cs.LG · 2026-11-02 · No. 161

Machine Learning, 2026-11-02.

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

01 — The papers

66 entries
  1. 01

    Counting on Dummies: Privacy Amplification with Observable Counts and Truncated Batches

    Vitaly Feldman, Audra McMillan, Kunal Talwar

    cs.LG · cs.CR

    In many practical deployments of Differentially Private Stochastic Gradient Descent (DP-SGD), the server processes data in batches. Randomly sampling these batches allows for better privacy bounds via privacy amplification results. Most sampling schemes with provable privacy amplification, including Poisson sampling and random allocation, result in variable batch sizes. The privacy guarantee of these schemes requires that the batch sizes are...

    arxiv.org/abs/2610.42139 · PDF

  2. 02

    Instance-dependent sequential limits as an exact information ledger

    Akshay Balsubramani

    cs.LG · cs.IT · math.ST · stat.ML

    Instance-dependent limits for bandit regret, sequential tests, best-arm identification and Markov decision processes are proved by one chain. It runs from a divergence decomposition through data processing to a statistic, a testing step that turns a class guarantee into a requirement on the history, and optimization over alternatives and allocations. Only the first step is an identity. This work writes the chain as an exact ledger. By the...

    arxiv.org/abs/2610.42133 · PDF

  3. 03

    Physiological Trajectories and Observation Patterns in ICU Mortality Prediction: Transportability Under Changing Measurement Conditions

    Benjamin T. Pike

    cs.LG · q-bio.QM

    Objective: To distinguish the transferable predictive value of physiological trajectories from that of observation patterns in ICU mortality prediction, and to assess whether measurement timing affects performance independently of observation quantity. Materials and Methods: In-hospital mortality was predicted at 48 hours using MIMIC-III, MIMIC-IV, and eICU. Nested representations progressively integrated latest-value snapshots (R1),...

    arxiv.org/abs/2610.42126 · PDF

  4. 04

    Gradient Descent on Two ReLU Neurons: Global Landscape and Bifurcation Dynamics

    Binghua Li, Mengzhe Li, Denny Wu, Tianhao Wang

    cs.LG · stat.ML

    Understanding how gradient descent learns features remains a central challenge in neural network theory. We study this question in a minimal multi-index setting that already exhibits rich multi-phase gradient dynamics: a well-specified two-neuron ReLU teacher-student model under isotropic Gaussian inputs, trained by population gradient flow from small initialization. We show that the dynamics are organized by two directions: the "easy"...

    arxiv.org/abs/2610.42118 · PDF

  5. 05

    Fast Privacy Loss Accounting for Random Allocation via Multiscale FFT

    Vitaly Feldman, Moshe Shenfeld

    cs.LG

    In the random allocation sampling scheme each user's data is used in $k$ steps chosen randomly and uniformly from a sequence of $t$ steps. This scheme is known to improve on Poisson sampling in private optimization and in communication-efficient aggregation. The existing privacy loss distribution (PLD) accountant for this scheme runs in time quadratic in the inverse of the accuracy parameter $α$ (Feldman and Shenfeld, 2026). We give an...

    arxiv.org/abs/2610.42117 · PDF

  6. 06

    Hyperbolic Inverse Reinforcement Learning

    Iulia Rusu, Anoop Praturu, Sergei Gepshtein, Tatyana Sharpee

    cs.LG

    Inverse reinforcement learning (IRL) seeks to infer the reward function underlying observed behavior, but its predictive accuracy and interpretability depend on how that reward is computed from the underlying features. While prior IRL reward functions have been implemented through a variety of techniques including linear weights, neural networks, or other function approximators, none make assumptions about the geometry of the underlying...

    arxiv.org/abs/2610.42104 · PDF

  7. 07

    Towards Understanding the Impact of Plasticity Loss on Reinforcement Learning in Stochastic Environments

    Philipp Bordne, André Biedenkapp

    cs.LG

    Deep reinforcement learning (RL) is a powerful framework for sequential decision-making. Despite many high-profile successes in recent years, deep RL is known to be brittle. An emerging line of research attributes this brittleness to a loss of plasticity in neural networks. However, plasticity loss has so far been understudied in stochastic environments. We study this interplay systematically, training SAC agents on three DeepMind Control...

    arxiv.org/abs/2610.42100 · PDF

  8. 08

    Cost-Aware Stopping for Freeze-Thaw Bayesian Optimization: When Does One-Step Lookahead Suffice?

    Long Vu, Jongseong Brad Choi, Binoy Ravindran, Zhoulai Fu

    cs.LG · math.OC · stat.ML

    Freeze-thaw Bayesian optimization (BO) evaluates candidates in resumable increments but usually stops at a budget or convergence test. We maximize a budget-free utility, the best observed score minus a cost per increment, with a one-step lookahead: advance the candidate with the largest expected improvement (EI) under a fixed surrogate, and stop once no EI exceeds the cost. This rule, EI stopping, never falls below stopping immediately in...

    arxiv.org/abs/2610.42085 · PDF

  9. 09

    Differentially Private Synthetic Data from Sliced Chebyshev Moments

    Lucas Rosenblatt, R. Teal Witter, Christopher Musco

    cs.LG · cs.CY · cs.DS

    Research on practical differentially private (DP) synthetic data has largely focused on methods that approximately preserve low-dimensional marginals of a target distribution, and can thus only hope to preserve higher-dimensional, global structure as a lucky by-product. At the same time, existing methods that target global accuracy guarantees pay a price exponential in the data dimension, $d$. In this work, we address this ``curse of...

    arxiv.org/abs/2610.42081 · PDF

  10. 10

    A richness-sensitive Acoustic Biodiversity Index from species-level detections: classifier benchmarking, false-positive control and validation against standard soundscape indices in an African savanna and an urban zoo

    Muhammad Saeed, Dmitry Mikhaylov, Mohamed H. Alhosani, Yasser F. Al Wahedi, Rory Pilossof, Ahmed Arabi, Walid Shaaba

    cs.LG · cs.SD · q-bio.QM

    Acoustic indices are widely used as rapid proxies for biodiversity, but they summarise the distribution of sound energy rather than the species producing it and often track richness poorly. We revise and test an Acoustic Biodiversity Index (ABI) built from species-level detections. Its primary form, ABI_E, is Shannon diversity of detections normalised by the regional species pool expected at the site and season (H / ln S_ref); temporal...

    arxiv.org/abs/2610.42042 · PDF

  11. 11

    There Is No Free Loop: When Does Looping a Trained Model Pay for Its Inference Compute?

    Rhea Sanjay Sukthanker, Pashmina Cameron

    cs.LG · cs.AI

    Looping re-applies a block of layers so that each token passes through the block more than once. Adding a loop to a trained model is attractive because it leaves the parameter count, and hence the disk size, the memory footprint and the training cost unchanged. Unfortunately, every generated token pays for the loop. Re-applying m of N layers makes each token about (N+m)/N times as expensive. Loops that a model learns during large-scale...

    arxiv.org/abs/2610.42030 · PDF

  12. 12

    CausalXRL: Explainable Reinforcement Learning through Causal Graph Reasoning

    Yanming Zhang, Eric Papenhausen, Klaus Mueller

    cs.LG

    Reinforcement learning is a powerful paradigm for training autonomous agents and has achieved impressive performance in complex environments. However, this success often comes at the cost of interpretability, diminishing trust and complicating efforts to debug and improve agent behavior. To address these challenges, we introduce CausalXRL, a novel framework for explainable reinforcement learning (XRL). A key feature of CausalXRL is its use of...

    arxiv.org/abs/2610.42028 · PDF

  13. 13

    GhostPrefix: Synchronization-Free Universal Prefix Attacks for ASR Manipulation

    Shengding Liu, Ashwin Innuganti, Qiben Yan

    cs.LG

    We present GhostPrefix, the first synchronization-free universal adversarial prefix attack against automatic speech recognition (ASR) systems. Unlike prior audio attacks, which are typically crafted per utterance, require synchronization with the victim waveform, or target narrow objectives, GhostPrefix learns a short, reusable audio segment prepended to arbitrary live speech to exert real-time control over ASR outputs without knowledge of...

    arxiv.org/abs/2610.42026 · PDF

  14. 14

    Know When the Equation Is Wrong: Structural Uncertainty for LLM-Guided Symbolic Regression under Extrapolation

    Mostafa Cham, Bayu Adhi Tama, Jianwu Wang

    cs.LG · cs.AI

    Large language models (LLMs) can help discover mathematical equations from data, but most LLM-guided symbolic regression methods focus on a single best equation without quantifying uncertainty over alternative functional forms. This is challenging because distinct equations can fit the observed data similarly while producing substantially different extrapolations. Our experiments on the LLM-SRBench dataset show that standard conformal...

    arxiv.org/abs/2610.42024 · PDF

  15. 15

    When Does Jacobian Fine-Tuning Stabilize Neural PDE Emulators?

    Shreeyam Bangera, Manya Gupta

    cs.LG

    Neural emulators of chaotic PDEs are trained on one-step prediction but deployed autoregressively, where their errors compound. Jacobian-based regularisers aim to curb this growth, but in a chaotic system part of the growth is physical and cannot be trained away. We separate the two sources of this error with a first-order decomposition. We compare the emulator's error with a counterfactual that propagates the same one-step mistakes through...

    arxiv.org/abs/2610.42013 · PDF

  16. 16

    Less Loss, Fixed Weights: Logit Corrections for Forgetting

    Hanyu Yang

    cs.LG · cs.CL

    When a language model is fine-tuned to follow instructions, or further pretrained on a new subject, it often gets worse at predicting standard web text. This is known as "forgetting." We ask whether this loss can be undone by simple, fixed adjustments to the model's final token predictions. Across 14 pairs of base and derived models, from 1B to 7B parameters, we fitted a fixed bonus or penalty for each token (a vocabulary bias), alone or...

    arxiv.org/abs/2610.42000 · PDF

  17. 17

    New Lower Bounds for Private PAC Learning

    Prathamesh Dharangutte, Amartya Sanyal

    cs.LG

    We study realizable PAC learning under approximate differential privacy, and give a framework for proving lower bounds on its sample complexity. We construct classes with identical VC and Littlestone dimensions but different private sample complexities at common privacy and accuracy parameters, showing that these dimensions do not determine the sample complexity, even for improper learners. To address this we introduce the Independent Class...

    arxiv.org/abs/2610.41980 · PDF

  18. 18

    Robust Learning of Multi-Index Models under Logconcave Distributions via Weak Partial Learners

    Ilias Diakonikolas, Daniel M. Kane, Mingchen Ma

    cs.LG

    We develop a new algorithmic framework for learning Boolean concept classes under logconcave distributions. Our framework applies to broad families of multi-index models and yields significantly faster robust learners than were previously known in these settings. It extends to partial classification and testable learning. Consequently, we obtain improved learners for several fundamental concept classes, including functions of halfspaces and...

    arxiv.org/abs/2610.41961 · PDF

  19. 19

    Chronos-2: A foundation model for multivariate time series forecasting

    Abdul Fatir Ansari, Oleksandr Shchur, Jaris Küken, Andreas Auer, Boran Han, Pedro Mercado, Syama Sundar Rangapuram,...

    cs.LG · cs.AI · stat.ML

    Accurate forecasting with uncertainty quantification is critical in a variety of scientific applications ranging from epidemiology and economics to energy systems and earth sciences. Yet most forecasting models still need to be developed and trained separately for each domain-specific dataset. Time series foundation models (TSFMs) aim to remove this limitation and enable the application of the same model across domains through large-scale...

    arxiv.org/abs/2610.41956 · PDF

  20. 20

    Observable-Specific Validity Tests for Scientific AI: Controlled Approximations and Intervention Response

    Xiaoshn Nee, Haobo Zhong, Xiaomin Ni

    cs.LG · cs.AI · cs.DL

    An approximation can reproduce an energy while missing its response, or satisfy an error tolerance without being certified by a given bound. We formulate observable-specific validity records for scientific artificial intelligence (AI) that link a reduction to its reference model, error measure, tolerance, parameter domain, and transformations or interventions. Explicit displacement and velocity certificates for a singularly perturbed...

    arxiv.org/abs/2610.41939 · PDF

  21. 21

    Binary Sparse Codes: Where are we on Energy Efficient Representations?

    Jose Miguel Ramos, Luis SaCouto, Andreas Wichert

    cs.LG

    Biological neural systems achieve impressive computational capabilities under tight metabolic constraints, relying on sparse cortical representations that minimise total neural activity. This combination of efficiency and representational power has motivated growing interest in sparse coding as a computational strategy for artificial systems. Artificial models designed to produce binary sparse codes span three broad paradigms, theoretical...

    arxiv.org/abs/2610.41933 · PDF

  22. 22

    On k-nearest-neighbor regression in metric spaces

    László Györfi, Vladimir G. Pestov, Harro Walk, Roi Weiss

    cs.LG

    We study mean squared consistency of k-nearest-neighbor regression for metric space valued features. For a fixed feature distribution mu with atomless distance distributions, we describe how neighbor selection redistributes probability through the density of a uniformly chosen neighbor of an independent query. This gives separate conditions for the approximation error and the conditional response variance, and weakens Stone's domination...

    arxiv.org/abs/2610.41896 · PDF

  23. 23

    Trainable Vector Quantization for Large Language Models

    Anastasiia Prutianova, Alina Bogdanova, Mikhail Shutov, Dmitrii Poimanov, Anatoly Gorkov, Ilia Zharikov

    cs.LG

    Vector Quantization (VQ) is a rapidly evolving technique for compressing Large Language Model (LLM) weights, reducing memory usage, and speeding up inference. However, extreme compression deteriorates model performance, which is challenging to recover due to the discrete nature of VQ parameters. Existing methods employ training-free approaches that are limited to layer-wise reconstruction loss minimization. To address this limitation, we...

    arxiv.org/abs/2610.41822 · PDF

  24. 24

    ElTab: Elastically Looped Tabular Transformers

    Mieszko Komisarczyk, Saurabh Mathur, Maurice Kraus, Sriraam Natarajan, Kristian Kersting

    cs.LG

    Tabular transformer models such as TabPFN achieve strong predictive performance but incur substantial inference costs, making them difficult to deploy in resource-constrained settings. While looped transformers reduce parameter count through weight sharing, they can lose significant predictive power when stopped early. To address this challenge, we introduce the Elastically Looped Tabular Transformer (ElTab), a single model trained to operate...

    arxiv.org/abs/2610.41819 · PDF

  25. 25

    DIFFNAS: Predictor-Free Neural Architecture Search via Diffusion Model and Zero-Cost Evaluation

    Yize Sun, Hanzhang Ma, Zixin Wu, Xuechun Li, Volker Tresp, Yunpu Ma

    cs.LG

    Neural Architecture Search (NAS) has advanced the automated design of deep networks, but remains computationally expensive due to reliance on model training or performance predictors. We introduce DIFFNAS, a predictor-free NAS framework that leverages a discrete graph diffusion model guided by zero-cost performance proxies. Unlike prior diffusion-based NAS methods that require learned performance predictors, DIFFNAS conditions the generative...

    arxiv.org/abs/2610.41818 · PDF

  26. 26

    Continual Reinforcement Learning Is an Integration Problem

    Nis Meinert, Jakob Gawlikowski, Edward Grant, Max Cyprys, Nathan Schurr, Neil Boston, Alfredo Ibias, Milind Tambe,...

    cs.LG

    This paper presents an opinionated set of ten desiderata intended to inform the design and evaluation of reinforcement-learning agents that learn continually throughout a continuing lifetime. We use these desiderata to examine a representative selection of existing approaches. Progress on individual desiderata is valuable, but the longer-term aim should be to integrate all ten within a single agent. To our knowledge, no published evaluation...

    arxiv.org/abs/2610.41809 · PDF

  27. 27

    Exploring Bad Minima in Transfer Learning

    Mina Remeli

    cs.LG · cs.CR

    Transfer learning is a highly successful learning paradigm in deep learning which leverages knowledge from one task and applies it to another. Despite its success, the necessary conditions for knowledge transfer are unclear and require further research. In this dissertation, our aim is to understand transferability a little better by answering the following question: Does good generalization on a source task necessarily imply transferability...

    arxiv.org/abs/2610.41807 · PDF

  28. 28

    Backpropagation-free Deep Learning for Time-Series Classification

    Miha Kapš, Blaž Bertalanič, Carolina Fortuna

    cs.LG · cs.AI

    Deep neural networks for time series classification (TSC) are trained almost exclusively by error backpropagation, which imposes update locking and the weight-transport problem. Backpropagation-free learning rules such as Direct Feedback Alignment (DFA), Direct Random Target Projection (DRTP), and the diffusion-inspired NoProp relax these constraints but have been evaluated almost exclusively on 2D image benchmarks. We adapt these three...

    arxiv.org/abs/2610.41805 · PDF

  29. 29

    Towards Out-of-Distribution Uncertainty Quantification for Machine-Learned Interatomic Potentials via Pre-Training

    Stefan P. Schmid, Tobias Setzer, Maximilian Scheurer

    cs.LG · physics.chem-ph

    Informative uncertainty quantification (UQ) is essential for the practical deployment of machine-learned interatomic potentials (MLIPs), as it provides insight into their reliability across chemical space. Although numerous post-hoc uncertainty estimation methods for MLIPs have recently been proposed, most have been trained in low-data regimes, restricting their applicability to narrow domains of interest. Inspired by the emergence of...

    arxiv.org/abs/2610.41801 · PDF

  30. 30

    On-device Detection of Wearer-induced NLOS on Chest-worn UWB Tags: Machine Learning, Link Indicators, or Both?

    Markus Gallacher, Vu Anh Minh Le, Matteo Trobinger, Davide Vecchia, Carlo Alberto Boano, Gian Pietro Picco

    cs.LG

    Ultra-wideband (UWB) enables high-accuracy distance estimation (ranging), but performance degrades in non-line-of-sight (NLOS) conditions. When UWB tags are worn, the human body inevitably causes a wearer-induced NLOS (WNLOS), yielding significant errors; still, despite its relevance, WNLOS remains understudied. Moreover, WNLOS detection typically must occur on-device at the tag, requiring classification accuracy to be balanced against tight...

    arxiv.org/abs/2610.41799 · PDF

  31. 31

    Accurate and Auditable ECG Classification via Visibility Graph Neural Networks

    Artur Xavier, Guilherme Henrique Gomes Evangelista, Aitazaz Khan, Antonio Luiz Pinho Ribeiro, Mohammed J. Zaki,...

    cs.LG · cs.AI

    Physicians read electrocardiograms (ECGs) from waveform shape and timing, yet accurate deep-learning classifiers rarely show which waveform regions drive their predictions. We introduce LYNX-ECG, a graph neural network that represents each heartbeat as a sparse visibility graph of timestamped turning points. On CODE-Test, LYNX-ECG reaches a macro F1 score of 0.889, outperforming five baselines including the ECGFounder foundation model, while...

    arxiv.org/abs/2610.41795 · PDF

  32. 32

    Implicit Assumptions About Tasks in Continual Learning

    Kathrin Korte, Christian Metzner, Akshay Bedhotiya, Viet Anh Khoa Tran, Eleni Nisioti, Sebastian Risi

    cs.LG · cs.AI

    Continual learning is the study of how a model accumulates knowledge from data received over time, rather than all at once. In practice, this is typically achieved by dividing a dataset into tasks and presenting them sequentially. Explanations of the resulting measurements take into account factors such as the similarity between segments, the rate at which the data changes and the order of presentation. Each of these factors can be defined in...

    arxiv.org/abs/2610.41792 · PDF

  33. 33

    Latent-Lens: Visual Perception in Small Language Models Through Latent Communication

    Mo Shahdloo, Faegheh Sardari, Mathew Salvaris

    cs.LG · cs.AI

    Bridging independent vision-language models (VLMs) and small language models (SLMs) in a modular multi-agent system typically forces them to exchange information as discrete text, which creates a lossy bottleneck for spatial and fine-grained details. This constraint is especially acute for SLMs, whose limited capacity makes them more reliant on the quality of the input signal. Recent latent-communication methods exchange states between...

    arxiv.org/abs/2610.41751 · PDF

  34. 34

    Do LLMs Know When They Are Wrong About Molecules? In-Context Property Prediction with Self-Reported Uncertainty

    Matthias Busch, Marius Tacke, Tamadur Albaraghtheh, Sviatlana V. Lamaka, Mikhail L. Zheludkevich, Roland C. Aydin,...

    cs.LG

    Large language models (LLMs) are used in many domains, but molecular property prediction with them is still little explored. In this work we evaluate the capability of 7 frontier LLMs to predict molecular properties together with a self-reported uncertainty from in-context examples, and compare them with classical baselines trained on the same molecules. We use 4 datasets and screen each for contamination of the training data of the LLMs:...

    arxiv.org/abs/2610.41745 · PDF

  35. 35

    Task Vectors Can Be Linearly Decomposed into Interpretable and Controllable Components

    Yukiya Horiba, Hakaze Cho, Takahiro Yakoh

    cs.LG · cs.CL

    Task Vectors (TVs) provide a compact way to steer large language models, but they are typically treated as atomized directions merely extracted from hidden states, leaving their internal functional structure poorly understood. Motivated by prior work that decomposes in-context learning into Task Recognition (TR) and Task Learning (TL), we ask whether a single TV can likewise be decomposed into functionally meaningful components. We formulate...

    arxiv.org/abs/2610.41743 · PDF

  36. 36

    Locality-Controlled Out-of-Distribution Guidance for Robust Regulatory DNA Sequence Design

    Dabin Jeong, Marwane Bourdim, Mo Lotfollahi

    cs.LG · q-bio.GN

    Designing regulatory DNA sequences for targeted gene-expression control is an important challenge. Recent approaches frame this task as property-conditioned generation, where diffusion models trained on natural sequences are guided toward high predicted activity through reward-based fine-tuning or sampling-time guidance. We show that single-reward evaluation can obscure reward-hacking behavior. By decomposing evaluation into property,...

    arxiv.org/abs/2610.41736 · PDF

  37. 37

    Leakage-Controlled Structure-Preserving Augmentation and Template-Aware Evaluation for Bangla Symptom-Based Disease Prediction

    Sabah Nushra, Shovan Samanta Turzo, Moinul Hossain, Md Kishor Morol

    cs.LG · cs.CL

    Bangla symptom-based disease prediction is commonly evaluated on a small, imbalanced structured benchmark containing many profiles that differ by only one or two symptoms. We test whether train-only asymmetric Bernoulli perturbation improves classification using 725 unique profiles spanning 85 diseases and 172 binary symptoms. Experiments use ten repeated real-profile holdouts and parent-grouped inner cross-validation. The 2 x 2 design...

    arxiv.org/abs/2610.41733 · PDF

  38. 38

    UpLoop: Training Looped Transformers via Upcycling at No Extra Pretraining Compute

    Cheng Liang, Tengfei Zhang, Ya Zhang, Yanfeng Wang, Zhouhan Lin

    cs.LG · cs.CL

    Looped language models can strengthen reasoning through repeated computation with shared weights, increasing execution depth without adding independent parameters. However, each additional pass increases training floating-point operations (FLOPs), making these gains more expensive to obtain. To address this cost, we propose UpLoop, a pretraining schedule that uses upcycling to train looped Transformers while exactly matching a standard dense...

    arxiv.org/abs/2610.41720 · PDF

  39. 39

    Hierarchical Residuals: A Minimalist Decoupling of Residual Width from Compute Width

    Ziteng Wang, Jianqiao Lu, Xinyu Ma, Jintao Zhang, Jun Zhu, Yiyuan Ma, Jianfei Chen

    cs.LG

    Residual connections underpin Transformers, but they implicitly tie the residual width carried through depth to the compute width used inside each layer. We introduce Hierarchical Residuals (HiRes), a minimalist architectural change that decouples these dimensions, widening the residual state while keeping per-layer compute width fixed. HiRes groups layers into blocks: standard d-dimensional PreNorm layers operate inside each block, while a...

    arxiv.org/abs/2610.41711 · PDF

  40. 40

    From fleet trajectories to single-cell promises: hierarchically weighted conformal prediction for battery SOH monitoring

    Xiangyu Liu

    cs.LG · stat.AP

    Interval predictions from conformal prediction are usually read as "90% coverage", but on nested battery data that number promises the fleet average, not the individual cell. We measure the gap on 124 public LFP cells under a preregistered three-tier reporting protocol (fleet, batch, cell): pooled coverage looks healthy (0.875-0.906 across splits) while the worst cell is covered 0.1-2% of the time, with exact binomial tests against the...

    arxiv.org/abs/2610.41696 · PDF

  41. 41

    From WSIs to Bulk, Spatial and Single-Cell Transcriptomics via Multi-Scale Graph Neural Networks

    Miguel Andrade, José Almeida, Carlos Santiago, Catarina Barata

    cs.LG

    Understanding the spatial distribution of gene expression is key to studying complex diseases like cancer. While Spatial Transcriptomics (ST) and Single-Cell Transcriptomics (scT) provide high-resolution insights, they remain expensive and scarce. In contrast, standard histopathology Whole-Slide Images (WSIs) paired with bulk RNA sequencing are widely available. This work introduces a Multi-Scale Graph Neural Network (GNN) capable of...

    arxiv.org/abs/2610.41692 · PDF

  42. 42

    Does Reasoning Make Language Models Agree?

    Mila Gorecki, Moritz Hardt

    cs.LG · cs.CL · cs.CY

    Language models face two opposing concerns: monoculture, in which models err in correlated ways, and multiplicity, in which equally accurate models make conflicting decisions. The rise of reasoning models has complicated this debate: On the one hand, reasoning may increase model agreement if models anchor on the same salient features or follow similar reasoning paths. On the other hand, stochastic reasoning traces may introduce additional...

    arxiv.org/abs/2610.41691 · PDF

  43. 43

    On the Generation Order in GFlowNets

    Timofei Gritsaev, Evgeny Katasonov, Ildus Sadrtdinov, Esmeralda S. Whitammer, Dmitry Vetrov

    cs.LG

    Sampling from discrete distributions specified by unnormalised probability functions is a classical statistical problem. GFlowNets address this problem by constructing objects sequentially, adding one discrete component at a time in either a predefined or a learned order. The choice of order determines which conditional distributions the neural network must learn, some of which may be easier to learn than others. Yet whether existing GFlowNet...

    arxiv.org/abs/2610.41648 · PDF

  44. 44

    A Constructive Orbit-Based Theory of Learnability

    Yunied Puig

    cs.LG

    We develop a finite-dimensional orbit framework for constructing data-adapted dictionaries, sparse realizations, and predictive extensions. The method uses wrapped weighted shifts to generate orbit dictionaries and local interpolation maps from independent orbit representatives, turning the learning problem into three explicit tasks: scaffold discovery, sparse realization on the discovered dictionary, and prediction through learned sparse...

    arxiv.org/abs/2610.41641 · PDF

  45. 45

    ORION: An Outcome-Driven LLM Advisor for RL-Based Multi-Domain 6G Slice Placement

    Mahmoud S. Mayaleh, Fetia Bannour, Nour-El-Houda Yellas

    cs.LG · cs.AI · cs.MA · cs.NI

    Deploying 6G network slices across administrative domains requires an online decision for each arriving request: which domain hosts each virtual network function, how traffic crosses domain boundaries, and whether the slice is admitted. This joint decision problem is NP-hard under continuous arrivals. Experiments on a parametrized family of multi-domain topologies show that no fixed planning strategy dominates across all regimes: the...

    arxiv.org/abs/2610.41563 · PDF

  46. 46

    FOCUS: First-Order Control Using Optimal Transport Supervision for Imitation from Observations

    João A. Cândido Ramos, Lionel Blondé, Hugues Vinzant, Jacopo Castellini, Alexandros Kalousis

    cs.LG · cs.RO

    Imitation learning from observations must recover behavior from state-only demonstrations without known temporal correspondence. Differentiable simulators provide pathwise policy gradients, but differentiating through long, contact-rich trajectories can make optimization unstable. We introduce FOCUS, which combines local optimal-transport alignment with short-horizon differentiable actor-critic learning. FOCUS matches agent rollout windows to...

    arxiv.org/abs/2610.41540 · PDF

  47. 47

    Enhancing Sentiment Classification on Small Datasets through Data Augmentation and Transfer Learning

    Mahmoud S. Mayaleh, Samer A. Mayaleh

    cs.LG · cs.CL

    Small-scale sentiment classification often suffers from data scarcity, which limits the generalization ability of the models. This study evaluates and compares the effectiveness of three data augmentation strategies: Easy Data Augmentation (EDA), back-translation, and contextual token substitution (nlpaug-style), with both traditional machine learning classifiers (Logistic Regression, Random Forest) and transformer-based models (BERT). We...

    arxiv.org/abs/2610.41518 · PDF

  48. 48

    Infinite ResNets are Gaussian Mean ODEs

    Louis-Pierre Chaintron, Lénaïc Chizat, Javier Maass

    cs.LG

    Understanding the training dynamics of deep neural networks in large-scale limits is a question both of foundational and of practical importance. Prior works have obtained such descriptions through dynamical mean field theory (DMFT), yielding a system of non-Markovian dynamics that are challenging to analyze and that obscure the optimization structure of the original dynamics. In this paper, we obtain instead a simple description of the...

    arxiv.org/abs/2610.41514 · PDF

  49. 49

    PCD: Predictive Coding Dynamics for Irregularly Sampled Time Series

    Tianrui Gou, Jiayangmei Xu, Xingzhou Zhang, Lei Sha

    cs.LG

    Dynamic systems evolve continuously over time, whereas observations in real scenarios are often sparse, irregular, and asynchronous. Accurate prediction therefore requires inferring latent state evolution between observations, as future values depend on partially observed trajectories. We propose Predictive Coding Dynamics (PCD), a residual feedback mechanism for irregular continuous dynamic models. PCD maintains previous reconstructions,...

    arxiv.org/abs/2610.41506 · PDF

  50. 50

    Spectral Haar Binarizer: Compact Boolean Contrast Representations for Hyperspectral Classification

    Michał Romaszewski, Przemysław Głomb

    cs.LG

    Tsetlin Machines offer an interpretable, hardware-friendly alternative to conventional machine learning models, but applying them to continuous hyperspectral measurements requires converting spectral signals into Boolean representations. Standard binarisation methods typically discretise each spectral band independently, potentially discarding important relationships between neighbouring wavelengths. In this paper, we introduce the Spectral...

    arxiv.org/abs/2610.41478 · PDF

  51. 51

    When Links Mislead: Label-Free Structural Polarity for Zero-Shot Classification of Web Pages

    Hanshuai Cui, Zhiqing Tang, Zhi Yao, Weijia Jia, Jun Luo

    cs.LG

    Web platforms such as digital libraries, online stores and university sites must classify new pages before anyone labels them. Large language models (LLMs) can label each page from its text, but these labels are noisy, and links could correct them only where linked pages share classes. Existing zero-shot methods ignore the links, or pull a page toward its neighbours as label propagation and graph neural networks (GNNs) do, which hurts where...

    arxiv.org/abs/2610.41464 · PDF

  52. 52

    RetryRoute: Retry-Aware Model Routing for Agentic Image Editing Services

    Hanshuai Cui, Zhiqing Tang, Zhi Yao, Jun Luo, Weijia Jia

    cs.LG

    Web agents increasingly edit images by calling open editing models in a critique-and-retry loop, and these models differ in cost by more than an order of magnitude. After every rejected edit, a service must decide whether to retry the same model, escalate to another or give up. Existing routers pick one model per request, and cascades escalate in a fixed order. However, both treat a retry as an independent draw. We identify sticky failures,...

    arxiv.org/abs/2610.41460 · PDF

  53. 53

    NormDetect: Anchored Calibration for Low-False-Positive Detection of Machine-Generated Web Text

    Hanshuai Cui, Fanshuai Meng, Zhiqing Tang

    cs.LG

    Web platforms screen posts for text written by large language models, and every false flag hits a human writer. Detectors must therefore run at a very low false positive rate (FPR). Zero-shot detectors such as Binoculars score text with frozen language models, but invisible edits such as homoglyphs shatter tokenization and push machine text into the human range. Unicode normalization undoes these edits, yet the normalized score still misses...

    arxiv.org/abs/2610.41459 · PDF

  54. 54

    Improving the Optimization Landscape of Matrix Completion with $ε$-close Surrogates

    Ziye Ma

    cs.LG · math.OC

    Non-convex optimization is central to machine learning and AI, yet unlike its convex counterpart, it still lacks a general and principled framework. In this work, we try to highlight the importance of combining noisy perturbations with over-parametrization mathematically under the classic setting of matrix completion (MC). Matrix completion is an important non-convex recovery problem in which only a subset of a matrix's entries is observed,...

    arxiv.org/abs/2610.41455 · PDF

  55. 55

    When Everyone Can Build a Model: Mechanistic Inference in the Age of AI

    Marvin van Aalst, Adèle H. Ribeiro, Anna Matuszyńska

    cs.LG

    Modern deep learning and AI-assisted code generation are lowering the technical barrier to building mathematical models of biological systems that reproduce complex observations. This lets researchers with domain expertise build semi-mechanistic, hybrid models constrained by prior knowledge and augmented with learned structure without extensive training in modelling. However, easier construction does not make interpretation easier: predictive...

    arxiv.org/abs/2610.41454 · PDF

  56. 56

    From Representation to Readout: Mechanistic Analysis of Membership Signals in Language Models

    Wenxi Dai, Jingyu Hu, Yanting Yang, Mengnan Du

    cs.LG · cs.AI

    Membership inference aims to determine whether a particular example was included in a model's training data. Such membership signals are often broadly described as memorization, while the underlying mechanisms can involve both how information is represented internally and how it is accessed by the model to produce behaviors. We therefore present a mechanistic study of membership signals in large language models (LLMs). We introduce Decodable...

    arxiv.org/abs/2610.41445 · PDF

  57. 57

    A Unified XAI-NSAI Analytical Framework for Trustworthy IoT Anomaly Detection

    Hugues Leger, Giovani R. Abuaitah

    cs.LG

    The widespread deployment of Internet of Things (IoT) devices in critical environments makes anomaly detection essential for identifying abnormal behaviors and supporting timely investigation and response. However, accurate detection alone is insufficient. Practical IoT deployments must also provide explanations, incorporate domain knowledge and reasoning, operate under IoT constraints, and support human decision making. Although recently...

    arxiv.org/abs/2610.41424 · PDF

  58. 58

    SparseGDN: Chunk Sparse in Linear Attention

    Jiedong Zhuang, Peng Zhao, Jialin Tang, Haoji Hu

    cs.LG · cs.AI · cs.CL

    Gated Delta Networks (GDNs) offer linear-time sequence modeling, yet long-context prefill remains costly due to dense intra-chunk computation, recurrent-state updates, and intermediate state storage. We introduce SparseGDN, a hardware-aware framework that combines structured intra-chunk approximation with input-dependent recurrent-event selection while producing outputs for every token. A block-adjacent solver reduces local Delta-solve...

    arxiv.org/abs/2610.41420 · PDF

  59. 59

    Certifying Retraining Regret through Residual Structure

    Peida Li

    cs.LG

    We certify a fixed ridge predictor's regret against full-feature retraining over every continuous completion of an interval-valued table. Our main theorem gives verifiable original-row residual conditions for a uniform error bound independent of nuisance-feature scale and rank, with an asymptotically sharp leading constant and quantitative leakage stability. Full-rank constructions separate certification dimension from feature dimension,...

    arxiv.org/abs/2610.41410 · PDF

  60. 60

    Anomalies as Residuals: Cauchy Displacement Networks

    Yang Cao, Zihao Guo

    cs.LG

    Neighbourhood-based detectors such as KNN, LOF and LUNAR remain among the most accurate methods for tabular anomaly detection, but they compare every query with the stored training set, so their inference cost and memory grow with the data and the data must be kept at deployment. We ask whether the neighbourhood geometry these detectors exploit can be compiled into a small network that scores a query with a single forward pass. We propose the...

    arxiv.org/abs/2610.41387 · PDF

  61. 61

    What We Keep of What We Forget: Replacement Operators and Routing Semantics in Training-Free MoE Expert Pruning

    Ezza Ali, Sehrish Bibi, Momina Moetesum, Faisal Shafait

    cs.LG · cs.CL

    Training-free expert pruning makes two decisions: which experts to remove, and what the layer computes when its router still selects a removed expert. Published methods change both at once. Holding the mask fixed, we compare replacement operators from three families on OLMoE, DeepSeek-V2-Lite and Moonlight, with Granite as a fourth router, using paired, multiplicity-corrected tests. The largest effect is a choice usually left implicit: the...

    arxiv.org/abs/2610.41366 · PDF

  62. 62

    Generated storm lifecycles fill the rare-event tail of short rainfall records

    Valerie Tsao, Nathaniel W. Chaney, Manolis Veveakis, Marta Zaniolo

    cs.LG

    Historically rare ''one-in-a-generation'' storms are occurring with unprecedented frequency, causing catastrophic damage and mounting casualties, and exposing the vulnerability of infrastructure designed around short observational records. To overcome the limitations of sparse data, we present StormGen, a generative machine-learning model that transforms an observed storm's initial hours into complete, physically plausible hourly rainfall...

    arxiv.org/abs/2610.41365 · PDF

  63. 63

    Quantile Propagation Planning for Budgeted Tail-Risk Evaluation

    Weijia Han, Zhi Huang, Lisha Qu, Yuanyuan Tan, Jiayue Meng

    cs.LG · eess.SY

    Risk-aware planning must identify rare, costly outcomes using only a limited number of model evaluations. Monte Carlo rollouts spend this budget on noisy tail estimates, while planning from the mean can miss the tail entirely. We introduce quantile propagation planning (QPP), which propagates a small, fixed set of disturbance quantiles through the dynamics and integrates the resulting cost curve to evaluate both mean and tail risk. The method...

    arxiv.org/abs/2610.41351 · PDF

  64. 64

    StrictPRMBench: A Replayable Evaluation Benchmark for Auditing Fixed-Score PRM Aggregation

    Pengfei Zheng

    cs.LG

    Process reward models (PRMs) score reasoning steps, but best-of-N selection also requires a rule that aggregates those scores. We study that rule as a separate evaluation object. StrictPRMBench fixes mathematical reasoning pools, scorer evidence, and the source trace of a deletion intervention, so that different reducers face the same recorded selection problem. Across seven scoring checkpoints and two benchmarks, the stored N=32 outcomes...

    arxiv.org/abs/2610.41333 · PDF

  65. 65

    Transferring Peptide-Target Interaction Patterns to Small Molecules by Joint Guidance for Diffusion and Flow Matching

    Jian Lee, Changyun Cho, Sun Kim

    cs.LG · cs.AI

    Generating small molecules to target protein-protein interactions (PPIs) remains one of the most challenging yet high-reward frontiers in drug discovery. Unlike traditional small-molecule drugs that bind to deep, well-defined pockets, PPI interfaces often comprise broad, shallow binding surfaces. A natural strategy is to leverage known peptide-target interactions, as peptide ligands provide rich interface-level constraints. However, a major...

    arxiv.org/abs/2610.41322 · PDF

  66. 66

    rl_rob: A transparent, from-scratch Python library for teaching reinforcement learning from tabular methods to deep policy gradients

    Abdulrahman Altahhan

    cs.LG

    rl_rob is an open-source Python library for teaching reinforcement learning. It implements the arc of a modern course, from tabular temporal-difference learning through linear function approximation to deep value-based and policy-gradient methods, with every algorithm written from first principles so that its update rule is visible in a few lines of code. All agents share one interface, so learners can compare methods directly across stages...

    arxiv.org/abs/2610.41303 · 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.