cs.LG · 2026-10-25 · No. 153

Machine Learning, 2026-10-25.

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

01 — The papers

56 entries
  1. 01

    Cost-Aware Abstention for LLM-Based IT Ticket Classification

    Sachin Gupta

    cs.LG · cs.AI · cs.CL

    Wrong automatic routing of an IT support ticket can cost more than human triage. We evaluate cost-aware abstention as an offline, cost-conditional readiness assessment: select a routing policy on calibration data under explicit costs, then evaluate that frozen policy once on held-out test data. On a public 47,837-ticket, eight-class corpus, we compare four trained non-LLM classifiers with four large language models under zero- and few-shot...

    arxiv.org/abs/2610.31657 · PDF

  2. 02

    METTA: Memory-Enhanced Test-Time Adaptation for Meta-Reinforcement Learning

    Ali Najar, Hosein Hasani, Amir Kooshan Fattah, Yasamin Rajabi, Mahdieh Soleymani Baghshah

    cs.LG · cs.AI

    Meta-reinforcement learning aims to train agents that adapt quickly to unseen tasks from limited interaction, but this requires representing task evidence over long interaction histories. Classical recurrent agents compress experience into hidden states, while transformer-based agents retrieve from past trajectories at quadratic memory and computation cost. In this work, we propose Memory-Enhanced Test-Time Adaptation (METTA) for...

    arxiv.org/abs/2610.31656 · PDF

  3. 03

    A Fix for Softmax and Cross-Entropy

    Jonathan T. Barron, Dan Gnanapragasam, Michael Tschannen

    cs.LG

    Standard implementations of softmax, log-softmax, and cross-entropy have two numerical bugs. The first bug, Softmax Collapse, rounds the target logit's gradient to zero if that one logit is much larger than the others, resulting in gradients that no longer sum to zero. The second bug, Softmax Cancellation, occurs when student logits are close to teacher logits, and produces non-zero gradients even when the student and teacher logits are...

    arxiv.org/abs/2610.31642 · PDF

  4. 04

    Causal Chain of Thought: Mitigating Attention Leakage via Context Isolation

    Gabriele Dominici, Teague McMillan, Fazl Barez, Marc Langheinrich, Martin Gjoreski

    cs.LG · cs.AI

    Chain-of-thought (CoT) prompting elicits step-by-step rationales that appear to explain model predictions, but there is no guarantee these steps causally drive the final answer. In standard autoregressive generation, dense attention allows each reasoning step to attend to the full context, enabling the model to derive answers through direct lookup while generating plausible-sounding but post-hoc justifications. We introduce CausalCoT, an...

    arxiv.org/abs/2610.31639 · PDF

  5. 05

    BeNice! Closing Looping's Inference-Memory Gap via KV-Sharing

    Hayden Prairie, Jacob M. Ryan, Taylor Berg-Kirkpatrick, Daniel Y. Fu

    cs.LG

    Looped architectures scale quality while keeping parameter count fixed by repeatedly passing activations through the same block of layers. During inference, each recurrence generates independent KV-caches, causing memory to grow linearly. To address this challenge, prior work has proposed KV-sharing (i.e., sharing KV-caches across depth), finding that, in some cases, the KV-cache of the final recurrence can be shared with the earlier...

    arxiv.org/abs/2610.31633 · PDF

  6. 06

    Questioning Instrumental Convergence in RLVR: Checkpoint Evidence from OLMo-3

    Barış Alparslan, Sandy Tanwisuth, Caleb Petterson

    cs.LG · cs.AI

    Instrumental convergence is often invoked as a general reason to expect capable agents to acquire or express power-seeking subgoals: self-preservation, resource acquisition, deception, or resistance to oversight which poses an important threat to the society. For LLMs, reinforcement learning is the first stage of post-training that directly optimizes behavior against an objective; therefore, it is the stage where convergent instrumental...

    arxiv.org/abs/2610.31624 · PDF

  7. 07

    OP-Print: Tracing Membership Imprints in On-Policy Self-Distillation

    Qi Li, Xinchao Wang

    cs.LG · cs.CR

    On-policy distillation has been widely adopted for alleviating the distribution mismatch of off-policy distillation. Self-distillation further removes the need for an external teacher by using the student's own outputs, enhanced with extra context or demonstrations, as supervision. This teacher-free design makes on-policy self-distillation practical, yet its privacy risks remain largely unexplored. In particular, using two coupled views of...

    arxiv.org/abs/2610.31616 · PDF

  8. 08

    Efficient and Optimal Learning of Discrete Distributions with Person-Level Privacy

    Gautam Kamath, Mahbod Majid, Ankur Moitra, Argyris Mouzakis, Jonathan Ullman

    cs.LG · cs.CR · cs.DS

    We study learning a discrete distribution under person-level differential privacy in the batched setting. Given $n$ users, each contributing $m$ i.i.d. samples from an unknown distribution $μ$ over $[k]$, we seek an estimate of $μ$ with total-variation error $α$. We give polynomial-time algorithms under both pure and approximate differential privacy. Under pure $(\varepsilon,0)$-DP, we achieve $n = \widetilde O(k/(α^2...

    arxiv.org/abs/2610.31612 · PDF

  9. 09

    CRANE: Post-Hoc Counterfactual Retrieval for Multimodal Classifiers

    Franco Rugolon, Ioanna Miliou, Panagiotis Papapetrou

    cs.LG

    Counterfactual explanation methods have been introduced for unimodal settings, with dedicated frameworks for tabular, timeseries, text, and image data, yet no post-hoc method has been designed for classifiers that learn from multimodal inputs. Existing explainability methods for multimodal classifiers apply modality-specific attribution independently to each input, which fails to yield coherent explanations and does not transfer to arbitrary...

    arxiv.org/abs/2610.31594 · PDF

  10. 10

    Action Learning for Scientific Simulations

    Sarah Liaw, Yuanqi Du, Yilun Du

    cs.LG

    Understanding and controlling how systems evolve is central to science and engineering, where many problems require reasoning about complete trajectories. Existing approaches typically learn local dynamics or generate trajectories, but objectives constructed from individual transitions can miss dependencies across time, while generative models do not necessarily provide a tractable objective for evaluating and optimizing arbitrary paths. We...

    arxiv.org/abs/2610.31585 · PDF

  11. 11

    Rethinking Sequence Compression with Test-time Regression

    Sihan Yang, Jie Huang, Hanxuan Li, Yifan Guo, Weigao Sun, Yiran Zhong, Weiyang Liu

    cs.LG

    Sequence compression is a fundamental operation and a critical information bottleneck in efficient language and multimodal models, and yet prevailing approaches based on pooling or learned projections remain underperforming. We revisit sequence compression through the lens of test-time regression: input tokens define a regression problem whose fitted coefficients form the compressed representation. Importantly, the regression problem is...

    arxiv.org/abs/2610.31574 · PDF

  12. 12

    On the Token Collapse Phenomenon in Textual Graph Soft Prompting

    Andrea Giuseppe Di Francesco, Rishabh Jain, Franco Maria Nardini, Nicola Tonellotto, Pietro Liò, Fabrizio Silvestri

    cs.LG

    Recent methods for injecting graph structure into LLM prompts include graph soft prompting, in which a graph neural network (GNN) learns graph-specific tokens that capture relational information among entities in textual prompts. G-Retriever combines graph soft prompts with a verbalized description of the same graph to strengthen the textual prompt signal. We conduct an in-depth analysis of G-Retriever and uncover a pathological behavior we...

    arxiv.org/abs/2610.31572 · PDF

  13. 13

    When to Follow the Teacher: Advantage-Based On-Policy Distillation

    Mahdi Sabbaghi, George Pappas, Hamed Hassani, Adel Javanmard

    cs.LG · cs.AI · stat.ML

    Online Policy Distillation (OPD) often outperforms offline methods such as Supervised Fine-Tuning (SFT) by training on trajectories sampled directly from the student. However, standard OPD trains the student to follow the teacher at every state of the trajectory, assuming that teacher guidance is always beneficial. We instead distinguish between what the teacher prefers and what actually helps the current student succeed. We introduce...

    arxiv.org/abs/2610.31563 · PDF

  14. 14

    Diffusion Language Models for Perplexity-Aware Black-Box Attacks

    Ali Al-Lawati, Elsayed Eshra, Dongwon Lee, Suhang Wang

    cs.LG · cs.CL · stat.ML

    \emph{Continuous-space} attacks find prompts that trigger malicious behaviors, such as overthinking, in black-box LLMs by sampling or optimizing over a low-dimensional Gaussian latent space and decoding it to text through a \emph{random mapping} and nearest-token lookup. Because the random mapping is not learned from language, its prompts can be highly unnatural, and as such easily detectable through input guardrails, and the risk estimates...

    arxiv.org/abs/2610.31558 · PDF

  15. 15

    Fine-Grained Analysis of Catastrophic Forgetting in Continual Learning via Algorithmic Stability

    Guodong Zheng, Yan Sun, Peng Wang, Enneng Yang, Jiafei Lyu, Xiaochun Cao, Dacheng Tao, Li Shen

    cs.LG

    Understanding catastrophic forgetting, in which learning new tasks degrades performance on previously learned tasks, remains a fundamental challenge in continual learning (CL). Existing theoretical analyses often rely on specific model or task structures and offer limited insight into how iterative optimization affects forgetting. To address this gap, we develop an algorithmic stability framework for expected population forgetting under SGD...

    arxiv.org/abs/2610.31551 · PDF

  16. 16

    Shift-aware score transport for privacy auditing without retraining

    Yaohong Yang, Gauri Pradhan, Qi Chen, Samuel Kaski, Antti Honkela

    cs.LG

    Privacy auditing empirically assesses privacy leakage by using membership inference attacks to establish lower bounds on privacy loss. We consider an observational auditing setting in which the auditor cannot insert canaries, control data inclusion, or retrain the target model, and must instead rely on fixed sets of known members and non-members. This setting is relevant for large deployed models or when retraining is restricted by...

    arxiv.org/abs/2610.31549 · PDF

  17. 17

    Same Luck, Different Choice: Twin-World Credit Assignment for Agentic Reinforcement Learning

    Zeropy Surio, Tobias Engel, Mei-Lin Chou, Anika Raman, Samuel Okafor

    cs.LG

    Agentic reinforcement learning rewards a complete episode, yet the policy needs credit for every decision inside it. Counterfactual branching estimates this credit by re-running the episode after an alternative decision, but the branch draws its sampling noise independently of the factual episode, so the outcome difference mixes the effect of the decision with the luck of everything that follows. We observe that in agentic search the...

    arxiv.org/abs/2610.31544 · PDF

  18. 18

    PYTHIA, a transferable machine learning framework for simultaneous defect and adsorption site optimization

    Ioannis Kouroudis, Zhaosu Gu, Poonam, Leon Muller, Alessio Gagliardi

    cs.LG

    Machine learning has permeated material optimization, providing high value. Nevertheless, its re- liance on a high number of samples limits its applicability to relatively explored material families. In this paper we extend the applicability of machine learning to novel and understudied materials through the PYTHIA (Probabilistic Yet Transferable Hierarchical Interaction Architecture) framework. Pythia utilizes two of the most powerful...

    arxiv.org/abs/2610.31539 · PDF

  19. 19

    Piano2Pop: Expanding Piano Demos into Pop Productions Without Aligned Supervision

    Yun-Chen Cheng, Chih-Pin Tan, Tzu-Hung Huang, Yi-Hsuan Yang

    cs.LG · cs.SD

    We introduce Piano2Pop, a music-expansion task that transforms piano MIDI into instrumental pop productions. Enabled by recent advances in music generation and polyphonic music transcription, our pseudo-supervised approach conditions a pretrained flow-matching model on piano rolls without aligned piano-pop supervision. We train on synthetic piano-instrumental pairs derived from pop recordings, comparing pair-free transcription, including...

    arxiv.org/abs/2610.31510 · PDF

  20. 20

    SABLE: Scalable Affine Bayesian Latent-space Exploration

    Erik Bodin, Lisa Maria Kreusser, Donney Fan, Henry Moss, Geoff Pleiss, Nicolas Durrande

    cs.LG · cs.AI · stat.ML

    Complex objects such as engineering designs and proteins are often difficult to optimise because their design spaces exhibit high-dimensional structure and objective evaluations are expensive. Generative models provide continuous latent representations, but their latent spaces can have thousands to tens of thousands of dimensions, creating a regime where the latent dimension $D$ greatly exceeds the evaluation budget. We introduce SABLE, a...

    arxiv.org/abs/2610.31495 · PDF

  21. 21

    Training Small Language Models for Transferable Tool Selection

    Zhaoyi Zhou, Akshayaa Magesh, Jeongyeol Kwon, Artem Kuriksha, Daniel Jiang, Andrea Zanette, Ayush Jain

    cs.LG · cs.AI

    As LLM agents operate over corpora of more than ten thousand tools, selecting the right tool for a complex multi-step user instruction becomes a central bottleneck. Tool descriptions cannot all fit in context, and relevance-based retrievers only rank tools without knowing which are actually needed. Recent work instead leverages proprietary commercial LLMs like Claude or ChatGPT to drive a retriever, decomposing the instruction and selecting...

    arxiv.org/abs/2610.31483 · PDF

  22. 22

    Evaluating the Applicability of LLM Inference Optimizations on Apple Silicon

    Kyrylo Yemets, Andrii Yaroshevych, Victor Muryn, Ostap Khomenko, Pavlo Kryven, Maksym Kmet, Maksym Shamrai, Mariya Hirna

    cs.LG · cs.DC

    Apple Silicon is among the most widely used platforms for local LLM inference, yet nearly all inference optimizations are designed and validated on server-class NVIDIA GPUs under batched serving workloads. Which optimizations retain their benefits when moved to Apple Silicon, and what determines their applicability? We evaluate four optimization families: KV-cache pruning, speculative decoding, ported GPU kernels, and post-training...

    arxiv.org/abs/2610.31479 · PDF

  23. 23

    RheOPT: Few-Shot Non-Newtonian Flow Prediction via Newtonian Pretraining of RheOFormer

    Maedeh Saberi, Amir Barati Farimani, Safa Jamali

    cs.LG

    Despite advances in numerical methods, non-Newtonian flow simulations remain computationally expensive because velocity and stress must be advanced together through rate- and history-dependent constitutive dynamics. Neural-operator surrogates reduce inference cost after training, but their training data require the same expensive simulations they aim to replace, making data-efficient learning essential when only a few non-Newtonian...

    arxiv.org/abs/2610.31465 · PDF

  24. 24

    Multi-Token Prediction and Alternative Tokenization Improve Learning Latent Multi-Hop Graph Search at Scale

    Arvid Frydenlund, FuTe Wong, Katherine Lambert, Frank Rudzicz

    cs.LG · cs.CL

    Given an unseen graph, a start node, $s$, and an end node, $t$, can a language model generate the shortest path from $s$ to $t$? Prior work has found that models fail to learn this task as the graphs scale. We show that this failure has two causes: the supervision signal and the tokenization. First, we theoretically establish and empirically demonstrate the effectiveness of multi-token prediction (MTP), and show that the benefits of MTP are...

    arxiv.org/abs/2610.31439 · PDF

  25. 25

    Drift Says When, Budget Says How Much: Rate-Controlled KV Re-selection for Long-Form Multimodal Generation

    Bin Hu, Yingbo Xiao

    cs.LG

    KV cache compression for multimodal LLMs usually selects the visual tokens to keep once, at prefill, and reuses that selection for the whole generation. We show that this one-shot selection is severely limited on long-form generation: on detailed image description, even an oracle that selects with the attention of the uncompressed answer reaches only 0.44-0.52 ROUGE-L agreement with the uncompressed output, because the evidence the model...

    arxiv.org/abs/2610.31433 · PDF

  26. 26

    Predictable performance is suboptimal, and what it means for model routing

    Guanhua Zhang, Moritz Hardt

    cs.LG · cs.AI

    Large language models (LLMs) differ in capabilities and costs, motivating routing methods that adaptively choose a model for each query. These methods rely on performance prediction: given a query, estimate whether a model will answer it correctly. Most routing methods train separate predictors of candidate models' performance to guide model selection. Our work shows that performance prediction only beats a simple baseline when the candidate...

    arxiv.org/abs/2610.31428 · PDF

  27. 27

    Post-Hoc Feature Selection Layer for Neural Networks Interpretability

    João Kenji Suwa, João Pedro Silveira e Silva, Bruno Iochins Grisci

    cs.LG · cs.AI

    The interpretability of complex neural networks remains a critical challenge, especially for models already deployed in high-stakes domains. To address this, we adapted the Feature Selection Layer (FSL) to optimize feature weights post-training, a variant we term post-hoc FSL. Our approach reframes the FSL as a lightweight, trainable module that integrates with already frozen pre-trained models on tabular datasets to highlight the features...

    arxiv.org/abs/2610.31420 · PDF

  28. 28

    SQuAT: Spherical Lattice Vector Quantization-Aware Training

    Sebastian Loeschcke, Ivan Koryakovskiy, Tycho F. A. van der Ouderaa, Riccardo Del Chiaro, Mart van Baalen, Paul N....

    cs.LG · cs.CL

    Vector quantization (VQ), in particular lattice-based methods, offer strong rate-distortion trade-offs for low-bit large language models (LLMs), yet their behavior under quantization-aware training (VQ-QAT) remains poorly understood. We scale lattice VQ-QAT to billion-parameter LLMs and multi-billion-token training regimes, and introduce SQuAT, a shape-gain VQ-QAT approach trained with a separate straight-through estimator (STE) for the shape...

    arxiv.org/abs/2610.31412 · PDF

  29. 29

    Harmful Responses Are Jailbreak Templates: Transferring Autoregressive Jailbreaks to Diffusion Language Models

    Jinwei Gan, Zifeng Cheng, Zhiwei Jiang, Ran Sun, Shiping Ge, Qing Gu

    cs.LG

    Diffusion large language models (dLLMs) enable flexible masked generation through bidirectional context modeling, but this capability also introduces a distinct jailbreak attack surface. We find that a harmful response to a given query naturally serves as an effective jailbreak template: simply masking a subset of its tokens can induce a dLLM to reconstruct harmful content. Based on this observation, we propose RMT-Jailbreak, a general...

    arxiv.org/abs/2610.31331 · PDF

  30. 30

    Bridging the Data Gap: Source-Informed Task-Space Transport for Few-Shot Diffusion

    Jinjian Jiang, Jing Peng, Rehan Deen, Yu Cao, Ziquan Liu, Muhammad Salman Haleem, David Mguni

    cs.LG · cs.AI

    Many deployments require a model to move rapidly from a data-rich source to a related target observed only sparsely. This is especially consequential where target data are difficult to collect, as in healthcare settings with persistent population or acquisition imbalances. Retraining or fine-tuning a high-capacity generator for every target is costly and can be statistically unreliable when target data are scarce. We introduce Meta-Diff,...

    arxiv.org/abs/2610.31323 · PDF

  31. 31

    The Value of System 1 Models in Medicine

    Xiufeng Yang, Konstantinos Vrettos, Zeineb Ben Chaaben, Fabian Drexel, Jannik Lübberstedt, Johannes Moll, Anirudh...

    cs.LG

    Many clinical decisions are a choice among a few fixed options, taken quickly and in large numbers. Such decisions also need an uncertainty estimate to know when to defer to a clinician. Generative language models answer slowly in free text and state their confidence poorly. In contrast, System 1 models answer zero-shot with a probability for each answer. However, their performance across clinical tasks is unknown. Four System 1 models, two...

    arxiv.org/abs/2610.31315 · PDF

  32. 32

    Identifiability from Stationary Covariance in Linear SDEs with Reciprocal Interactions and Latent Variables

    Ana Leticia Garcez Vicente, Saber Salehkaleybar

    cs.LG

    Continuous-time stochastic differential equations (SDEs) provide a natural framework for modeling interacting dynamical systems. A central causal question is whether causal effects among observed variables can be identified from observational data drawn from the stationary distribution, given the underlying causal graph. For linear SDEs, existing identifiability results are largely restricted to fully observed systems whose causal graphs...

    arxiv.org/abs/2610.31310 · PDF

  33. 33

    Extreme Compression of Per-Layer Embeddings via Cross-Layer Latent Vector Quantization

    Jiacheng Liu, Yaxin Luo, Xiaohan Zhao, Peng Sun, Xinyi Shang, Jiacheng Cui, Yonina C. Eldar, Zhiqiang Shen

    cs.LG

    Scaling embedding capacity alongside Transformer computation is an emerging trend in language models. This trend makes high compression rates increasingly useful. Hence, we study extreme compression of Per-Layer Embeddings (PLE), which follow this trend and account for 35.2% of the parameters of Gemma 4 E4B. At such rates, conventional quantizers give each token-layer vector its own code of only a few bytes. We instead develop OPTIC...

    arxiv.org/abs/2610.31309 · PDF

  34. 34

    The Bayesian Learning Rule Beyond KL Geometry

    Sophia Sklaviadis, Wu Lin, Andre Martins, Mario Figueiredo, Thomas Moellenhoff, Emtiyaz Khan

    cs.LG

    The Bayesian learning rule (BLR) exploits the relation between Bayesian updating and the Kullback-Leibler divergence (KLD) to derive efficient additive updates for exponential family variational posteriors. The KL geometry is powerful but restrictive: combined with posterior families with parameter-dependent support it can lead to undefined objectives. We ask which non-KL geometries preserve the additive BLR computation. Previous works that...

    arxiv.org/abs/2610.31308 · PDF

  35. 35

    DISC: Domain-Integrated Similarity Cost for Time-Series Dataset

    Yen-Ku Liu, Hongjie Chen, Franck Dernoncourt, Ryan Rossi

    cs.LG

    The growth of time-series datasets and foundation models has made dataset comparison a central part of time-series workflows. Assessing source-target transfer before model adaptation requires similarity estimates that can summarize transfer-related dataset relationships without exhaustive source-target evaluation over every candidate pair. Many metrics have been proposed for measuring similarity between time-series datasets, but they capture...

    arxiv.org/abs/2610.31288 · PDF

  36. 36

    Deep ReLU Networks Can Fail the Best Approximation Property for Continuous Target Functions

    Sebastian Kassing

    cs.LG · math.NA · math.OC

    We construct a continuous function on $[-1,1]^2$ for which the infimum of the $L^2$-approximation error over the class of ReLU networks with two hidden layers and architecture $(2,2,1,1)$ is not attained. This counterexample shows that deep ReLU networks do not satisfy the best approximation property in general, even for continuous target functions. Our findings contrast with existing results for shallow neural networks, where ReLU is one of...

    arxiv.org/abs/2610.31264 · PDF

  37. 37

    QuaNAS: Uncertainty-Aware Reuse of Earlier Searches for Query-Efficient Neural Architecture Search

    Brama Seta Putra Satria, Dwi Sunaryono, Misbakhul Munir Irfan Subakti, Irzal Ahmad Sabilla, Laras Suciningtyas

    cs.LG

    Neural architecture search (NAS) on a new dataset often affords only tens of short training runs, so what the search knows before its first evaluation decides its efficiency. Yet most NAS methods start from scratch, guided at best by zero-cost proxies that are unreliable across tasks. QuaNAS treats a library of earlier searches in the same search space as explicit knowledge of uncertain reliability. Main effects of design decisions learned...

    arxiv.org/abs/2610.31251 · PDF

  38. 38

    Local Discrimination of Binary Classifiers via Localized AUC Functional

    Wenye Ren, Anna Győrffy-Kerekes, Tancredi Dorsi, Olga Demler

    cs.LG · stat.ML

    Risk prediction models are frequently evaluated with AUC, which means that during evaluation we are blind to performance heterogeneity, which could lead to bad decision-making in clinics. In this work, we aim to change that by identifying patient cohorts for whom the risk prediction model might not be an accurate predictor. To separate these cohorts, we use a version of the weighted AUC, which we call Localized AUC or LAUC. Using this metric,...

    arxiv.org/abs/2610.31216 · PDF

  39. 39

    Evolving Federated Density Peak Clustering with Shuffler-Based Anonymization

    Miha Ožbot, Seiichi Ozawa, Igor Škrjanc

    cs.LG · cs.CR · cs.DC · eess.SY

    In this paper, we present a novel one-shot federated clustering approach based on evolving fuzzy clustering and shuffler-based anonymization. In federated clustering, each round reveals additional model information to the server and can increase the opportunity to infer properties of local datasets, even when raw samples are never shared. Consequently, fewer client-server exchanges and unlinkability are important, so transmitted information...

    arxiv.org/abs/2610.31211 · PDF

  40. 40

    Inside-JEPA: From Prediction to Relational Grounding in Latent World Models

    Hossein Resani, Javen Qinfeng Shi

    cs.LG · cs.CV · cs.RO

    Joint-embedding predictive architectures (JEPAs) learn world models in latent space, enabling efficient planning but leaving their states opaque. We introduce Inside-JEPA, which places relational channels for ordered entity pairs inside the JEPA latent state and shows that an inverse-dynamics loss can ground these channels without relation labels when its head reads them, either alone or as part of the full state. On Something-Something V2,...

    arxiv.org/abs/2610.31194 · PDF

  41. 41

    Towards Field-Deployable Wearable IMU Systems for Occupational Activity Recognition

    Ihsan Ul Haq, James Brusey, Farnad Nasirzadeh, Chandan Karmakar

    cs.LG

    Wearable inertial measurement unit (IMU) systems for construction worker human activity recognition (HAR) routinely report aggregate accuracies above 95 %, rarely under genuine subject-independent (leave-one-subject-out (LOSO)) evaluation, masking the per-individual variance that determines real-world deployment performance. Using the VTT-ConIoT benchmark, we conduct, to our knowledge, the first systematic configuration sensitivity analysis...

    arxiv.org/abs/2610.31160 · PDF

  42. 42

    PriMa-Causa: Causal Prescriptive Maintenance

    Felix Saretzky

    cs.LG · cs.AI · eess.SY

    Automated manufacturing requires data-driven decisions under uncertainty and sparse observational records, and unplanned downtime is a dominant cost of getting them wrong. Prescriptive maintenance recommends a maintenance action against an impending failure rather than merely predicting it. Established reference architectures couple predictive models with rule- or optimisation-based recommendation logic. Selecting a corrective action,...

    arxiv.org/abs/2610.31147 · PDF

  43. 43

    Is High Local Accuracy Enough? Geometric Consensus with Support-Adaptive Statistical Personalization for Federated Learning

    Junyu Lin, Bosong Qi, Pengcheng Shi, Jiayuan Li

    cs.LG

    Real-world deployment of personalized federated learning (PFL) requires adapting to heterogeneous clients while preserving recognition performance beyond local distributions. However, high local accuracy can conceal failures on locally rare or absent classes. Effective personalization should therefore strengthen global recognition while raising the personalization floor. We propose an evaluation protocol that complements local accuracy with...

    arxiv.org/abs/2610.31133 · PDF

  44. 44

    What Comes Next and Why: Interpretable Next-Event Prediction with Neuro-Symbolic Rules

    Tim Nico Bauerschmidt, Isabel Valera, Jilles Vreeken

    cs.LG · stat.ML

    In real-world applications, being able to explain why a prediction is made is often as important - if not more so - than achieving high accuracy. However, existing methods for sequential event prediction typically prioritize accuracy at the expense of interpretability. Neuro-symbolic approaches offer a promising direction to improve interpretability by learning symbolic rules, but extending them to next-event prediction is challenging due to...

    arxiv.org/abs/2610.31114 · PDF

  45. 45

    Transformers with CoT Discover Heavy-Ball Acceleration Through Training

    Yuxuan Cai, Yuan Cao

    cs.LG · stat.ML

    Modern transformers often solve complex tasks by generating and refining intermediate reasoning steps. To develop a theoretical understanding of chain-of-thought (CoT) reasoning, a notable recent study by Huang et al. (2025) viewed transformers as algorithm executors, and showed that when transformers' CoT steps are supervised by a gradient-descent teacher, they can learn to implement gradient descent for in-context weight prediction in...

    arxiv.org/abs/2610.31089 · PDF

  46. 46

    nnInteractive Agent: General-Purpose AI Agents as 3D Image Annotators

    Fabian Isensee, Constantin Ulrich Harsy, Maximilian Rokuss, Benjamin Hamm, Philipp Vollmuth, Klaus H. Maier-Hein

    cs.LG

    Interactive 3D segmentation models such as nnInteractive turn a few spatial prompts (points, scribbles, boxes) into accurate volumetric masks, but a human still does the annotation work: finding each structure, placing prompts, checking the result and correcting it. General-purpose AI agents can now interpret images, call tools and write code. We show that such agents can take over this work without additional training: the user makes a...

    arxiv.org/abs/2610.31080 · PDF

  47. 47

    When Forgotten Facts Return: Relearning-Aware Unlearning of Knowledge Graph Embeddings via Tangent Kernels

    Parnian Fazel, Mohammad M Maheri, Hamed Haddadi

    cs.LG

    Approximate unlearning answers a deletion request on a knowledge-graph embedding by editing entity embeddings instead of retraining from scratch. These embeddings keep being updated as new facts arrive. Under this ordinary continued training, forgotten facts recover more than in the exact-unlearning baseline, a model retrained without them and given the same continued training. We trace this relearning to entity embeddings shared by forgotten...

    arxiv.org/abs/2610.31075 · PDF

  48. 48

    Federated Differentially Private Optimization via Low-Rank Preconditioning

    Antti Koskela, Tejas Kulkarni, Elias Raninen

    cs.LG · cs.CR

    We consider differentially private (DP) optimization in the federated setting, where a DP estimate of the feature covariance matrix serves as a Hessian approximation for generalized linear models. Existing central DP methods for covariance estimation have communication cost that grows quadratically in the data dimension, while existing baselines with linear communication cost are only locally private and degrade as the number of clients...

    arxiv.org/abs/2610.31069 · PDF

  49. 49

    XALIGN: A Human-centric Visual Analytics Approach to Understanding AI Explanations

    Bahavathy Kathirgamanathan, Gennady Andrienko, Natalia Andrienko, Tatiana von Landesberger

    cs.LG

    Explainable AI (XAI) methods are usually evaluated for technical properties such as faithfulness and robustness. In high-stakes application domains, experts also need to check model predictions against established knowledge. Different explanation methods often disagree, and this disagreement is hard to see when each method is inspected separately. We propose XALIGN, a visual analytics workflow and interface that shows rule-based explanations...

    arxiv.org/abs/2610.31015 · PDF

  50. 50

    EEG-Captioner: Grounding EEG-to-Language Generation in Fundamental Signal Properties

    Leyuan Jin, Jiahao Mei, Yixuan Yuan, Shimeng Li, Mengyue Wu, Yue Ding

    cs.LG

    Electroencephalography (EEG)-to-language models offer a promising approach for translating complex neural signals into human-readable descriptions. However, generating plausible language does not necessarily demonstrate that a model has learned to interpret the underlying EEG signal. A direct way to assess such interpretation is to probe objectively measurable properties of the signal itself. We therefore construct an EEG-language dataset of...

    arxiv.org/abs/2610.30995 · PDF

  51. 51

    Energy and Latency First: How Deployment Constraints Reverse Time-Series Model Selection

    Shruti Bothe, Illyyne Saffar, Ayush K. Varshney

    cs.LG

    AI models, and time-series models in particular are typically ranked by accuracy and deployed under constraints absent from that ranking. In this paper, we study a real setting where these constraints are explicit and measurable: handover pre- diction in cellular networks, where inference runs on the User Equipment once per 40 ms measurement period and must complete before the network acts. The timing bound is hard: a late prediction is...

    arxiv.org/abs/2610.30965 · PDF

  52. 52

    Doubly Robust Distributionally Robust Offline Contextual Pricing

    Min Xu, Xinyi Yin, Yunfan Zhang, Yuxuan Han, Houcai Shen, Caihua Chen

    cs.LG

    Offline contextual pricing often relies on logged observational data, but faces challenges from distributional shifts between training and deployment environments. Distributionally robust optimization (DRO) provides a principled approach to off-policy evaluation and learning (OPE/L). However, existing methods are mostly limited to discrete actions. Recent work has explored DRO for continuous treatments using inverse propensity weighting...

    arxiv.org/abs/2610.30957 · PDF

  53. 53

    Ground-Truth-First Benchmarking of Reinforcement Learning via Converse Optimality: From Smooth, Fully Observed Systems to Non-Smooth or Partially Observed Tasks

    Grégoire Ouerdane, Sinan Ibrahim, Stefan Streif, Pavel Osinenko

    cs.LG

    Benchmarking in reinforcement learning is usually relative: algorithms are ranked by returns on environments with unknown optima. Distance to optimality remains unobservable, while conclusions may depend on reward design, stochasticity, hyperparameters, environment selection, and simulator details. We develop a ground-truth-first framework based on converse optimality, constructing each task with a certified optimal value function and policy...

    arxiv.org/abs/2610.30954 · PDF

  54. 54

    OpenAttention: One Percent Is Enough with Tail Proxies

    Hyunwoo Yu, Jongbeom Lee, Hyunbin Kim, Sungjun Park, Jaemin Choi, Jincheol Yang, Seunghun Moon, Bongjoon Hyun,...

    cs.LG · cs.AI

    Sparse attention reads only a small subset of the key-value cache, and how to treat the unread tail remains a key challenge. Discarding the tail biases the output, and sampling it as previous approaches do fails at reads of 5% and below, where the tail likely holds heavy hitters the selector missed. Our key observation is that the simplest possible proxy, one token that folds the entire tail into its count-aware mean, already yields lower...

    arxiv.org/abs/2610.30951 · PDF

  55. 55

    Random Initialization as a Baseline for Time Series Foundation Models

    Pinar Sungu Isiacik, Georgiana Ifrim

    cs.LG

    Time series foundation models (TSFMs) use self-supervised pretraining to learn reusable representations for downstream tasks such as multivariate time series classification (MTSC). We examine how much pretrained weights contribute beyond the architecture itself by comparing seven publicly released instances from four TSFM families with architecture-matched randomly initialized counterparts on the UEA MTSC archive. For each instance, five...

    arxiv.org/abs/2610.30950 · PDF

  56. 56

    Same molecule, Different Knowledge: Cross-Identifier Binding Failures in LLMs For chemistry

    Khouloud Saadi, Di Wang

    cs.LG

    Chemical AI workflows routinely represent the same molecule through multiple co-referring identifiers, including trivial names, IUPAC names, and SMILES, implicitly assuming that these representations provide interchangeable access to an LLM's knowledge. To test this assumption, we first introduce a controlled benchmark and show that it does not hold. Across foundation models spanning 7B to 72B parameters, even chemistry-specialized models,...

    arxiv.org/abs/2610.30903 · 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.