cs.LG · 2026-11-09 · No. 168

Machine Learning, 2026-11-09.

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

    On the (non-)closedness of the set of ReLU networks

    Eloi Martinet, Leon Bungert

    cs.LG · math.CA · math.FA · math.NA · math.OC

    We prove that the set $\mathcal{N}_N$ of ReLU networks with one hidden layer and at most $N\in\mathbb N$ neurons is weakly closed in $W^{1,p}(Ω)$, for a bounded connected open set $Ω\subset\mathbb R^d$. Consequently, every coercive, weakly lower semicontinuous energy on this space attains its minimum over $\mathcal{N}_N$. This covers the Deep Ritz method with convex integrands, Sobolev training and Lipschitz learning. Moreover, as the weight...

    arxiv.org/abs/2611.09887 · PDF

  2. 02

    Computer-Use Agents Are Embodied Generalists

    Pranjal Aggarwal, Sean Welleck

    cs.LG · cs.RO

    In many areas of AI, general-purpose models have increasingly replaced handcrafted methods and specialized architectures. We argue that embodied AI is poised for a similar shift. Our key claim is that embodied tasks can be formulated as interaction with digital software. Specifically, we build software that renders an agent's view of a physical environment on a screen and provides controls for acting in it. A computer-use agent (CUA) can then...

    arxiv.org/abs/2611.09847 · PDF

  3. 03

    Fifty-Two Shades of Pig: What the UCR Archive Really Asks Classifiers to Do

    Omar Swelam, Jaris Küken, Shi Bin Hoo, Frank Hutter, Lennart Purucker

    cs.LG

    The pigs were bleeding. That was the point of the experiment: could airway pressure reveal hemorrhage before it became critical? In the UCR Time Series Classification Archive, the de facto benchmark for Time Series Classification (TSC), the same recordings ask a different question: "which of 52 pigs produced this trace?". The dataset PigAirwayPressure is not an outlier. Although the archive is the standard benchmark, whether its tasks...

    arxiv.org/abs/2611.09843 · PDF

  4. 04

    Designing Effective Monitor-Based Interventions for Mitigating Reward Hacking During RL

    Aria Wong, Joshua Engels, Neel Nanda

    cs.LG

    Reinforcement learning (RL) rewards are notoriously difficult to design and control, often leading to the model learning unintended behaviors such as reward hacking. One potential solution is to monitor for reward hacking and penalize it when detected; however, training against a monitor could lead to evasive behavior, and our general understanding of how to apply monitors effectively during training is limited. To study how best to use...

    arxiv.org/abs/2611.09828 · PDF

  5. 05

    The Cost of Coverage: Conformalising Native Forecast Uncertainty for GB Wind Power

    Carl Somers

    cs.LG · stat.AP

    Uncertainty bands on wind forecasts help inform reserve sizing and commitments, but often fail to deliver their stated coverage. We apply conformal prediction to native bands to see what it adds and what it costs, comparing six forecasters, from ARIMA to a tabular foundation model, at two wind farms in Great Britain (GB) under operational conditions. We show that calibration repairs empirical marginal coverage with little or no Winkler Score...

    arxiv.org/abs/2611.09815 · PDF

  6. 06

    Scaling Graph Similarity Learning with Graph Neural Networks

    Ahmet Tüzen, Helge Langseth, Kjetil Nørvåg

    cs.LG · cs.AI

    Graph similarity computation is an important task in graph mining with many real-world applications, such as chemical compound analysis, program dependence comparison, and social network analysis. The problem is computationally challenging: Graph Edit Distance (GED) and Maximum Common Subgraph (MCS), which quantify the distance between graphs and identify their largest shared subgraph, respectively, are NP-hard to compute exactly. To address...

    arxiv.org/abs/2611.09805 · PDF

  7. 07

    Quasipolynomial-Time Planning for Decodable POMDPs

    Oliver Dobson, Varun Jog, Chun Hei Michael Shiu

    cs.LG · cs.DS · cs.IT · math.OC · stat.ML

    We study planning for finite-horizon partially observable Markov decision processes (POMDPs) with finite state, action, and observation spaces. Both exact and approximate planning are known to be computationally hard in this setting, so prior work has explored natural assumptions that make planning easier. Recent work [Golowich et al., 2022] gave quasipolynomial-time approximate planning algorithms for observable POMDPs. These are POMDPs...

    arxiv.org/abs/2611.09778 · PDF

  8. 08

    From PAC bounds on the generalization error to PAC bounds on the parameter estimation error

    Mihaly Petreczky, John Leth, Emma Lukacs, Balint Daroczy

    cs.LG

    Probably Approximately Correct (PAC) bounds on the generalization error (the difference between the empirical prediction error computed on the training data and its expectation) are an established tool in statistical machine learning, and there have been several extensions of these bounds to dynamical systems. In parallel, in recent years, there has been growing interest in finite-sample bounds for parameter estimation error in system...

    arxiv.org/abs/2611.09767 · PDF

  9. 09

    A Tight Hierarchy for Chain of Thought

    Thanh Le, A. Pavan, N. V. Vinodchandran

    cs.LG · cs.CC

    Chain-of-thought (CoT) reasoning has emerged as a powerful paradigm for extending the computational capabilities of transformer-based models. A growing body of recent work has begun to uncover the fine-grained computational power of transformers augmented with CoT, as well as their connections to classical complexity theory. In this work, we advance this line of research by establishing the first hierarchy theorem for CoT-based computation....

    arxiv.org/abs/2611.09759 · PDF

  10. 10

    Diffusion-free Score Matching for Exact Conditional Sampling

    Thomas Wedenig, Robert Peharz

    cs.LG

    The success of modern score-based generative models is largely driven by diffusion-based training, which learns a collection of time-indexed score fields corresponding to progressively noisier versions of the data distribution. While such models achieve impressive sample quality, leveraging them for sound probabilistic inference---particularly sampling from arbitrary conditionals of the model---remains challenging. This difficulty arises...

    arxiv.org/abs/2611.09758 · PDF

  11. 11

    Safety Is $Π_1$, Provability Is $Σ_1$: One Framework for the Hugging Face Breach and OpenAI's Mathematical Proofs

    Jinho Kang, Jinsook Kim

    cs.LG · cs.AI · cs.CR · cs.LO

    OpenAI's agents breached Hugging Face in July 2026, and the same company released more than seven hundred AI-generated manuscripts of mathematical proofs in October 2026. The first event is read as a failure in security, while the second is read as a success in mathematics. We read both through the hybrid framework of our paper (Kim and Kang, 2026, arXiv:2611.01914), in which coding, proving, and learning from data are derivations by one and...

    arxiv.org/abs/2611.09736 · PDF

  12. 12

    One Run Is Not Enough: A Monte Carlo Protocol for Evaluating LLM Reliability on Data Science Tasks

    Wenqi Ni, Ozan Evkaya, Miguel de Carvalho

    cs.LG

    Large language models (LLMs) have shown promise for automating data science workflows. Benchmarks such as DataSciBench (2026) and StatQA (2024) have advanced the evaluation of their capabilities across data science and statistical tasks. Assessing their reliability requires examining both the quality of their responses and their consistency across repeated runs, task types, and difficulty levels. With this aim, we introduce DS-QA, a benchmark...

    arxiv.org/abs/2611.09729 · PDF

  13. 13

    Fixed-point acceleration on Neural Wasserstein manifolds

    Vitalii Aksenov, Sebastian Gutierrez Hernandez

    cs.LG · math.OC

    We propose a fixed-point acceleration method for optimization over probability measures, applicable to a variety of machine learning problems. We represent probability measures as pushforwards of a reference distribution through a neural network map and equip the parameter space with the induced pullback Wasserstein metric. This geometric structure allows natural gradient descent to be interpreted as a fixed-point iteration on a Riemannian...

    arxiv.org/abs/2611.09723 · PDF

  14. 14

    ICare-DT: a digital twin for continuous forecasting of intensive care patients

    Yanke Li, Yahang Qi, Yumi Kim, Diego Paez-Granados

    cs.LG · cs.AI

    Digital twins of intensive care patients should update forecasts from new data and predict treatment responses, but few studies test whether routine records support either. We propose ICare-DT, a Transformer-based twin that updates a latent patient state hourly from new measurements without retraining. We trained it on 140,473 eICU-CRD stays to forecast vital signs crossing their training-set means. On test data, later forecasts beat the...

    arxiv.org/abs/2611.09712 · PDF

  15. 15

    RidgeFill: Training-Free Late-Layer Cache Fill for Faster Prefill

    Dylan Johnston, Robert P. Smith

    cs.LG · cs.CL

    RidgeFill shortens prefill by not computing most of a model's upper layers for the prompt. The lower layers run exactly up to a split; from the split layer's state a per-layer map predicts every later layer's input, and each late layer writes its cache from that prediction. The design follows YOCO and DeepSeek-V4.1-Flash, where the projection is trained into the model. RidgeFill retrofits it onto models that were never trained for it: the map...

    arxiv.org/abs/2611.09709 · PDF

  16. 16

    Verifier Noise Is a Temperature: Testing Tilt Predictions in Finite-Budget LLM-Guided Search

    Xiaomeng Li

    cs.LG · cs.AI

    LLM-guided search loops propose a modified candidate, score it, and keep it if it scores better. When both candidates are re-scored at every comparison, as in the minibatch acceptance gates of prompt optimizers, keep-if-better is a reversible Markov chain if the score difference is logistic and the mutation operator is reversible, and its stationary law is the mutation equilibrium tilted by exp(beta f). For any continuous symmetric noise...

    arxiv.org/abs/2611.09699 · PDF

  17. 17

    The Multiscale Single-Index Model: An MSE Analysis of Online SGD Dynamics

    Gordon Dai, Joan Bruna

    cs.LG

    We consider the Multiscale Single-Index Model (MSIM), first introduced by Oymak and Soltanolkotabi (2021), as a stylized model for hierarchical feature learning with scale separation. Each layer extracts a shared single-index feature at one physical scale and passes it to the next, giving a tractable setting in which to study how deep architectures learn local multiscale representations. Under non-degeneracy and delocalization assumptions on...

    arxiv.org/abs/2611.09682 · PDF

  18. 18

    Online Causal Configuration for Networked Systems via Doubly Robust Steady-State Learning

    Yuli Liu, Zhiheng Zhang

    cs.LG · stat.ML

    Measurement exploration reveals how interacting local controllers behave but can alter the steady-state regime a network operator seeks to optimize. Exploratory rewards may therefore fail to identify a configuration's no-exploration value: under full exploration, identical logged distributions can conceal different target values. Identification requires accounting for both action-law mismatch and shifts in the steady-state context...

    arxiv.org/abs/2611.09665 · PDF

  19. 19

    Encoding Memory in World Models with Test-Time Model-Based Reinforcement Learning

    Maxime Burchi, Fayaz Ali Dharejo, Radu Timofte

    cs.LG · cs.AI · cs.CV

    Model-based reinforcement learning (MBRL) approaches have demonstrated strong performance across diverse domains in recent years, learning accurate world models from agent trajectories. However, despite this progress, current approaches still struggle to surpass human performance on tasks that require long-term memory and planning. In this work, we address the problem of long-term memory in MBRL by moving beyond the commonly studied paradigm...

    arxiv.org/abs/2611.09625 · PDF

  20. 20

    BiLoGo: Reducing the smoothing of localised responses in Transformer-based neural operators for sheet forming

    Yingxue Zhao, Haoran Li, Haosu Zhou, Tobias Pfaff, Nan Li

    cs.LG

    Trial-and-error FE simulation of sheet forming could be computationally expensive, neural-network-based learned simulators are used to accelerate design. These simulators should accurately capture localised responses, such as thinning around sharp radii or near ribs, which determine formability. Transformer-based neural operators propagate information efficiently across the mesh but tend to smooth such localised responses. This study develops...

    arxiv.org/abs/2611.09621 · PDF

  21. 21

    Know What You Need: Efficient Activation Checkpointing

    Danyal Ahmed, Edouard Oyallon

    cs.LG

    Large Language Model training is increasingly constrained by memory capacity. Activation Checkpointing reduces activation memory by avoiding storing selected activations during the forward pass and rematerializing them during backpropagation, thereby trading memory savings for additional computation. Although this technique is essential for scaling model training, rematerialization can be prohibitively expensive, particularly for Transformer...

    arxiv.org/abs/2611.09614 · PDF

  22. 22

    Transformer-FSQ: Discrete Latents for Extreme KV-Cache Compression

    Andras Szecsenyi, Thomas L. Lee, Vaisakh Shaj, Elliot J. Crowley, Amos Storkey

    cs.LG

    Private, personalised language models must run on local consumer devices, where large, frontier models cannot fit, and retaining long contexts is non-trivial. Sequential memory---the key-value (KV) cache---grows linearly with context length, quickly dwarfing the model itself: for Qwen3-1.7B-Base, 200,000 retained entries require 21 GiB, 6.6 times more than the model weights. Yet the KV-cache compression methods we measured fail beyond...

    arxiv.org/abs/2611.09598 · PDF

  23. 23

    Coordinatewise Epsilon for Adam under Heavy-Tailed Noise

    Jeonseong Kim, Chenglin Fan

    cs.LG

    Adam estimates second moments coordinatewise, but typically adds the same scalar $\varepsilon$ to every denominator. We study how this shared $\varepsilon$ affects convergence when noise and curvature vary across coordinates. Under coordinatewise finite-$p$ noise moments, $p\in(1,2]$, choosing $\varepsilon_i$ from the noise and curvature bounds for each coordinate attains the heavy-tailed $\ell_1$-stationarity benchmark...

    arxiv.org/abs/2611.09596 · PDF

  24. 24

    Solver-Free Decision-Focused Learning via Neural Optimization Layers for Edge Selection

    Xinda Zheng, Frits de Nijs, Edward Lam

    cs.LG · math.OC

    Decision-focused learning trains prediction models to improve downstream decisions, but many methods require costly solver calls or optimal-solution labels. We propose a solver-free decision-focused learning framework for NP-hard edge-selection problems with uncertain parameters in the objective, the constraints, or both. A neural optimization layer is first trained without solution labels, using problem objectives and constraint penalties....

    arxiv.org/abs/2611.09560 · PDF

  25. 25

    Marginal-Nonuniform Multiclass Learning

    Nataly Brukhim, Steve Hanneke, Amirreza Shaeiri, Maximilian Thiessen

    cs.LG · stat.ML

    We study the problem of multiclass classification within the framework of marginal-nonuniform learning, where the constants in the learning rate may depend on the unknown marginal distribution generating the unlabeled data, but not on the target concept. This notion naturally lies between classical PAC learning, where the constants are independent of both the marginal distribution and the target concept, and more recent universal learning,...

    arxiv.org/abs/2611.09543 · PDF

  26. 26

    Active Learning of Robust Models

    Mizu Nishikawa-Toomey, Andreas Kirsch, Laurent Charlin, Dhanya Sridhar

    cs.LG · cs.AI

    In many domains from species identification to cancer diagnosis, labelling data to train predictive models remains costly. In these settings, it is advantageous to intelligently select data points to label to achieve the maximum information gain, this is the motivation behind the study of active learning. In addition to the high costs of acquiring data in such domains, there is the added complexity of data distribution shifts that can occur...

    arxiv.org/abs/2611.09520 · PDF

  27. 27

    TimeOmni-Live: Continuous & Live Time Series Understanding with Long-Term Memory

    Tong Guan, Alexander Chemeris, Yiqing Guo, Warren Jin, Zhao Li, Shirui Pan, Ming Jin

    cs.LG

    Existing time series language models predominantly follow an offline, one-shot paradigm: they process fixed-length inputs independently and do not retain past states. This static approach fundamentally misaligns with the continuous nature of real-world streaming applications and fails to maintain the historical context necessary for cross-window reasoning. To bridge this gap, we introduce Continuous & Live Time Series Understanding (CL-TSU),...

    arxiv.org/abs/2611.09502 · PDF

  28. 28

    e-means: outperforming k-means (in one-tenth the epochs)

    Swann Bessa, Pierre-Emmanuel Mazaré, Gergely Szilvasy, Hugo Touvron, David Picard, Matthijs Douze, Hervé Jégou

    cs.LG · cs.AI

    Lloyd's k-means algorithm is a long-running standard for clustering, with applications in data analysis, retrieval, quantization and data curation. In this paper, we introduce an exponential moving average optimization procedure for k-means with mini-batch updates, powered by modern deep learning machinery and new insights. We also establish scaling laws on how decay rate and batch size impact final performance. This gives us a formulaic rule...

    arxiv.org/abs/2611.09501 · PDF

  29. 29

    Mosaic:montage-agnostic self-distilled graph learning for subject-independent ieeg seizure detection

    Arya Jalali, Mahsa Shoaran

    cs.LG · cs.AI

    Intracranial EEG provides high signal-to-noise ratio and spatial resolution for detecting epileptic activity, but patient-specific electrode placement leads to substantial variation in channel count and montage across subjects. This heterogeneity limits the generalization of existing detectors under subject-independent evaluation. We propose MOSAIC (MOntage-agnostic Self-distilled Adaptive Inter-channel Connectivity learning), a lightweight...

    arxiv.org/abs/2611.09496 · PDF

  30. 30

    Consensus-Constrained Dual-Federated Optimization with Client-Side Dynamic Weighting for Distributed Android Device Threat Defense

    Mohammad Mashreghi, Mohammad Hossein Badiei, Md. Jalil Piran

    cs.LG · cs.CR · cs.DC

    Smartphones, smart-home hubs, and wearables are increasingly exposed to sophisticated Android malware, while centralized threat-detection pipelines raise privacy and scalability concerns. Federated learning (FL) enables collaborative model training without collecting raw client data, but partial participation and variation in local model quality can destabilize global optimization. In this article, we propose DW-FedADMM, a federated...

    arxiv.org/abs/2611.09470 · PDF

  31. 31

    From LoRA to Language: Recovering Skill Documents from Adapters

    Saranya Dintakurthy Swadha, Pavan C Shekar, Arpit Jain, Aswanth Krishnan

    cs.LG · cs.AI · cs.CL

    Agent skills are short documents that teach a language model how to handle a class of tasks. To save context, these documents are increasingly compiled into low-rank adapters (LoRA), which keep the behaviour but lose the text: an adapter is hard to read, edit, audit or move to another model. We ask whether the skill document can be written back from the adapter and its frozen base model, and we score each recovered document by the share of...

    arxiv.org/abs/2611.09454 · PDF

  32. 32

    Not All Entropy Is Equal: Reward-Conditioned Entropy Allocation in Reinforcement Learning with Verifiable Rewards

    Song Yu, Li Li, Zhenming Wu

    cs.LG · cs.AI

    Policy entropy is widely used to diagnose exploration and collapse in reinforcement learning with verifiable rewards (RLVR), yet its magnitude does not consistently track reasoning performance across optimization objectives. We argue that this ambiguity arises because global token entropy conflates uncertainty in successful and failed trajectories, diversity across correct solution strategies, and variation among realizations of the same...

    arxiv.org/abs/2611.09448 · PDF

  33. 33

    DexDICE: Reconciling Safety and Performance via Dual-Expert Offline Imitation Learning

    Seokin Seo, Sung Kuk Shyn, Minji Seo, Yisoo Lee, Jongmin Lee, Hongseok Yang, Kee-Eung Kim

    cs.LG · cs.AI

    Ensuring high returns while providing strong safety guarantees in offline imitation learning (IL) is a fundamental challenge: existing approaches either rely on explicitly specified cost functions, which are difficult to design for complex real-world safety constraints, or require expert demonstrations that are both safe and high-performing, which are often too costly to collect. We introduce dual-expert imitation learning, which leverages...

    arxiv.org/abs/2611.09433 · PDF

  34. 34

    MINE: Entropy-coded Memory-efficient Inference for Large Language Models

    Emil Schätzle, Markus Püschel

    cs.LG · cs.DC · cs.PF

    We present MINE (Memory-efficient INference Engine), a lossless entropy coding approach that accelerates single-stream inference of Large Language Models (LLMs), where existing lossless compression methods yield speedups only for specific data types such as \texttt{BF16} and under additional assumptions. Single-stream decoding is strongly memory-bound, since every weight must be streamed from memory once per generated token and requests...

    arxiv.org/abs/2611.09424 · PDF

  35. 35

    Future-Informed Actor-Critic

    Samy Mokeddem, Gaspard Lambrechts, Adrien Bolland, Damien Ernst

    cs.LG

    Actor-critic methods optimize a parametrized policy using a critic to estimate their policy gradient. Recent work has shown that conditioning the critic on features of future states, actions, or rewards can improve sample efficiency and even final return. However, existing methods are restricted to exogenous features (independent of actions), overlooking endogenous features (dependent on actions) that may be crucial for credit assignment. We...

    arxiv.org/abs/2611.09417 · PDF

  36. 36

    Causal graph fusion across incomplete cohorts and causal steering of tabular foundation models

    Yanke Li, Dhiyaa Al Jorf, Ayoube Ighissou, Anke Scheel-Sailer, Robert Riener, Diego Paez-Granados

    cs.LG · stat.AP

    Clinical data are split across incomplete cohorts with partially overlapping variables. Causal discovery under missing data and multi-source integration have developed separately, and tabular foundation models cannot take a causal graph without retraining. We make two contributions. First, causal graph fusion learns a partial ancestral graph on the shared variables, fuses it into each cohort's graph under temporal background knowledge and...

    arxiv.org/abs/2611.09400 · PDF

  37. 37

    k-Volume: a Unified Active Learning Strategy

    Eliot Tron, Alain Celisse, Christophe Denis, Alice Le Brigant

    cs.LG · stat.ML

    In the classification context where revealing the label of a sample is costly, we propose a new Active Learning (AL) strategy named k-Volume (kVol). It randomly selects a sample to be labeled according to the volume of the smallest ball containing its k nearest neighbors. kVol is highly versatile: it can be applied either in the input space or in the output space, and it enjoys the desirable properties of the three main families of AL...

    arxiv.org/abs/2611.09388 · PDF

  38. 38

    Calibrated Uncertainty Quantification on TDC ADMET: What Can Be Resolved Across 22 Datasets?

    Laurits Fredsgaard, Michael Riis Andersen, Mikkel N. Schmidt

    cs.LG

    Absorption, distribution, metabolism, excretion, and toxicity (ADMET) models often rely on small, heterogeneous datasets, yet their predictions may determine which compounds proceed to experimental testing. Useful evaluations must therefore assess both predictive performance and the quality of reported probabilities or intervals. We ask which common uncertainty-quantification methods perform best on the 22 Therapeutics Data Commons (TDC)...

    arxiv.org/abs/2611.09373 · PDF

  39. 39

    Mitigating Covariate Shift in Behavioral Cloning via Robust Stationary Distribution Correction

    Seokin Seo, Byung-Jun Lee, Jongmin Lee, HyeongJoo Hwang, Hongseok Yang, Kee-Eung Kim

    cs.LG · cs.AI

    We consider offline imitation learning (IL), which aims to train an agent to imitate from the dataset of expert demonstrations without online interaction with the environment. Behavioral Cloning (BC) has been a simple yet effective approach to offline IL, but it is also well-known to be vulnerable to the covariate shift resulting from the mismatch between the state distributions induced by the learned policy and the expert policy. Moreover,...

    arxiv.org/abs/2611.09371 · PDF

  40. 40

    Open-world Reconstruction-based Classifier with Concept Drift Adaptation

    Joanna Komorniczak

    cs.LG

    Data stream processing has become central to modern machine learning applications, with concept drifts and novel class appearances posing the primary challenges faced by recognition methods. Adapting to concept drifts while also being able to detect novel classes and distinguish groups of unknown observations can be seen as a step towards effective method employment in an open-world setting. This work proposes a classification method that can...

    arxiv.org/abs/2611.09350 · PDF

  41. 41

    Reliable Prerequisite-Aware Learning Path Recommendation under Uncertain Knowledge Structures: Long-Horizon Evaluation and Cross-Dataset Validation

    M. M. Golam Hafiz, A. K. M. Shahriar Emtiaz, Safia Akter Simi, Md. Rakibul Islam, M. F. Mridha

    cs.LG · cs.AI

    Learning-path recommendation often depends on knowledge and prerequisite structures whose reliability is uncertain, while high accuracy in predicting the next resource does not necessarily establish reliable long-horizon path behaviour. This study develops a reliability-oriented prerequisite-aware framework that jointly examines immediate next-resource ranking, uncertain prerequisite structures, hard, soft and hybrid constraint semantics,...

    arxiv.org/abs/2611.09345 · PDF

  42. 42

    Probably Safe Sim-to-Real Transfer with Lifetime Validity

    Rohan Tangri, Jan-Peter Calliess

    cs.LG

    We present Dejavu, a method that makes sim-to-real transfer of reinforcement learning policies in constrained Markov decision processes probably safe over an unbounded sequence of policy updates. An anytime valid test using per-decision importance sampling and a betting confidence sequence certifies each candidate trained in simulation before deployment using only trajectories collected by a safe baseline policy operating in production....

    arxiv.org/abs/2611.09344 · PDF

  43. 43

    Robust Deep Prototype Classifiers via Lipschitz Training

    Adia Khalid Khan, Lukas Gehring, Benjamin Paaßen

    cs.LG · cs.AI

    Robustness and interpretability are interrelated and desirable properties of machine learning models. A model that changes its prediction under imperceptible perturbations behaves inconsistently with human-recognizable features and, hence, is not interpretable. Prototype-based classifiers such as Generalized Learning Vector Quantization address both at once. They are interpretable by construction because they make the decision by comparing...

    arxiv.org/abs/2611.09327 · PDF

  44. 44

    DiffusionRef-IMU: Guide-Anchored Residual Diffusion for Cross-Location Inertial Measurement Unit Translation

    M. M. Golam Hafiz, Safia Akter Simi, Md. Rakibul Islam, M. F. Mridha

    cs.LG

    Cross-location variation in wearable inertial measurement unit (IMU) signals complicates learning when measurements from a desired target body location are unavailable. This study proposes DiffusionRef-IMU, a two-stage cross-location translation framework in which a deterministic Direct network first estimates the target signal and a guide-anchored residual diffusion model refines only the remaining target discrepancy. The evaluation...

    arxiv.org/abs/2611.09323 · PDF

  45. 45

    Stable probabilistic emulation of partially observed dynamical systems

    Mario Massy, Sara Shamekh, Fabrizio Falasca, Matthieu Blanke

    cs.LG

    Predicting the long-term behavior of partially observed dynamical systems such as the coupled atmosphere-ocean system is central to societal decision-making. Motivated by the success of deep learning emulators in weather forecasting, data-driven climate models represent a promising approach. However, climate emulation poses distinct challenges: models must account for unresolved fast variables introduced by unavoidable coarse-graining,...

    arxiv.org/abs/2611.09317 · PDF

  46. 46

    A Stochastic Latent Variable Generalization of Joint-Embedding Predictive Architectures

    Nikos Giakoumoglou, Paschalis Giakoumoglou, Tania Stathaki

    cs.LG · cs.AI · cs.CV

    Joint-Embedding Predictive Architectures (JEPAs) predict the latent embedding of a target signal from a context signal, with an auxiliary variable z conditioning the predictor. Despite its prominent role in the original JEPA blueprint, the content of z has been carried by a fixed deterministic token in every prior instantiation, collapsing the genuinely multi-modal distribution over plausible completions to a point estimate. We present...

    arxiv.org/abs/2611.09302 · PDF

  47. 47

    Safety-Focused Reinforcement Learning Using Risk-Sensitive Coupled Control

    Jiabin Lin, Shana Moothedath, Vivek S. Borkar

    cs.LG

    This paper investigates the coupled control problem within safety-focused reinforcement learning, addressing the essential challenge of achieving optimal performance while ensuring safety in systems operating within complex environments. We define safety as the requirement to prevent the system from reaching undesirable or hazardous states. We propose a coupled control approach that prioritizes safety by employing effective methods to prevent...

    arxiv.org/abs/2611.09290 · PDF

  48. 48

    Diversity-Aware Client Selection for Communication-Efficient Federated Learning

    Max Halleryd, Sima Sinaei, Viktor Valadi, Mina Alibeigi, Henrik Abrahamsson, Salman Toor

    cs.LG · cs.AI

    Federated learning (FL) enables collaborative model training across distributed clients without sharing raw data, making it attractive for privacy-sensitive and bandwidth-constrained applications. A practical challenge in deploying FL is reducing communication overhead while maintaining model performance under heterogeneous, non-IID data distributions. Selecting which subset of clients participates in each training round offers a principled...

    arxiv.org/abs/2611.09278 · PDF

  49. 49

    FondueQA: A Benchmark for Evaluating Parametric Compensation in Multi-Hop QA

    Laurin Schmid, Suman Navaratnarajah, Koel Dutta Chowdhury, Steffen Eger, Josif Grabocka, Christian Frey

    cs.LG · cs.AI · cs.IR

    Large Language Models (LLMs) show remarkable performance based on parametric knowledge, i.e., their ability to recall facts and associations internalized in a model's weights. However, studying this capability on established multi-hop benchmarks may be complicated by benchmark contamination, as many of these benchmarks have been publicly available long enough to plausibly appear in modern pretraining corpora. As a result, correct answers...

    arxiv.org/abs/2611.09274 · PDF

  50. 50

    Rethinking MoE Load Balancing as Feedback Control

    Yikun Hou, Alp Yurtsever

    cs.LG

    Mixture-of-Experts (MoE) training is often not stationary: the data mixture changes at curriculum boundaries, when a model is continually pretrained into a new domain, or when the corpus is refreshed. Routing is learned against the current mixture, so each such change disrupts the per-expert load the router had converged to, and existing balancing methods have only been evaluated on stationary data. We rethink load balancing as a feedback...

    arxiv.org/abs/2611.09252 · PDF

  51. 51

    Topology as a Design Variable: Joint Architecture-Sizing Optimization for SRAM

    Can Wang, Junhao Ma, Yikai Wang, Shan Shen, Wei W. Xing

    cs.LG

    SRAM's power-performance-area-stability trade-off depends on bitcell topology, a discrete architectural choice that interacts non-separably with continuous transistor sizing. However, no open-source SRAM tool treats topology as an optimizable variable. OpenRAM generates SRAM layouts but performs no optimization, while OpenYield supports high-fidelity evaluation but restricts optimization to scalar sizing of a fixed 6T cell. OpenYieldV2 closes...

    arxiv.org/abs/2611.09248 · PDF

  52. 52

    Practical Non-Stationary Graph Gaussian Processes

    K. S. Sesh Kumar, Viacheslav Borovitskiy

    cs.LG

    Gaussian processes on graphs provide a principled framework for prediction and uncertainty estimation, but modelling heterogeneous networks often requires balancing flexibility against parameter complexity. Variability at a node may reflect useful signal that should propagate through the network, unreliable measurements whose influence should be suppressed, or a mixture of both. We introduce Variance-Coupled Gaussian Processes (VCGPs), which...

    arxiv.org/abs/2611.09247 · PDF

  53. 53

    Controllable Molecule Generation via Sparse Representation Editing: An Interpretability-Driven Perspective

    Zhuoran Li, Xu Sun, Chang Wen Chen, Wanyu Lin

    cs.LG

    Controllable molecule generation is crucial for diverse scientific applications, such as drug discovery and materials design. While large language models (LLMs) show great promise, their dense and entangled representations impede precise control over the generation of molecules with bespoke substructures or properties. To address this, we propose Sparse Representation Editing (SpaRE), an interpretability-driven framework for fine-grained and...

    arxiv.org/abs/2611.09235 · PDF

  54. 54

    Cross-Layer TopK: Dynamically Allocating Sparsity Across Layers in Transcoders

    Anyuan Zhuo, Shaohua Zhang, Tianle Li, Yuan Lin, Diandian Gu, Jun Wang, Pinyan Lu

    cs.LG

    Cross-Layer Transcoders (CLTs) decompose the activations of large language models into human-interpretable features and trace how these features interact across layers to form circuits, making them a key tool for mechanistic interpretability. However, training CLTs effectively and efficiently is hard. Beyond the usual trade-off between sparsity and reconstruction quality, a CLT has to model all layers at once, which adds a further challenge....

    arxiv.org/abs/2611.09233 · PDF

  55. 55

    Provable State Estimation with Recurrent Models

    Vincent Andrieu, Pauline Bernard, Lucas Brivadis, Laurent Praly, Christian Wolf

    cs.LG · math.OC

    Estimating the state of a dynamical system from observations is a crucial task in robotics and control. Despite their empirical success, recurrent models often lack formal guarantees on convergence and robustness. We bridge this gap by analyzing these architectures from the perspective of nonlinear observer theory. By framing RNNs as a time-discretization of continuous latent dynamical systems (Neural ODEs), we demonstrate that enforcing a...

    arxiv.org/abs/2611.09217 · PDF

  56. 56

    Copies Are Not Confirmations: How News Duplication Fools LLM Judges and Continual Learners

    Abdelkader Metales, Miriam Grabenheinrich, Aicha Metales

    cs.LG · cs.CL

    Many news reports of a claim are copies of one source: wire stories, content shared within a media group, syndicated features and press releases. We ask whether LLM judges and continual learners treat such copies as independent confirmations. In 200,000 English CC-News articles, a MinHash detector links 18.9% to a near-copy from another outlet (precision 0.97 once boilerplate is removed); it links 92% of edited reprints to their originals but...

    arxiv.org/abs/2611.09203 · PDF

  57. 57

    RoleWalk: training-free structural embeddings for directed graphs that transfer across graphs

    Maixent Chenebaux

    cs.LG · cs.SI

    Structural role embeddings describe a node by its position in a graph - hub, bridge, periphery, member of a dense core - rather than by who its neighbours are. Existing methods either learn coordinates specific to one graph (struc2vec, Role2Vec, RESD) or read scales and normalisations from the whole graph (GraphWave, Digraphwave), so the same neighbourhood can receive different coordinates in two graphs; few read edge directions. We present...

    arxiv.org/abs/2611.09189 · PDF

  58. 58

    A Taxonomy-Driven Synthetic Dataset for Retrieval Effectiveness and Fairness Evaluation in Hiring

    Warren Jouanneau, Morgane Hoffmann, Emma Jouffroy, Marc Palyart

    cs.LG · cs.CL · cs.CY · cs.IR

    While machine learning hiring tools are ubiquitous, the need for privacy regulations and commercial sensitivity makes public benchmarking legally complex. We address this challenge by introducing a taxonomy-driven, fully synthetic HR dataset and benchmark to simultaneously evaluate retrieval effectiveness and algorithmic fairness. Grounded in 660K real freelancer profiles, our multilingual corpus features candidate profiles, project briefs,...

    arxiv.org/abs/2611.09178 · PDF

  59. 59

    SIFR: Stable Signed Flow Reinforcement

    Runlong Liao, Xingqian Xu, Lizhang Chen, Baiyu Su, Yufan Huang, Zhonghao Wang, Akash Gokul, Ming-Yu Liu, Qiang Liu,...

    cs.LG · cs.AI · cs.CV

    Reward-based fine-tuning of flow models requires effectively promoting high-reward and suppressing low-reward samples, while maintaining training stability over many rounds of iterative updates. Iterative reward fine-tuning provides a simple starting point, but the positive reward weights give no explicit unfitting signal, and repeated reweighting concentrates the distribution and causes collapse. We propose Stable Signed Flow Reinforcement...

    arxiv.org/abs/2611.09175 · PDF

  60. 60

    TOFS: Feature Selection with Swap Supervision for Semantic Communication over Fading Channels

    Zhiyi Zhu, Eiji Takimoto, Patrick Finnerty, Chikara Ohta

    cs.LG

    Task-oriented semantic communication transmits task-relevant features to support receiver inference under limited communication resources and adverse channel conditions. For image classification, the usefulness of a feature depends on the other transmitted features and the distortion introduced by the wireless channel. Effective feature prioritization should therefore account for classification performance after transmission of the selected...

    arxiv.org/abs/2611.09173 · PDF

  61. 61

    Beyond the Grid: Continuous Dictionary Pursuit for Interpretable Signal Decomposition

    Nick Janssen, Melanie Schaller, Bodo Rosenhahn

    cs.LG

    Deep learning models achieve high accuracy in signal processing but remain difficult to interpret. Classical matching pursuit and sparse coding offer better transparency but suffer from spectral leakage and grid mismatch when signal frequencies do not align with discrete dictionary atoms. We introduce Continuous Dictionary Pursuit (CDP), a framework for interpretable signal decomposition that learns the parameters of closed-form functions...

    arxiv.org/abs/2611.09163 · PDF

  62. 62

    Augmentron: Scalable Multi-View Visual Augmentation for Robot Learning

    Tanay Tandon, Aseem Doriwala, Jade Choghari, Catherine Weaver, Pragna Mannam

    cs.LG · cs.RO

    Robot policies trained on visually narrow demonstrations can become sensitive to environmental features that are unrelated to the task. Existing generative augmentation methods can introduce visual diversity, but processing every frame and camera view independently is computationally expensive and may produce inconsistent observations across cameras and timestamps. We introduce Augmentron, a geometry-aware pipeline that generates reusable...

    arxiv.org/abs/2611.09145 · PDF

  63. 63

    Learning Generative Dynamics for 3D Molecule Generation via Sequential Neuro-Symbolic Constraints

    Zhuoran Li, Xu Sun, Chang Wen Chen, Wanyu Lin

    cs.LG · cs.AI

    Diffusion models have recently emerged as a powerful paradigm for generating 3D molecules. Despite showing remarkable promise, they learn atomic distributions implicitly from data, without explicitly encoding physicochemical rules into their generative dynamics. Consequently, generated samples can be statistically plausible yet chemically invalid. To address this problem, we propose Sequential Neuro-Symbolic Constrained Diffusion (SensDiff),...

    arxiv.org/abs/2611.09142 · PDF

  64. 64

    Deep Learning Theory through the Lens of Statistical Physics: A Topical Review

    Wei Huang, Youjin Deng

    cs.LG · cond-mat.stat-mech

    The recent success of foundation models and deep learning has far outpaced our theoretical understanding of how large-scale neural networks learn and generalize. Statistical physics provides a natural language for describing such high-dimensional, nonlinear, and strongly coupled systems, offering concepts such as mean-field theory, phase transitions, energy landscapes, and stochastic dynamics. In this topical review, we adopt a...

    arxiv.org/abs/2611.09138 · PDF

  65. 65

    COMBIO: Automated Discovery of a Compact Composite Biomarker from Routine Laboratory Panels

    Tomer D. Meirman, Lior S. Rokach, Bracha Shapira, Seffi Cohen, Noa Dagan

    cs.LG

    Machine-learning models based on routine laboratory panels can predict outcomes but can be difficult to interpret and operationalize. We present COMBIO, an automated pipeline that distils such a model into a compact closed-form biomarker. The composite is a product of laboratory measurements raised to integer powers from minus two to two. The sign of each exponent indicates whether higher values of the corresponding measurement increase or...

    arxiv.org/abs/2611.09102 · 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.