cs.LG · 2026-07-07 · No. 46

Machine Learning, 2026-07-07.

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

01 — The papers

73 entries
  1. 01

    Weak-to-Strong Generalization via Direct On-Policy Distillation

    Shiyuan Feng, Huan-ang Gao, Haohan Chi, Hanlin Wu, Zhilong Zhang, Zheng Jiang, Bingxiang He, Wei-Ying Ma, Ya-Qin...

    cs.LG · cs.AI · cs.CL

    Reinforcement learning with verifiable rewards (RLVR) is a powerful recipe for improving language-model reasoning, but it is expensive to repeat on every new strong model because the target model must generate many rollouts during training. As models scale, post-training itself becomes a bottleneck. We study a weak-to-strong alternative: run RL on a smaller model where rollouts are cheaper, then reuse what that RL run learned to improve a...

    arxiv.org/abs/2607.05394 · PDF

  2. 02

    What Does a Discrete Diffusion Model Learn?

    Rodrigo Casado Noguerales, Bernhard Schölkopf, Thomas Hofmann, Aran Raoufi

    cs.LG · cs.AI · cs.CL · cs.IT · stat.ML

    What does a discrete diffusion model learn: a denoiser, a score ratio, or a bridge plug-in predictor? At the level of jump rates, these are one object in different coordinates, and reading a neural network in the wrong coordinate changes the process being trained and sampled. Starting with a rigorous derivation of the continuous-time Markov chain (CTMC) ELBO for any noising process, boundary terms included, we prove the \emph{Oracle Distance}...

    arxiv.org/abs/2607.05381 · PDF

  3. 03

    TabPack: Efficient Hyperparameter Ensembles for Tabular Deep Learning

    Yury Gorishniy, Akim Kotelnikov, Ivan Rubachev, Artem Babenko

    cs.LG

    In deep learning for tabular data, efficient ensembles of multilayer perceptrons (MLPs) have recently emerged as effective and practical architectures. Existing methods of this kind use the same hyperparameters for all underlying MLPs, which requires hyperparameter tuning for achieving the best performance. In this work, we introduce TabPack, an efficient MLP ensemble with strong out-of-the-box performance and reduced reliance on traditional...

    arxiv.org/abs/2607.05380 · PDF

  4. 04

    CompactionRL: Reinforcement Learning with Context Compaction for Long-Horizon Agents

    Yujiang Li, Zhenyu Hou, Yi Jing, Jie Tang, Yuxiao Dong

    cs.LG

    Long-horizon agentic LLMs are increasingly limited by finite context windows, as extended interaction trajectories can exceed the maximum context length before a task is completed. Context compaction offers a natural solution by summarizing previous interaction states and continuing the rollout under a compressed context, but incorporating compaction into reinforcement learning remains underexplored. We propose CompactionRL, a reinforcement...

    arxiv.org/abs/2607.05378 · PDF

  5. 05

    TREK: Distill to Explore, Reinforce to Refine

    Yuanda Xu, Zhengze Zhou, Kayhan Behdin, Jelena Markovic-Voronov, Hejian Sang, Xiaomin Li, Wenhui Zhu, Xinchen Du,...

    cs.LG · cs.AI · stat.ML

    Group Relative Policy Optimization (GRPO) is effective when the current policy already samples useful reasoning trajectories, but it stalls on hard prompts whose correct solution modes lie outside the student's on-policy support. We propose TREK (Teacher-Routed Exploration via Forward KL), a simple staged procedure that uses distillation not for imitation but for exploration support expansion. A key advantage of TREK is its generality:...

    arxiv.org/abs/2607.05339 · PDF

  6. 06

    How Far is Too Far? Defining the Distance Threshold for Verification Siamese Networks

    Heloísa Dias Viotto, Cauê Samonek, Lucas Garcia Pedroso, Marcos Sunye, André Abed Grégio, Paulo Lisboa de Almeida

    cs.LG

    Siamese verification networks are widely used to compare items such as faces, cars, or signatures. In these scenarios, the network is trained to learn an embedding space in which similar objects are mapped closer together, while dissimilar objects are mapped further apart. Two objects are considered to belong to the same class (e.g., the same person in two different images) when the distance between their embeddings falls below a predefined...

    arxiv.org/abs/2607.05329 · PDF

  7. 07

    Biologically Informed Deep Neural Networks for Multi-Omic Integration, Pathway Activity Inference and Risk Stratification in Cancer

    Pedro Henrique da Costa Avelar, Le Ou-Yang, Min Wu, Sophia Tsoka

    cs.LG

    Integrating complex, multi-omics data presents significant challenges. Existing approaches often face a trade-off between model interpretability and representational capacity, with most either relying on post-hoc interpretation or use linear models that may overlook complex interactions. We report Pathway Activity Autoencoders for the multi-omics setting, which embed prior knowledge via pathway-informed architectural constraints, fostering...

    arxiv.org/abs/2607.05306 · PDF

  8. 08

    Learning Only What Valid Adapters Can Express: Subspace-Constrained Adaptation Against Fine-Tuning Poisoning

    Fabien Polly

    cs.LG · cs.CR

    Parameter-efficient fine-tuning still leaves a broad space of behavior-changing updates reachable, so a poisoned objective can be represented and optimized. We study an alternative: adaptation constrained to the subspace estimated from a trusted pool of existing task adapters. On flan-t5-large with 196 public LoRA adapters, we show that (1) the functionally relevant content of an adapter lies in a low-dimensional shared subspace, 30 to 38...

    arxiv.org/abs/2607.05300 · PDF

  9. 09

    Air Quality Downscaling with Station-Guided Pseudo-Supervision

    Guorun Wang, Simone Foti, Andreas D. Demou, Leonidas Kotoulas, Theodoros Christoudias, Alexandros Koliousis, Mihalis...

    cs.LG · cs.AI · cs.CV

    Super-resolving coarse atmospheric fields to local PM$_{2.5}$ variations is uniquely challenged by a mismatch in spatial support: while pixels represent regional averages, ground-truth observations are discrete, unaligned samples of a continuous spatial signal. To bridge this gap, we present a station-guided framework for high-resolution PM$_{2.5}$ downscaling over Europe. Taking coarse CAMS atmospheric composition fields alongside...

    arxiv.org/abs/2607.05292 · PDF

  10. 10

    Advances in Neural Controlled Differential Equations

    Benjamin Walker

    cs.LG

    Many real-world systems evolve continuously, yet most machine learning models interpret time series as discrete sequences. Continuous-time approaches instead treat time series as samples from an underlying input path, a formulation that naturally accommodates irregularly sampled or oversampled data. Among these, Neural Controlled Differential Equations (NCDEs) are a maximally expressive class of models that parametrise a vector field using a...

    arxiv.org/abs/2607.05280 · PDF

  11. 11

    Adaptive Inference Batching using Policy Gradients

    Ruslan Sharifullin

    cs.LG · cs.AI · cs.DC · cs.PF

    Inference serving systems must balance throughput and latency under bursty, heterogeneous workloads, yet the industry standard remains static batching policies that require manual tuning and cannot adapt to shifting traffic. We investigate whether reinforcement learning (RL) can learn adaptive batching and routing policies that outperform these heuristics, training REINFORCE and PPO agents on a discrete-event simulator validated against...

    arxiv.org/abs/2607.05272 · PDF

  12. 12

    Target-Guided Selective Reweighting for Physics-Informed Neural Network Inverse Problems: A Transfer Learning Approach

    Qian Hu, Bin Fan, Yao Xiao, Zhicheng Lin, Meixin Xiong

    cs.LG

    Physics-informed neural networks (PINNs) encounter ill-posed optimization, loss competition, and parameter compensation in partial differential equation (PDE) inverse problems. Transfer learning can reuse representations from source tasks, but direct fine-tuning may introduce negative transfer when dominant physical mechanisms, governing parameters, or observation noise differ between source and target domains: the model achieves low field...

    arxiv.org/abs/2607.05271 · PDF

  13. 13

    GeoFlow: Geo-Aware Modeling of Inter-Area Relationships in Origin-Destination Flow Prediction and Generation

    Zherui Huang, Guanjie Zheng, Hao Xue, Linghe Kong

    cs.LG

    Origin-destination (OD) flow modeling underpins urban planning and mobility analysis, but prevailing graph-based methods often neglect salient geographic attributes, limiting their ability to model long-range and multi-area dependencies. In this paper, we introduce GeoFlow, a novel framework that (i) augments area representations with geospatial attributes, including relative positions, k-hop and geodesic distances, (ii) employs a specialized...

    arxiv.org/abs/2607.05257 · PDF

  14. 14

    FUSE: FK-Steered Multi-Modal Flow Matching for Efficient Simulation-Based Posterior Estimation

    Weichen Qin, Yufan Xie, Peihao Wang, Chia-Jui Chou, Minghui Du, Peng Xu, Ziren Luo, Yi Yang, Jingyi Yu, Bo Liang,...

    cs.LG

    Simulation-Based Inference (SBI) is critical for scientific discovery, with generative models offering a promising path toward efficient inference. However, existing methods struggle with effective multimodal modeling. They often rely on brute-force fusion strategies that ignore the structural disparities between parameters and observations, thus limiting estimation fidelity. In this work, we introduce FUSE (Feynman-Kac steered mUlti-modal...

    arxiv.org/abs/2607.05252 · PDF

  15. 15

    CanniUplift: A Holistic Framework for Mitigating Seller and Incentive Cannibalization in E-commerce Uplift Modeling

    Zuwang He, Shihao Shu, Yuli Qu, Hanyu Gao, Ziliang Zhang, Diwei Chen, Xiangda Yan, Buyu Gao, Tanchao Zhu, Yumeng Li,...

    cs.LG · cs.AI · cs.IR

    Personalized incentive allocation is vital for e-commerce, where uplift modeling is the standard for estimating Individual Treatment Effects (ITE). However, traditional models often fail in complex multi-seller environments with violations of the Stable Unit Treatment Value Assumption (SUTVA). We identify two critical challenges: Seller-level Cannibalization, where incentives shift expenditure between shops without growing the platform, and...

    arxiv.org/abs/2607.05242 · PDF

  16. 16

    Video-based detection of cessation of breathing in pre-term infants using machine learning

    Dineo Serame, Lionel Tarassenko, Mauricio Villarroel

    cs.LG

    Pre-term infants are susceptible to potentially harmful apnoea-related cessations of breathing due to immature respiratory control. However, reliable respiratory monitoring in the neonatal intensive care unit (NICU) remains challenging because motion artefacts, sensor displacement, and skin fragility can compromise contact-based measurements. Non-contact video monitoring offers a complementary approach that does not depend on adhesive sensors...

    arxiv.org/abs/2607.05230 · PDF

  17. 17

    FlatManifold: Robust Continual Learning under Severe Label Noise and Domain Shifts via Intrinsic Manifold Flattening

    Rai Hisada, Kanji Tanaka

    cs.LG

    In non-stationary streaming environments, simultaneously adapting to complex, non-linear domain shifts via continual learning while mitigating the catastrophic effects of severe, uncalibrated label noise poses a fundamental mathematical challenge. In this paper, we propose \FlatManifold{}, a novel, streamlined robust continual learning framework that utilizes a Nyström manifold flattening map based on the kernel trick and projection onto an...

    arxiv.org/abs/2607.05201 · PDF

  18. 18

    Noisy-Channel Minimum Bayes Risk Decoding

    Yusuke Sakai, Hidetaka Kamigaito, Taro Watanabe

    cs.LG · cs.AI · cs.CL

    Minimum Bayes Risk (MBR) decoding yields more robust and higher-quality text generation than maximum a posteriori (MAP) decoding by selecting hypotheses that maximize expected utility over sampled pseudo-references. However, there exists a discrepancy in the design: hypothesis selection calculates expected utility scores conditioned on given pseudo-references, while commonly used evaluation metrics, e.g., BLEU and COMET, are asymmetric....

    arxiv.org/abs/2607.05198 · PDF

  19. 19

    Latent Programming Horizons in Coding Agents

    André Silva, Han Tu, Martin Monperrus

    cs.LG · cs.SE

    A coding agent solving a software-engineering task spends dozens of steps reasoning, editing code, and running tests, yet little is known about what the underlying language model internally represents about the program it is working on. We show that the residual streams of language models under coding agents linearly encode properties of the evolving program: a logistic-regression probe on hidden states is able to decode whether the current...

    arxiv.org/abs/2607.05188 · PDF

  20. 20

    SMART: A Machine Learning and Monte Carlo Framework for Rapid Analysis of Stochastic Transistor Aging and Process Variation in Digital Circuits

    Arash Esshaghi, Siavash Es'haghi, Gholamreza Shahabadi, Alireza Moradi

    cs.LG · cs.AR

    As CMOS technology scales into the deep nanometer regime, digital circuit reliability is increasingly threatened by the combined stochastic effects of Bias Temperature Instability (BTI) and Process Variation (PV). Traditional reliability analysis methods, which rely on computationally intensive simulations or extensive lookup tables, fail to scale efficiently for large designs, creating a critical bottleneck in design space exploration. To...

    arxiv.org/abs/2607.05187 · PDF

  21. 21

    Relational Multi-Agent Reinforcement Learning for Dynamic Pricing in High-Speed Railway Markets

    Enrique Adrian Villarrubia-Martin, David Muñoz-Valero, Luis Rodriguez-Benitez, Giovanni Montana, Luis Jimenez-Linares

    cs.LG · cs.AI · cs.MA

    In liberalised railway systems, operators must set prices dynamically in an environment with partial observability, as they retain private information about their objectives and performance, where regulatory constraints prohibit communication or direct information exchange between competitors to prevent explicit collusion. Consequently, agents must learn to infer strategic interactions only from observable market data which presents a...

    arxiv.org/abs/2607.05179 · PDF

  22. 22

    Platonic Projection Structures: Operator-Induced Observability in Representation Learning

    Kazuo Ishii, Bishnu Prasad Gautam, Jieling Wu, Javaid Saher

    cs.LG

    We characterize observability in representation learning through Platonic Projection Structures (PPS), an operator-theoretic framework for analyzing representation accessibility under partial observation. Rather than treating observable outputs as direct reflections of latent representations, PPS models observation through a self-adjoint positive semidefinite operator acting on a latent representation space. A system is represented as a...

    arxiv.org/abs/2607.05175 · PDF

  23. 23

    MeGA-MP: Metric Graph Advection Message Passing -- A Physics-Informed Message Passing Operator for Advection-Dominated Metric Graphs

    Janine Strotherm, Luca Hermes, André Artelt, Barbara Hammer

    cs.LG

    Many real-world systems are organized as networks where spatio-temporal dynamics unfold along connections and not discretely between nodes. Examples include utility networks such as water distribution systems or gas networks, electrical grids, and traffic flow networks. Such systems are naturally modeled as metric graphs, where edges correspond to one-dimensional Euclidean subspaces connected at vertices. Metric graphs are independent of an...

    arxiv.org/abs/2607.05167 · PDF

  24. 24

    Physiological Noise Augmentation Improves Non-Invasive Brain-to-Speech

    Benjamin Ballyk, Teyun Kwon, Miran Özdogan, Oiwi Parker Jones

    cs.LG

    Non-invasive brain-to-speech decoding aims to restore communication to patients suffering from neurodegenerative disease, without the risks of neurosurgery. Existing MEG- and EEG-based methods, while scalable, continue to suffer from high word error rates driven by relatively low signal-to-noise ratios compared to invasive recordings. We propose physiological noise augmentation (PNA), a data augmentation method that explicitly trains decoders...

    arxiv.org/abs/2607.05165 · PDF

  25. 25

    PDEFlow: Autonomous Agentic PDE Pipelines for Neural Operator Learning and Solver-Free Inference

    Akshat Jani, Prathamesh Gadekar, Sakhinana Sagar Srinivas, Venkataramana Runkana

    cs.LG · cs.AI

    We present PDEFlow, an autonomous agentic framework that turns user-level ODE and PDE descriptions into solver-backed neural-operator pipelines. The workflow links problem specification, data generation, operator training, and checkpoint-based inference. A stateful input graph converts multi-turn natural-language input and user edits into validated problem specifications. The data-generation module then samples parameters, solves the...

    arxiv.org/abs/2607.05134 · PDF

  26. 26

    Localized LoRA-MoE: Block-wise Low-Rank Experts With Adaptive Routing

    Babak Barazandeh, Subhabrata Majumdar, Vinay Prithyani, George Michailidis

    cs.LG · cs.AI · cs.CL

    Large Language Models (LLMs) and high-dimensional perception networks increasingly rely on parameter-efficient fine-tuning (PEFT) to adapt to diverse operational contexts. However, standard methods like LoRA are structurally limited by a monolithic bottleneck, making them highly susceptible to gradient warfare. Interleaved multi-task streams may trigger destructive optimization feedback, collapsing adapter weights into unspecialized averages....

    arxiv.org/abs/2607.05114 · PDF

  27. 27

    Grokking Is Conditional and Fragile: A Fully-Tractable, Multi-Seed Study at 12K Parameters

    Yoshiyuki Ootani

    cs.LG · cs.AI

    Grokking -- the delayed onset of generalization long after a network has fit its training set - -is usually studied in models too large to read completely and reported from single training runs. We instead study a publicly released ~11,856-parameter Llama-style transformer (Glimmer-1-Base) on modular arithmetic, small enough to enumerate its weights, attention, and full input-output map, and we measure grokking as a multi-seed rate rather...

    arxiv.org/abs/2607.05104 · PDF

  28. 28

    Choosing a parallel heterogeneous ensemble method for tabular classification

    Vassili Maillet, Gustavo, Angulo, Pierre Jouvelot

    cs.LG

    Parallel ensemble methods were compared on $56$ small-to-medium tabular classification tasks drawn from OpenML CC18. A set of ``best practice'' recommendations on the use of ensemble methods was derived from these observations. It was later validated on 28 additional tasks using TabArena's precomputed data, where the recommendation set significantly outperformed Single Best and matched or exceeded individual ensemble methods. Two key...

    arxiv.org/abs/2607.05103 · PDF

  29. 29

    Counterfactual Methods for Detecting Unfairness in Anti-Money Laundering Algorithms

    Lea Multerer, Michele Inchingolo, David Kletz, Adrian Cosma, Alessandro Antonucci, Martina Gogova

    cs.LG

    The application of machine learning-based predictive algorithms to Anti-Money Laundering (AML) has grown rapidly, driven by the vast volume of financial transaction data available to banks. These algorithms are typically trained not only on transactional data but also on sensitive client information, which may raise fairness concerns. Despite this, AML detection systems remain largely underexplored from a fairness perspective, even though...

    arxiv.org/abs/2607.05101 · PDF

  30. 30

    Functional Bilevel Optimization for Predictive Fairness

    Ieva Petrulionyte, Julien Mairal, Michael Arbel

    cs.LG · stat.ML

    When sensitive attributes are continuous and high-dimensional $-$ demographic score vectors, posteriors over attributes, age or income profiles $-$ enforcing full statistical independence is often too restrictive, and existing relaxations rely on indirect dependence penalties or adversarial schemes that do not directly target the fairness-accuracy trade-off. We instead consider mean demographic parity through DPVar, the variance of the...

    arxiv.org/abs/2607.05098 · PDF

  31. 31

    FAST: A Holistic Framework for Optimizing Memory-I/O, Computation, and Sampling in Temporal GNN Training

    Yushu Cai, Qingrui Zhu, Lei Liu, Kai Sheng, Hao Chen, Xin He

    cs.LG

    Temporal Graph Neural Networks (TGNNs) are widely used for learning from dynamic graphs in applications such as recommendation, social network analysis, and traffic forecasting. However, scaling TGNN training to large dynamic graphs remains challenging due to three intertwined bottlenecks: memory I/O, irregular computation, and temporal neighbor sampling. Existing systems often optimize these stages in isolation, leaving substantial...

    arxiv.org/abs/2607.05095 · PDF

  32. 32

    Computing Monetary Risk Measures in Linear Time

    Palash Agrawal, Gersi Doko, Maeve Burwell, Marek Petrik

    cs.LG · cs.MS · stat.AP

    Monetary risk measures have gained popularity for expressing decision-makers' risk aversion. Value-at-Risk (VaR) and Conditional-Value-at-Risk (CVaR), in particular, are used commonly for this purpose. This paper proposes new efficient algorithms to compute these risk measures for a discrete random variable in expected linear time with respect to the size of its domain. First, we propose a QuickVaR algorithm that computes the VaR of a...

    arxiv.org/abs/2607.05078 · PDF

  33. 33

    KVpop -- Key-Value Cache Compression with Predictive Online Pruning

    Lukas Hauzenberger, Niklas Schmidinger, Anamaria-Roberta Hartl, David Stap, Thomas Schmied, Sebastian Böck, Günter...

    cs.LG

    Key-value (KV) cache growth is a major bottleneck in autoregressive decoding, as memory and bandwidth scale linearly with context length. Existing KV eviction methods often rely on static heuristics or proxy scores, which poorly track future token utility and cause brittle eviction as relevance shifts. To address this, we introduce KVpop, which learns a fixed-budget KV eviction policy by directly supervising the keep-or-drop decision. The...

    arxiv.org/abs/2607.05061 · PDF

  34. 34

    CollabEval: Statistically Efficient Collaborative Model Evaluation via Matrix Completion

    Adam Fisch, Daniel Deutsch, Joshua Maynez, Alekh Agarwal, Jonathan Berant, William Cohen, Amir Globerson, Jacob Eisenstein

    cs.LG

    Evaluating generative AI models is a routine, but resource-intensive, process that is conducted over and over again during the course of model development. In this work, we propose Collaborative Evaluation (CollabEval), a simple, effective, and principled method for exploiting dependencies between historical runs of different models on the same tasks to improve statistical efficiency. Specifically, our approach treats model evaluation as a...

    arxiv.org/abs/2607.05046 · PDF

  35. 35

    Uncertainty-aware damage identification in short-span bridges via physics-informed variational autoencoder

    Ana Fernandez Navamuel, A. Javier Omella, Diego Zamora-Sanchez, David Pardo

    cs.LG

    Vibration-based damage identification in civil infrastructure is a challenging, ill-posed inverse problem due to measurement noise, sparse sensor arrays, and environmental variability. While deep learning is powerful for system identification, deterministic approaches lack reliable uncertainty quantification and can yield physically inconsistent results. This work proposes a robust probabilistic Scientific Machine Learning (SciML) framework:...

    arxiv.org/abs/2607.05025 · PDF

  36. 36

    Beyond Modality Fusion: Deep Ensembles for Multimodal Classification

    Ilya Burenko, Dmitry Vetrov

    cs.LG · cs.CV

    In multimodal classification, late-fusion approaches classify concatenated modality-specific features extracted by unimodal neural networks. When modality imbalance is pronounced, various regularization techniques have been proposed to balance the learning process and overcome the inferior performance of late-fusion networks. In contrast, this work demonstrates that multimodal data can be effectively classified without any explicit modality...

    arxiv.org/abs/2607.05019 · PDF

  37. 37

    Hyperparameter Transfer in Graph Neural Networks

    Gage DeZoort, Boris Hanin

    cs.LG · cs.AI

    The performance of deep learning models crucially depends on the settings of hyperparameters like learning rate, initialization scale, and weight decay. Hyperparameter transfer aims to make near-optimal hyperparameter settings consistent across model scale, so that large models can be optimized by proxy tuning their smaller, cheaper-to-optimize counterparts. While transfer principles are well-studied in the context of dense neural networks in...

    arxiv.org/abs/2607.05017 · PDF

  38. 38

    ImputeECG: Deep Learning Reconstruction of Complete 12-Lead Electrocardiograms from Incomplete Recordings for Cardiac Assessment

    Xiaocheng Fang, Haoyu Wang, Jieyi Cai, Qinghao Zhao, Jun Li, Shanwei Zhang, Guangkun Nie, Yujie Xiao, Shun Huang,...

    cs.LG · cs.AI

    Complete digital 12-lead electrocardiograms (ECGs) are essential for AI-enabled cardiovascular assessment, yet many clinical ECG records, particularly those digitized from ECG images, remain incomplete because of short display formats, incomplete waveform digitization, lead loss, or signal corruption. We developed ImputeECG, a mask-conditioned one-dimensional Transformer autoencoder that completes 12-lead, 10-s ECGs while retaining all...

    arxiv.org/abs/2607.05009 · PDF

  39. 39

    The Map Behind the Flow: Finite-Step Gradient Descent as a Dynamical System

    Thomas Hofmann

    cs.LG · cs.AI

    Many phenomena of deep learning are dynamical: they concern not only which minima exist, but how gradient descent reaches, avoids, or selects among them. Edge-of-stability behavior, sharpness oscillations, catapult phases, balancing, and movement toward flatter representations are effects of the training map itself, and are poorly captured by the small-step gradient-flow limit. This paper studies fixed-step gradient descent as a discrete...

    arxiv.org/abs/2607.04993 · PDF

  40. 40

    Non-Convex Sparse Reinforcement Learning via Non-Monotone Inclusions

    Kyohei Suzuki, onstantinos Slavakis

    cs.LG

    This work delivers two key contributions: one to efficient feature selection in reinforcement learning (RL), the other to the theory of non-monotone inclusions. On the RL side, the estimation bias inherent in conventional regularization schemes is addressed by augmenting classical least-squares temporal-difference (LSTD) policy evaluation with the sparsity-inducing, non-convex projected minimax concave (PMC) penalty. Because the PMC penalty...

    arxiv.org/abs/2607.04990 · PDF

  41. 41

    Data-Driven Soft Labeling Scales DNA Read Classification to Whole-Body Cell-Type Deconvolution

    Dmytro Rizdvanetskyi, Nathan Ross, Pavlo Lutsik

    cs.LG · q-bio.GN · q-bio.QM

    Cell-type deconvolution, the task of estimating the proportions of constituent cell types in a heterogeneous biological sample, is a core problem in computational biology. Methods that rely on epigenetic marks such as DNA methylation typically operate on aggregated methylation estimates, discarding the pattern-level information carried by individual DNA reads. Existing read-level approaches that exploit this information are scarce, and all...

    arxiv.org/abs/2607.04987 · PDF

  42. 42

    Qantara: Bridge-Flow Training for Multi-Paradigm JEPA Control

    Ruslan Rakhimov, George Bredis, Yuriy Maksyuta, Daniil Gavrilov

    cs.LG · cs.CV · cs.RO

    Joint-Embedding Predictive Architectures (JEPAs) underpin a growing family of latent world models for control from raw pixels, but every existing JEPA world model commits at training time to a single inference paradigm: either trajectory optimisation in a learned dynamics model, or direct behaviour cloning. A single checkpoint that serves both would defer this choice to inference, when deployment constraints (rollout cost, observation...

    arxiv.org/abs/2607.04978 · PDF

  43. 43

    Geometry-Aware Bayesian Quantification via Compositional Data Analysis

    Alejandro Moreo, Pablo González, Juan José del Coz

    cs.LG · stat.ML

    Accurately estimating the unknown target label distribution is the critical first step for adapting to label shift. This task, widely known as quantification or class prevalence estimation, has recently seen significant advances through continuous KDE-based methods which model the density of multiclass classifier posteriors. Posterior vectors might be regarded as compositional data, since they lie on the probability simplex. However, existing...

    arxiv.org/abs/2607.04977 · PDF

  44. 44

    Train Smarter, Not Longer: Memorization-Guided Data Reuse for Efficient LLM Training

    Jingwei Zuo, Cong Zeng, Ilyas Chahed, Maksim Velikanov, Dhia Eddine Rhaiem, Pasquale Balsebre, Abhay Kumar, Younes...

    cs.LG · cs.CL

    The training paradigm of large language models has shifted from traditional one-pass training to multi-epoch training, as reasonable reuse of limited high-quality data can improve both model performance and sample efficiency. Meanwhile, excessive repetition introduces the risk of overfitting and diminishing returns. Determining when and how to reuse data effectively thus emerges as a natural but under-explored question. Through a novel...

    arxiv.org/abs/2607.04969 · PDF

  45. 45

    Sensitivity Sampling with Predictions for k-Means Clustering

    Cristian Boldrin, Fabio Vandin

    cs.LG · cs.DS

    We study the problem of k-means clustering on large datasets. The state-of-the-art for the problem is given by coresets-based approaches, which build small weighted summaries of the input and derive approximate solutions with rigorous quality guarantees from them. One of the most popular and advanced approaches to derive coresets for k-means is sensitivity sampling. However, sensitivity sampling requires to compute the importance of each...

    arxiv.org/abs/2607.04949 · PDF

  46. 46

    Lightweight ML-Based Automatic Sleep Staging Framework with Constrained CNN and Mamba for Small-Sample EEG Datasets

    Zihao Wei, Yulin Gong, Yudan Lv

    cs.LG

    Automatic sleep staging is a key technology for precise diagnosis and treatment of sleep disorders as well as long-term home sleep monitoring. Portable electroencephalogram (EEG) devices have become the focus of research due to their convenience in data collection. However, current methods still face three major challenges: large parameter sizes that easily lead to overfitting on small datasets, low accuracy in classifying difficult stages...

    arxiv.org/abs/2607.04934 · PDF

  47. 47

    Input Pathways Shape Few-Shot, Not Zero-Shot, Binding in Tiny Transformers: A Fully-Enumerable Study

    Yoshiyuki Ootani

    cs.LG · cs.AI

    How does the way information reaches a transformer -- as symbolic tokens, a clean per-factor "oracle" code, or an entangled perceptual vector -- shape whether it binds that information compositionally? We study ~6-10K-parameter transformers on finite factored worlds enumerated exhaustively, so every measurement covers the whole input space (zero sampling variance) and the informative routes are information-matched (exact Bayes ceiling 1.0)....

    arxiv.org/abs/2607.04926 · PDF

  48. 48

    When Do Foundation Models Pay Off? A Break-Even Analysis of Pretrained Time Series Forecasters

    Nicholas Tan Jerome, Frank Simon

    cs.LG

    Deploying a time series foundation model requires GPU infrastructure, engineering overhead, and carries no guarantee of improvement over XGBoost. We provide the first systematic break-even analysis answering when this investment pays off. Across 30 benchmark datasets, we compare zero-shot and LoRA fine-tuned foundation models (Chronos, Moirai, Lag-Llama) against classical baselines (Naive, ETS, ARIMA, XGBoost) at six training set sizes from...

    arxiv.org/abs/2607.04919 · PDF

  49. 49

    RL-Ballast: Ship Ballast Water Path Planning and Clog Prediction via Reinforcement Learning

    Ming-Kuan Lin, Yi-Chung Lai, Ming-Hsin Chiang, Tsung-Wei Pan, Jung-Hua Wang

    cs.LG

    Under the Shipping 4.0 paradigm, autonomous and reduced-crew vessels require intelligent internal systems to maintain operational safety and structural stability. Ballast-water control is essential for ship trim and integrity, but conventional rule-based or manual approaches have limited adaptability to hydraulic anomalies such as valve failures and pipe blockages, and often depend on dense pressure or flow sensors for diagnosis. To address...

    arxiv.org/abs/2607.04906 · PDF

  50. 50

    Active Learning on Adversarially Corrupted Graphs

    Marco Bressan, Nicolò Cesa-Bianchi, Tommaso d`Orsi, Emmanuel Esposito, Silvio Lattanzi

    cs.LG · stat.ML

    Motivated by real-world scenarios where malicious entities tamper with existing networks, we define a model where an adversary seeks to hide a set of \emph{corrupted vertices} inside a graph $G^*$. To this end, the adversary can add edges between the corrupted vertices, as well as edges between the corrupted vertices and $G^*$, and its power is then measured by the size of the \emph{neighborhood} of the corrupted vertices in $G^*$. Our goal...

    arxiv.org/abs/2607.04869 · PDF

  51. 51

    Enhancing the Forecasting Capability of Multi-Model Blending Algorithms for Extreme Precipitation via Joint Use of Station and Gridded Observations

    Yu Wang, Yong Cao, Kan Dai, Yue Shen, Xiaoqing Zeng, Ruixia Zhao

    cs.LG

    Accurate extreme precipitation forecasting is critical for disaster mitigation but remains challenging for numerical weather prediction (NWP) models due to systemic intensity underestimation and spatial displacement. Traditional precipitation multi-model blending algorithms perform pixel-by-pixel blending on the forecast field based on weights, which may lead to the expansion of precipitation areas and the smoothing of extreme values. This...

    arxiv.org/abs/2607.04862 · PDF

  52. 52

    Framework for Grouping Local Process Models

    Viki Peeva, Wil M. P. van der Aalst

    cs.LG

    Local Process Models (LPMs) are an underexplored concept in process mining. LPMs describe patterns in event data considering sequence, choice, concurrency, and loop. In recent years, process mining has proved successful in the analysis and improvement of operational processes. More often than not, surprising findings are found when one does not consider the full process, making LPMs and their discovery highly valuable. However, similar to...

    arxiv.org/abs/2607.04856 · PDF

  53. 53

    Pretraining Curricula Enable Selective Fine-tuning

    Sebastian A. Bruijns, Jirko Rubruck, Mia H. Whitefield, Kai J. Sandbrink, Fazl Barez, Christopher Summerfield

    cs.LG · cs.AI

    Transformers follow implicit curricula whereby some tasks are learned before others. However, how explicit pretraining curricula influence learning, generalization, and the selectivity of fine-tuning is unclear. This is important for AI safety, where fine-tuning is used to selectively suppress misaligned behaviors. Here, we compare curricula that pretrain tasks in a balanced (sampled uniformly) or an imbalanced (one task early, the other...

    arxiv.org/abs/2607.04846 · PDF

  54. 54

    Representing and Detecting Label Ambiguity in IMU-Based Exercise Evaluation

    Andreas Spilz, Heiko Oppel, Michael Munz

    cs.LG

    Home-based physiotherapy is performed without supervision, which leads to incorrect execution and motivates systems that assess movement automatically from inertial measurement units (IMUs). Such systems assign each repetition to a category, yet a relevant share of repetitions falls near a class boundary, where even trained raters disagree. Classifiers trained with one-hot labels collapse these borderline repetitions onto a single class and...

    arxiv.org/abs/2607.04842 · PDF

  55. 55

    Probably Correct Optimal Stable Matching under Two-Sided Uncertainty

    Andreas Athanasopoulos, Anne-Marie George, Christos Dimitrakakis

    cs.LG · stat.ML

    We study a sequential learning problem for stable matchings in two-sided markets where preferences on both sides are initially unknown. We focus on a centralized setting where an algorithm matches agents at each time step and receives noisy rewards that reflect the preferences of the matched agents, following a semi-bandit feedback structure. We adopt a pure exploration perspective, aiming to efficiently identify the optimal stable matching...

    arxiv.org/abs/2607.04824 · PDF

  56. 56

    KinEMbed: Decoding Kinematics from Electromyography via Cross-Modal Contrastive Learning

    Sofia Gilardini, Chenfei Ma, Kianoush Nazarpour

    cs.LG

    Decoding hand kinematics from surface electromyography (EMG) is a core challenge in wearable biosignal processing with clinical relevance for prosthetic control and motor rehabilitation. Most representation learning approaches for EMG focus on discrete gesture classification, and few focus on continuous regression. We present KinEMbed, a cross-modal contrastive learning framework for hand kinematics regression that jointly trains dual...

    arxiv.org/abs/2607.04820 · PDF

  57. 57

    Layer-Parallel Inference Reduces Encrypted Nonlinear Depth in Transformers

    Ligong Han, Kai Xu, Hao Wang, Ruijiang Gao, Akash Srivastava

    cs.LG · cs.CR

    Fully homomorphic encryption (FHE) enables computation on encrypted data, but practical encrypted Transformer inference is bottlenecked by the sequential composition of many nonlinear blocks. We study whether Structured Newton Layer Parallelism (SNLP) can make this inter-layer composition more FHE-friendly: each Transformer block still requires polynomial approximations for operations such as softmax and RMSNorm, but SNLP reduces the...

    arxiv.org/abs/2607.04819 · PDF

  58. 58

    Compressed Computation under $L^4$ Loss is likely Computation in Superposition

    Francisco Ferreira da Silva, Stefan Heimersheim

    cs.LG

    Neural networks are thought to represent concepts as directions in their activation space, and superposition lets them encode more concepts than they have dimensions. It is natural to ask whether they can also compute more functions than they have neurons, i.e., perform computation in superposition. In this regime many functions of sparse inputs are evaluated by a layer with fewer neurons than there are functions to compute. Representation in...

    arxiv.org/abs/2607.04800 · PDF

  59. 59

    Predicting Drafted Deck Strength for "Magic: the Gathering"

    Tomas Rigaux, Hisashi Kashima

    cs.LG · cs.AI

    Many real-world games do not admit a fixed, compact rule set: instead, their dynamics are defined by interactions among a large and often evolving collection of game pieces, making general-purpose policy learning impractical. Magic: the Gathering (MTG) exemplifies this setting, where the cards themselves define and alter gameplay rules, strategic constraints, and long-term outcomes, while the pool of available cards is ever-changing. We study...

    arxiv.org/abs/2607.04782 · PDF

  60. 60

    Towards Personalized Differentially Private Learning for Decentralized Local Graphs

    Longzhu He, Peng Tang, Chaozhuo Li, Jinhu Fu, Litian Zhang, Li Sun, Philip S. Yu, Sen Su

    cs.LG · cs.CR

    Graph-structured data is increasingly generated and stored in decentralized environments, such as social platforms, mobile applications, and edge networks, where users maintain control over their local graph data. However, collecting and analyzing such decentralized graph data for downstream learning tasks raises significant privacy concerns, as nodes and their attributes often contain sensitive personal information. Local Differential...

    arxiv.org/abs/2607.04777 · PDF

  61. 61

    MARLIN: De Novo Molecular Structure Elucidation from Tandem Mass Spectra without a Ground-Truth Formula

    Xujun Che, Xiuxia Du, Depeng Xu

    cs.LG

    Untargeted tandem mass spectrometry (MS/MS) detects thousands of small molecules per biological sample, yet most go unidentified because they are absent from spectral libraries. These uncharacterized metabolites and natural products are precisely the compounds that matter for drug discovery, biomarker research, and exposomics. Computational de novo structure elucidation could close this gap, but almost all state-of-the-art methods assume the...

    arxiv.org/abs/2607.04774 · PDF

  62. 62

    Multi-Turn On-Policy Distillation with Prefix Replay

    Baohao Liao, Hanze Dong, Christof Monz, Xinxing Xu, Li Dong, Furu Wei

    cs.LG · cs.AI · cs.CL · stat.ML

    We study on-policy distillation (OPD) for agentic tasks, where an LLM agent interacts with an environment over multiple turns and a student imitates a teacher over these multi-turn interaction histories. Fully online OPD is costly because each update requires fresh student rollouts through the environment and teacher queries at visited histories. We propose Replayed-Prefix On-Policy Distillation (ReOPD), an off-environment alternative that...

    arxiv.org/abs/2607.04763 · PDF

  63. 63

    Trust Region Policy Distillation

    Zhengpeng Xie, Li Lyna Zhang, Zeke Xie, Mao Yang

    cs.LG · cs.AI

    Big goals are hard to achieve all at once; breaking them into small steps is wiser. We present Trust Region Policy Distillation (TOP-D), which transforms the notoriously unstable, high-variance On-Policy Distillation (OPD) into a stable training paradigm by dynamically constructing a proximal teacher. Theoretically, we establish a rigorous framework demonstrating that TOP-D inherently controls gradient variance. By providing a formal global...

    arxiv.org/abs/2607.04751 · PDF

  64. 64

    RSPO: Reward-Swap Policy Optimization for Multi-Turn LLM Agents

    Qiang Liu, Taian Guo, Ruizhi Qiao, Xing Sun

    cs.LG · cs.AI

    Reinforcement learning holds significant potential for training large language models (LLMs) to handle multi-turn interactive tasks. However, in long-horizon, multi-turn tasks characterized by sparse outcome rewards, directly training with outcome rewards often results in slow convergence due to the sparsity of signals and the lack of fine-grained feedback. Furthermore, the model may fail to learn successful trajectories that are not sampled...

    arxiv.org/abs/2607.04713 · PDF

  65. 65

    Hierarchical Scaffolding Enables Human-Like Cognitive Selectivity under Data Scarcity

    Juhyoung Park, Jaehyuk Bae, Hyeonbo Yang, Se-Bum Paik

    cs.LG · cs.CV

    Modern machine learning systems demand extensive datasets for visual recognition. Conversely, humans learn with high efficiency despite severe data limitations, often by acquiring broad categorical structures before refining finer distinctions. Inspired by this contrast, we introduce SCALA (Scaffolded Cognitive Architecture for Learning under limited dAta), a hierarchical learning framework grounded in cognitive psychology that guides models...

    arxiv.org/abs/2607.04709 · PDF

  66. 66

    F-ACVAE: A Federated Adaptive Conditional Variational Auto-Encoder for Privacy-Preserving Intrusion Detection in IoT Networks

    Mohammad Ansarimehr, Somayeh Changiz, Ehsan Baghishani, Ali Mousavi

    cs.LG · cs.CR

    The rapid proliferation of Internet of things (IoT) devices has significantly expanded the cyber-attack surface, necessitating robust and privacy-preserving intrusion detection systems (IDS). However, centralized learning approaches often suffer from severe performance degradation due to high-dimensional traffic data, extreme class imbalance, and highly non-independent and identically distributed (non-IID) data across heterogeneous edge...

    arxiv.org/abs/2607.04698 · PDF

  67. 67

    URSA: Chemistry-Aware Benchmark for Utilitarian Retrosynthesis Assessment

    Bogdan Zagribelnyy, Ivan Ilin, Nikita Bondarev, Anton Morgunov, Arkadii Lin, Maksim Kuznetsov, Rim Shayakhmetov,...

    cs.LG · cs.AI · cs.CE · cs.CL

    Synthesis planning aiming to find pathways of reactions for a target molecule is one of the most important and challenging tasks in drug discovery. Recent progress has produced both specialized deep-learning retrosynthesis systems and general-purpose large language models, but objective comparison remains difficult due to the lack of flexible, chemically interpretable benchmarking protocols. In the current study, we are introducing the URSA...

    arxiv.org/abs/2607.04688 · PDF

  68. 68

    A Physics-Regulated Neural Framework for Learning 3D Grain Growth Dynamics

    Zhihui Tian, Kang Yang, Michael Tonks, Amanda R. Krause, Joel B. Harley

    cs.LG · physics.comp-ph

    Grain growth is governed by the reduction in grain boundary energy and exhibits well-established statistical scaling laws. Developing data-driven surrogates that preserve these physical invariants while remaining computationally scalable remains challenging, especially in 3D. We present 3D-PRIMME (Physics-Regulated Interpretable Machine Learning for Microstructure Evolution) for learning three-dimensional grain growth dynamics. The model is...

    arxiv.org/abs/2607.04680 · PDF

  69. 69

    Machine Learning for Depression Screening and Intervention: an Original Circadian Rhythm Score-based Methodology

    Bin Wang, Shuo Lian, Yuanyuan Hou, Dexian Wang, Peilan He, Feng Hong, Yanwei Yu, Tianrui Li

    cs.LG · cs.AI

    Depression screening from large-scale behavioral data is challenged by fragmented circadian indicators, limited interpretability, and the lack of intervention-oriented analysis. Existing approaches typically analyze sleep, activity, and social behaviors in isolation, failing to capture their joint circadian structure. To address this limitation, we first propose the Circadian Rhythm Score (CRS), a composite index that compresses multi-domain...

    arxiv.org/abs/2607.04648 · PDF

  70. 70

    Reliability and Identifiability in Persona-Trained Monte Carlo: Variance Decomposition, Stability Bounds, and the Identifiability of Heterogeneous News Reaction

    Salavat Ishbulatov

    cs.LG · cs.CE

    Persona-Trained Monte Carlo (PTMC) estimates distributions of market-outcome functionals by repeatedly simulating limit-order-book interaction among $K$ neural policy bots whose behavioral personas are drawn from a learned heterogeneity distribution $\mathcal{P}$. This paper develops the statistical theory that makes the word "reliable" precise for such estimators. We decompose estimator variance into a persona-draw component $σ_P^2$ and a...

    arxiv.org/abs/2607.04627 · PDF

  71. 71

    Measuring What Matters: A Unified Evaluation Framework for GNN Explainability

    Francesco Paolo Nerini, Mirko Zaffaroni, Paolo Baracco, Gabriele Ciravegna, Alan Perotti

    cs.LG

    Graph eXplainable AI (G-XAI) is increasingly important for making Graph Neural Networks interpretable and accountable. While a growing number of explainers are available, choosing the right method and assessing the trustworthiness of its outputs remains unclear. Consistent evaluation practices and actionable guidance are still missing, hindering practical adoption. In this paper, we introduce a unified, quantitative benchmarking framework for...

    arxiv.org/abs/2607.04600 · PDF

  72. 72

    Minimum Block Width for Universal Approximation by Residual Neural Networks with Inner Width One

    Qi Zhou, Xuan Zhou, Xiao-Song Yang

    cs.LG · stat.ML

    In this paper, we study the universal approximation property of residual neural networks, and obtain some new results. For input and output dimensions $d_x$ and $d_y$, and LeakyReLU, ReLU, ReLU-like activation functions, the upper and lower bounds of the block width are established. To achieve $L^p$ approximation $(1\leq p <+\infty)$ on any compact domain, we show that the exact minimum block width is $\max\{d_x,d_y\}$ when the inner width is...

    arxiv.org/abs/2607.04597 · PDF

  73. 73

    Score Distributions, Not Cells: Evaluating Single-Cell Perturbations Under Class Overlap

    Youssef Marrakchi, Davide D'Ascenzo, Sebastiano Cultrera di Montesano

    cs.LG · q-bio.QM · stat.ML

    Most classification problems assume the classes are roughly separable, so that an individual sample can usually be assigned to one class. Single-cell perturbation data violates this assumption: two perturbations can produce different populations of cells while overlapping so much that an individual cell could belong to either. Per-cell accuracy then measures this overlap rather than model quality. We see this on Tahoe-100M and the Virtual...

    arxiv.org/abs/2607.04595 · 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.