adversarial perturbations can be provably detected from a geometric perspective. Our key insight is that adversarial perturbations cannot simultaneously preserve geometric structure across spaces with...

    arxiv.org/abs/2610.26102 · PDF

  61. 61

    FinHopper: A Progressive Multi-Hop Benchmark for Complex Financial Agents with Step-wise Evaluation

    Yuxuan Wu, Miaopei Lin, Qiufang Ying, Lin Dan, Hao Zhao, Qirui Liu, Yuteng Shen, Shaohui Wu, Weihong Luo, Xiku Du,...

    cs.LG

    Financial agents are expected to answer factual investment queries through open search, but financial multi-hop retrieval differs from general multi-hop QA in a crucial way: each intermediate answer is an exact typed constraint, such as an entity, date, event, or numerical indicator, that determines the next query. This strict answer-conditioned dependency makes small intermediate errors cascade into plausible but wrong search trajectories....

    arxiv.org/abs/2610.26036 · PDF

  62. 62

    Truncate or Not? A Scaling Study of Looped Transformers with Truncated Backpropagation

    Xuyuan Xiong, Siyu Chen, Zihan Shao, Yue Su, Weiyang Jin, Ping Luo, Zhuoran Yang, Xihui Liu

    cs.LG · stat.ML

    Looped transformers reuse a shared block of layers to increase computational depth without increasing the number of trainable parameters. They are trained with backpropagation through time (BPTT). To save training compute, BPTT is often truncated to the last few loops, although this may weaken the benefit of looping. To study this tradeoff systematically, we jointly vary model size, training tokens, and forward and backward loop counts across...

    arxiv.org/abs/2610.26030 · PDF

  63. 63

    ProtGlycanDock: Towards Accurate Protein-Glycan Docking with Tailored Dataset, Benchmark and Model

    Minghao Xu, Jiaze Song, Jinyu Wang, Wentao Zhang

    cs.LG

    Performing accurate protein-glycan docking is critical to understanding protein-glycan interactions which govern many biological processes. However, this important task is hampered by the lack of tailored datasets, benchmarks and models for protein-glycan docking. To fill this blank, we first curate a standard benchmark ProtGlycanDock for training and evaluating protein-glycan docking models under both protein and glycan generalization...

    arxiv.org/abs/2610.26009 · PDF

  64. 64

    Scenario Generalization and Sector Transfer in Real Surface Irrigation Surrogate Modelling with Tabular Models

    Alessandro Cagiano, M. Aybars Seyhan, Fabiola Gangi, Matteo Zignani, Daniele Masseroni

    cs.LG

    Surface irrigation planning requires screening many operating conditions and spatial configurations, while high-fidelity shallow-water-equation (SWE) simulations are expensive. We study tabular surrogate models for predicting maximum water height from drone-derived field morphology, terrain, hydrological, spatial, and scenario descriptors, where each row represents a spatial cell under one irrigation scenario. To assess deployment...

    arxiv.org/abs/2610.26005 · PDF

  65. 65

    PnP-CF: A Consensus Flow-Matching Solver for General Inverse Problems

    Dibyanshu Kumar, Magda Gregorova

    cs.LG

    Plug-and-Play (PnP) methods using continuous-time deep generative models as learned data distribution priors enable synthesizing high-quality solutions of model-based inverse problems. The quality of the solutions depends critically on the interplay and consensus between the inverse problem solver and the generative model sampler. We revisit a recently introduced flow-based PnP framework, which we refer to as PnP-CF (Consensus Flow), and...

    arxiv.org/abs/2610.26000 · 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.