Machine Learning

Papers filed under cs.LG on arXiv, each one already summarized by Paperlayer. Open any of them to read the summary beside the original PDF, with every point linked to the line, figure, or table it came from.

Search paper metadata (including unsummarized papers)

20,041 to 20,100 of 20,205

GPT-Red: Automated Red Teaming via Self-Play at Scale
Eric Wallace, Christopher A. Choquette-Choo, Nikhil Kandpal +15
cs.CRcs.AIcs.CLarXiv:2607.26115v12026
Stale but Stable: Staleness-Adaptive Trust Regions for Stabilizing Asynchronous Reinforcement Learning
Junyao Yang, Yucheng Shi, Zongxia Li +6
cs.LGcs.CLarXiv:2607.18722v32026
MindForge: Teaching Small Language Models Whole-Life-Cycle Software Engineering via Source-Free Program Synthesis
Yihao Chen, Shi Chang, Khaled Chawa +4
cs.SEcs.CLcs.LGarXiv:2607.27146v12026
QQWorld: Quantile-Quantile Matching for World Model Regularization
Zhoushun Yu, Xiaoyu Hu, Xiangyu Xu
cs.LGcs.AIcs.CVarXiv:2607.28415v12026
You Only Pass Once: Answering and Abstaining Together in a Single Forward Pass of a Frozen Language Model
Ziyang Luo, Zhongyao Chu, Xinjie He +4
cs.CLcs.LGarXiv:2608.14465v12026
More Correct Mass, Worse Answers: Why Power Sampling Can Fail and How to Fix It
Haohui Yang, Jiaxing Sun, Xiujun Ma
cs.LGarXiv:2608.14420v12026
Non-Shattering at and Above the Dynamical Temperature in the Spherical Pure p-Spin Model
Taegyun Kim
math.PRcond-mat.dis-nncs.LGarXiv:2608.14369v12026
Style or Signature? Artist-Disjoint Evaluation of Style Classification in Frozen Vision Embeddings
Rory Ashton
cs.CVcs.LGarXiv:2608.14435v12026
Beyond Euclidean Clipping: Overcoming Exploration Collapse in LLM RL via Riemannian Isometric Policy Optimization
Zhicheng Cai, Xinyuan Guo, Hanlin Wu +4
cs.LGcs.AIarXiv:2607.10169v12026
$π\mathbf{R}^2$: Reactive Real-time Flow Policies
Sungjae Park, Shubham Tulsiani
cs.ROcs.AIcs.LGarXiv:2607.26055v12026
Can AI agents conduct open-ended AI research? Early evidence from two case studies
Peter Kirgis, Sayash Kapoor, Andrew Schwartz +21
cs.AIcs.CYcs.LGarXiv:2607.27191v22026
NexForge: Scaling Agent Capabilities through Requirement-Driven Task Synthesis for LLMs
Jiarong Zhao, Zhikai Lei, Zhiheng Xi +5
cs.SEcs.AIcs.LGarXiv:2607.14186v62026
CLBench-V: Evaluating Multimodal Context Learning from Grounding to Knowledge Acquisition
Lai Wei, Chengqi Li, Jiapeng Li +3
cs.CVcs.AIcs.CLarXiv:2607.25294v12026
Flux-OPD: On-Policy Distillation with Evolving Contexts
Yuran Wang, Zekun Wang, Bohan Zeng +10
cs.LGcs.AIarXiv:2607.28022v12026
Molt: A Scalable PyTorch-Native Training Framework for Agentic Reinforcement Learning
Jian Hu, Huiying Li, Hao Zhang +8
cs.LGcs.CLcs.DCarXiv:2607.21653v12026
ReToken: One Token to Improve Vision-Language Models for Visual Retrieval
Yao Xiao, Reuben Tan, Zhen Zhu +3
cs.CVcs.AIcs.LGarXiv:2607.28627v12026
HiFi-UMI: Learning Deployable Manipulation Policies from High-Fidelity UMI Data Alone
Simple AI, :, Yuteng Wei +16
cs.ROcs.CVcs.LGarXiv:2607.25895v12026
Echoverse: Deep, Evolving Environments for Training Computer-Use Agents at Scale
Yash Pandya, Sahil Gupta, Sarthak Harne +10
cs.AIcs.LGarXiv:2607.28074v12026
EMBL AI Librarian: Life-Sciences Knowledge Layer for AI Agents
Luigi Sigillo, Matteo Silvestri, Francesco Tabaro +9
cs.CLcs.AIcs.IRarXiv:2607.28229v12026
SkillRise: Agentic Reinforcement Learning for Cross-Task Skill Evolution
Zhiyuan Yao, Yuxin Chen, Zhengxi Lu +13
cs.LGcs.AIarXiv:2607.26784v12026
MedMix: Specialization-Consistent Federated Sparse MoEs under Modality Heterogeneity
Adiba Orzikulova, Dong Min Kim, Jaehong Yoon +1
cs.LGarXiv:2608.13911v12026
Explorative Modeling: Unlocking a Third Pretraining Axis and End-to-End Generation
Alexi Gladstone, Heng Ji, Yilun Du
cs.LGcs.AIcs.CLarXiv:2607.27372v12026
Robust Dual-Model Collaborative Random Vector Functional Link Network
A. Quadir, A. Rahaman, Mushir Akhtar +1
cs.LGarXiv:2608.13628v12026
Catching the Imposter: Self-Supervised Learning of Physical Coherence with Cross-Entity Feature Permutations
Aleksei Rozanov, Arvind Renganathan, Vipin Kumar
cs.LGarXiv:2608.14372v12026
Weak-to-Strong On-Policy Distillation
Fangxu Yu, Weijia Xu, Michael Xu +2
cs.LGarXiv:2607.26246v22026
A Vocabulary for Multi-Agent Automated Research Systems
Bardiya Akhbari
cs.AIcs.LGcs.MAarXiv:2607.22682v12026
ODEWorld: A Continuous Predictive Architecture via Physical-Time Flow
Dongxiu Liu, Haoyi Niu, Peng Cheng +5
cs.LGcs.CVcs.ROarXiv:2607.27924v22026
When Top-K Misses the Decision: Tool-Call Drift in Multi-Teacher On-Policy Distillation
Jiabin Shen, Guang Chen, Chengjun Mao
cs.CLcs.LGarXiv:2607.07050v42026
AskChem: Claim-Centered Infrastructure for Chemistry Literature Synthesis
Bing Yan, Gregory Wolfe, Stefano Martiniani +1
cs.CLcs.AIcs.IRarXiv:2607.28618v12026
An Exam for Active Observers
Jiarui Zhang, Muzi Tao, Shangshang Wang +3
cs.CVcs.AIcs.CLarXiv:2607.16165v12026
Interactive Training 2: Auditable Control Plane for Live Model Training
Wentao Zhang, Xuanhe Pan, Han Zhou +2
cs.LGarXiv:2607.18314v12026
Metis: Memory Foundation Model
Zeyu Zhang, Ziliang Guo, Yihang Sun +14
cs.CLcs.LGarXiv:2607.26760v22026
CADENCE: Closing the Reasoning Gap via Coverage-Adaptive On-Policy Distillation
Satyam Kumar, Saurabh Jha
cs.LGcs.AIarXiv:2607.16955v12026
Multi-Turn On-Policy Distillation with Prefix Replay
Baohao Liao, Hanze Dong, Christof Monz +3
cs.LGcs.AIcs.CLarXiv:2607.04763v32026
Agentic Transaction: Towards ACID-Compliant Agent Systems
Zhaoyan Sun, Xiaoxiao Wang, Guoliang Li
cs.DBcs.AIcs.CLarXiv:2608.13900v12026
Engineering Signals of Human-AI Collaboration in the Agentic Coding Era: A Longitudinal Analysis of 33,228 Pull Requests from vLLM and SGLang with Implications for Biomedical AI Agents and Bioinformatics Pipeline Developmen
Jiada Li, Xuesong Ye, Olamide Olowoniyi
cs.SEcs.AIcs.ETarXiv:2608.13884v12026
Split the Labor: Separating Evidence Interpretation from Decision Aggregation
Zhelun Wu
cs.AIcs.CLcs.LGarXiv:2608.14509v12026
Attributing Preprocessing Invariance in Spectral Foundation Models
Dongjun Wei, Hongyi Wu, Yinuo Zou
cs.AIcs.CEcs.LGarXiv:2608.14227v12026
Evaluating Agentic Learning Harness Capabilities Without Labels via the Scaling Hypothesis
Aryan Luthra, Kshitij Jain, Siddharth Arya +2
cs.AIcs.CRcs.LGarXiv:2608.13608v12026
The Query Knows What to Forget: A Second Erase Direction for Linear Attention
Dhruman Gupta, Aritra Das, Debayan Gupta
cs.LGarXiv:2608.13668v12026
Helping Music Co-Creation Agents 'Listen' Well: Hierarchical Self-Supervised World Models for Understanding and Generation
Scott H. Hawley
cs.SDcs.LGeess.ASarXiv:2608.04378v12026
FreeBalance: Pre-Routing Online Moe Load Balancing via Residual Workload Prediction
Pengfei Chen, Yize Wu, Shouxu Kuang +2
cs.AIcs.LGarXiv:2608.14205v12026
Removing Temporal Note Redundancy Improves Multimodal Reinforcement Learning for Medicine
Chenran Weng, Joo Seung Lee, Malini Mahendra +1
cs.AIcs.LGarXiv:2608.14157v12026
DRIFT: Derailing Denoising Trajectories of Flow-Matching VLAs with Adversarial Patch Attack
Hoseong Tae, Jong-Seok Lee
cs.CVcs.LGarXiv:2608.03207v12026
Adjacency-Based Spectral Proxy Control of Mobile Communication Agents
Mariana del Castillo, Federico Larroca
cs.ROcs.LGcs.MAarXiv:2608.13616v12026
Omega-S: A Functional Resilience Index for LLM Fine-Tuning
Alberto Acedo
cs.LGcs.NEq-bio.MNarXiv:2608.03887v12026
FactorJEPA: Factorizing Monolithic Futures into Layout-Agent-Interaction Channels for Crowded and Chaotic Global South Urban Worlds
Kapil Wanaskar, Gaytri Jena, Aman Chadha +3
cs.AIcs.CVcs.LGarXiv:2608.01049v12026
To Add Is Machine, To Delete Is Human: Measuring and Mitigating Deletion Avoidance in LLM Code Editing
Amir M. Ebrahimi, Mohammed Mehedi Hasan, Aaditya Bhatia +2
cs.SEcs.AIcs.LGarXiv:2607.28887v12026
Towards Interpretable Foundation Models for Retinal Fundus Images
Samuel Ofosu Mensah, Camila Roa, Kerol Djoumessi +1
cs.CVcs.LGstat.COarXiv:2603.18846v42026
EEG-PRISM: Physiologically-Grounded Interpretability of Predictions by EEG Foundation Models
Deeksha M Shama, Punnisa Amornsirikul, Archana Venkataraman
cs.LGarXiv:2608.13676v12026
Sequence prediction under a lying oracle
Puspabeethi Samanta, Nikhil Karamchandani, Jayakrishnan Nair
cs.LGcs.ITarXiv:2608.14102v12026
Model-agnostic Retrieval-Augmented Extended Forecasting for time series
Juan Pablo Villa Serna, Rohan Asthana, Vasileios Belagiannis
cs.LGarXiv:2608.14054v12026
Geometric Filtering of LLM-Generated Samples for Few-Shot Text Classification
Benjamín Schindler, Gonzalo A. Ruz
cs.LGcs.CLarXiv:2608.13866v12026
Structure-Guided Spatiotemporal Attention Graph Neural Network for Traffic Flow Prediction
Xuanmian He, Can Li, Wanjing Ma
cs.LGcs.AIarXiv:2608.14177v12026
RestoreKV: Recovering Full-Cache Behavior Under Aggressive Query-Agnostic KV Cache Eviction
Changwoo Baek, Seungjun Shin, Kyeongbo Kong
cs.CLcs.LGarXiv:2608.01247v12026
ARCHead: Activation-Metric Residual Correction for Large Language Model Output Heads
Şuayp Talha Kocabay, Talha Rüzgar Akkuş, Kamer Ali Yuksel
cs.CLcs.LGarXiv:2608.02703v12026
Round-Trip Consistency: Bidirectional Diffusion Models Can Predict Their Own Rollout Errors
Alexander Scheinker
stat.MLcs.LGphysics.comp-pharXiv:2608.00675v22026
SAGE: Surrogate-gradient Adaptation via Attention-Guided Entropy for Spiking Transformers
Kiran Nair, Rodrigue Rizk, KC Santosh
cs.LGcs.AIcs.CVarXiv:2608.13702v12026
AgentStream: How Well Do Self-Evolving LLM Agents Perform Under Streaming Tasks?
Dong Yan, Jian Liang, Dapeng Hu +4
cs.AIcs.LGarXiv:2608.00155v12026
TRUE-Colon: Exposing a Consistent Transfer Asymmetry in Real-Time Polyp Detection
Sebastian Doerrich, Andreas Franz Schwab, Francesco Di Salvo +3
eess.IVcs.CVcs.LGarXiv:2608.13711v12026

Machine Learning

GPT-Red: Automated Red Teaming via Self-Play at Scale

Stale but Stable: Staleness-Adaptive Trust Regions for Stabilizing Asynchronous Reinforcement Learning

MindForge: Teaching Small Language Models Whole-Life-Cycle Software Engineering via Source-Free Program Synthesis

QQWorld: Quantile-Quantile Matching for World Model Regularization

You Only Pass Once: Answering and Abstaining Together in a Single Forward Pass of a Frozen Language Model

More Correct Mass, Worse Answers: Why Power Sampling Can Fail and How to Fix It

Non-Shattering at and Above the Dynamical Temperature in the Spherical Pure p-Spin Model

Style or Signature? Artist-Disjoint Evaluation of Style Classification in Frozen Vision Embeddings

Beyond Euclidean Clipping: Overcoming Exploration Collapse in LLM RL via Riemannian Isometric Policy Optimization

$π\mathbf{R}^2$: Reactive Real-time Flow Policies

Can AI agents conduct open-ended AI research? Early evidence from two case studies

NexForge: Scaling Agent Capabilities through Requirement-Driven Task Synthesis for LLMs

CLBench-V: Evaluating Multimodal Context Learning from Grounding to Knowledge Acquisition

Flux-OPD: On-Policy Distillation with Evolving Contexts

Molt: A Scalable PyTorch-Native Training Framework for Agentic Reinforcement Learning

ReToken: One Token to Improve Vision-Language Models for Visual Retrieval

HiFi-UMI: Learning Deployable Manipulation Policies from High-Fidelity UMI Data Alone

Echoverse: Deep, Evolving Environments for Training Computer-Use Agents at Scale

EMBL AI Librarian: Life-Sciences Knowledge Layer for AI Agents

SkillRise: Agentic Reinforcement Learning for Cross-Task Skill Evolution

MedMix: Specialization-Consistent Federated Sparse MoEs under Modality Heterogeneity

Explorative Modeling: Unlocking a Third Pretraining Axis and End-to-End Generation

Robust Dual-Model Collaborative Random Vector Functional Link Network

Catching the Imposter: Self-Supervised Learning of Physical Coherence with Cross-Entity Feature Permutations

Weak-to-Strong On-Policy Distillation

A Vocabulary for Multi-Agent Automated Research Systems

ODEWorld: A Continuous Predictive Architecture via Physical-Time Flow

When Top-K Misses the Decision: Tool-Call Drift in Multi-Teacher On-Policy Distillation

AskChem: Claim-Centered Infrastructure for Chemistry Literature Synthesis

An Exam for Active Observers

Interactive Training 2: Auditable Control Plane for Live Model Training

Metis: Memory Foundation Model

CADENCE: Closing the Reasoning Gap via Coverage-Adaptive On-Policy Distillation

Multi-Turn On-Policy Distillation with Prefix Replay

Agentic Transaction: Towards ACID-Compliant Agent Systems

Engineering Signals of Human-AI Collaboration in the Agentic Coding Era: A Longitudinal Analysis of 33,228 Pull Requests from vLLM and SGLang with Implications for Biomedical AI Agents and Bioinformatics Pipeline Developmen

Split the Labor: Separating Evidence Interpretation from Decision Aggregation

Attributing Preprocessing Invariance in Spectral Foundation Models

Evaluating Agentic Learning Harness Capabilities Without Labels via the Scaling Hypothesis

The Query Knows What to Forget: A Second Erase Direction for Linear Attention

Helping Music Co-Creation Agents 'Listen' Well: Hierarchical Self-Supervised World Models for Understanding and Generation

FreeBalance: Pre-Routing Online Moe Load Balancing via Residual Workload Prediction

Removing Temporal Note Redundancy Improves Multimodal Reinforcement Learning for Medicine

DRIFT: Derailing Denoising Trajectories of Flow-Matching VLAs with Adversarial Patch Attack

Adjacency-Based Spectral Proxy Control of Mobile Communication Agents

Omega-S: A Functional Resilience Index for LLM Fine-Tuning

FactorJEPA: Factorizing Monolithic Futures into Layout-Agent-Interaction Channels for Crowded and Chaotic Global South Urban Worlds

To Add Is Machine, To Delete Is Human: Measuring and Mitigating Deletion Avoidance in LLM Code Editing

Towards Interpretable Foundation Models for Retinal Fundus Images

EEG-PRISM: Physiologically-Grounded Interpretability of Predictions by EEG Foundation Models

Sequence prediction under a lying oracle

Model-agnostic Retrieval-Augmented Extended Forecasting for time series

Geometric Filtering of LLM-Generated Samples for Few-Shot Text Classification

Structure-Guided Spatiotemporal Attention Graph Neural Network for Traffic Flow Prediction

RestoreKV: Recovering Full-Cache Behavior Under Aggressive Query-Agnostic KV Cache Eviction

ARCHead: Activation-Metric Residual Correction for Large Language Model Output Heads

Round-Trip Consistency: Bidirectional Diffusion Models Can Predict Their Own Rollout Errors

SAGE: Surrogate-gradient Adaptation via Attention-Guided Entropy for Spiking Transformers

AgentStream: How Well Do Self-Evolving LLM Agents Perform Under Streaming Tasks?

TRUE-Colon: Exposing a Consistent Transfer Asymmetry in Real-Time Polyp Detection