世界模型 - arXivDaily 专题

2602.10476 2026-02-12 cs.LG cs.AI 56%

Driving Reaction Trajectories via Latent Flow Matching

通过潜在流匹配驱动反应轨迹

Yili Shen, Xiangliang Zhang

机构 * University of Notre Dame（诺丁汉大学）

专题命中模型式强化学习：latent dynamics(abstract)；分类 cs.AI、cs.LG

AI总结 LatentRxnFlow通过连续潜在轨迹建模提升反应预测的准确性与透明性，实现轨迹级诊断与不确定性分析。

2602.03511 2026-02-04 cs.RO cs.AI 56%

CMR: Contractive Mapping Embeddings for Robust Humanoid Locomotion on Unstructured Terrains

CMR：用于在无结构地形上稳健人形运动的收缩映射嵌入

Qixin Zeng, Hongyin Zhang, Shangke Lyu, Junxi Jin, Donglin Wang, Chao Huang

机构 * University of Southampton（索姆塞特大学）； Westlake University（西lake大学）； Nanjing University（南京大学）

专题命中模型式强化学习：latent dynamics(abstract)；分类 cs.AI、cs.RO

AI总结 CMR通过结合对比表征学习与Lipschitz正则化，提升人形机器人在无结构地形中抗扰动的运动能力。

2601.21988 2026-01-30 cs.LG cs.AI cs.MA cs.RO cs.SY eess.SY 56%

Generalized Information Gathering Under Dynamics Uncertainty

在动态不确定性下的一般信息收集

Fernando Palafox, Jingqi Li, Jesse Milzman, David Fridovich-Keil

机构 * University of Texas at Austin（德克萨斯大学奥斯汀分校）； DEVCOM Army Research Laboratory（陆军研究实验室）

专题命中模型式强化学习：分类 cs.AI、cs.LG、cs.RO；dynamics model(abstract)

AI总结本文提出了一种统一框架，用于在动态不确定性下一般化信息收集，通过解耦动态模型、信念更新等选择与信息收集成本，提供理论依据和实验验证。

2601.19551 2026-01-28 cs.LG cs.AI 56%

Scale-Consistent State-Space Dynamics via Fractal of Stationary Transformations

通过稳定变换的分形实现尺度一致的状态空间动力学

Geunhyeok Yu, Hyoseok Hwang

机构 * Department of Software Convergence, Kyung Hee University（软件融合系，庆熙大学）

专题命中模型式强化学习：latent dynamics(abstract)；分类 cs.AI、cs.LG

AI总结本文提出FROST方法，通过分形归纳偏置实现状态空间模型的尺度一致潜在动力学，提升模型的自适应效率和稳定性。

Comments 8 pages (excluding 2 pages of references), 3 tables, 2 figures. Appendix: 4 pages

2512.15430 2025-12-18 cs.LG cs.AI 56%

FM-EAC: Feature Model-based Enhanced Actor-Critic for Multi-Task Control in Dynamic Environments

基于特征模型的增强型Actor-Critic：用于动态环境多任务控制

Quanxi Zhou, Wencan Mao, Manabu Tsukada, John C. S. Lui, Yusheng Ji

机构 * The University of Tokyo（东京大学）； National Institute of Informatics（日本信息处理学会）； The Chinese University of Hong Kong（香港中文大学）

专题命中模型式强化学习：model-based reinforcement learning(abstract)；分类 cs.AI、cs.LG

AI总结 FM-EAC通过整合规划、行动和学习，提升动态环境多任务控制的泛化能力与任务迁移性能。

2503.10253 2025-12-16 cs.LG cs.AI 56%

PIMRL: Physics-Informed Multi-Scale Recurrent Learning for Burst-Sampled Spatiotemporal Dynamics

PIMRL：基于突发采样的多尺度递归学习用于时空动态

Han Wan, Qi Wang, Yuan Mi, Rui Zhang, Hao Sun

专题命中模型式强化学习：latent dynamics(abstract)；分类 cs.AI、cs.LG

AI总结 PIMRL通过物理引导的多尺度递归学习，有效处理突发采样的稀疏时空动态数据，显著提升建模精度和效率。

Comments To appear in AAAI 2026

2509.23846 2025-12-03 cs.LG cs.AI 56%

Adversarial Diffusion for Robust Reinforcement Learning

对抗扩散用于鲁棒强化学习

Daniele Foffano, Alessio Russo, Alexandre Proutiere

机构 * Division of Decision and Control Systems KTH, Royal Institute of Technology（决策与控制系统系瑞典皇家理工学院）； Faculty of Computing and Data Sciences（计算与数据科学学院）

专题命中模型式强化学习：model-based RL(abstract)；分类 cs.AI、cs.LG

AI总结本文提出AD-RRL，通过对抗扩散方法提升强化学习在环境动态不确定性下的鲁棒性和性能。

2511.16427 2025-11-21 cs.LG cs.AI 56%

Generative Modeling of Clinical Time Series via Latent Stochastic Differential Equations

通过潜在随机微分方程进行临床时间序列生成建模

Muhammad Aslanimoghanloo, Ahmed ElGazzar, Marcel van Gerven

机构 * Department of Machine Learning and Neural Computing, Donders Institute for Brain, Cognition and Behaviour, Radboud University（机器学习与神经计算系，多纳德斯脑认知行为研究所，拉德堡德大学）

专题命中模型式强化学习：latent dynamics(abstract)；分类 cs.AI、cs.LG

AI总结本文提出基于潜在神经随机微分方程的生成建模框架，用于处理临床时间序列数据中的不规则采样和不确定性，通过模拟和真实数据验证，实现了比传统方法更准确的预测和不确定性估计。

2510.24757 2025-10-30 eess.SY cs.AI cs.LG cs.SY 56%

Stable-by-Design Neural Network-Based LPV State-Space Models for System Identification

Ahmet Eren Sertbaş, Tufan Kumbasar

机构 * Artificial Intelligence and Intelligent Systems Laboratory（人工智能与智能系统实验室）； Istanbul Technical University（伊斯坦布尔技术大学）

专题命中模型式强化学习：latent dynamics(abstract)；分类 cs.AI、cs.LG

Comments In the 12th International Conference of Image Processing, Wavelet and Applications on Real World Problems, 2025

2506.19885 2025-10-28 cs.LG cs.AI cs.SY eess.SY 56%

FlightKooba: A Fast Interpretable FTP Model

Jing Lu, Xuan Wu, Yizhun Tian, Songhan Fan, Yali Fang

机构 * School of Computer Science and Artificial Intelligence（计算机科学与人工智能学院）； Civil Aviation Flight University of China（中国民航飞行大学）； College of Computer Science（计算机科学学院）

专题命中模型式强化学习：latent dynamics(abstract)；分类 cs.AI、cs.LG

Comments Version 2: Major revision of the manuscript to refine the narrative, clarify the model's theoretical limitations and application scope, and improve overall presentation for journal submission

2510.03360 2025-10-07 cs.LG cs.AI math.OC physics.flu-dyn 56%

Physics-informed Neural-operator Predictive Control for Drag Reduction in Turbulent Flows

Zelin Zhao, Zongyi Li, Kimia Hassibi, Kamyar Azizzadenesheli, Junchi Yan, H. Jane Bae, Di Zhou, Anima Anandkumar

机构 * NVIDIA

专题命中模型式强化学习：model-based RL(abstract)；分类 cs.AI、cs.LG

2506.12996 2025-09-24 physics.comp-ph cs.LG 56%

Latent Representation Learning of Multi-scale Thermophysics: Application to Dynamics in Shocked Porous Energetic Material

Shahab Azarfar, Joseph B. Choi, Phong CH. Nguyen, Yen T. Nguyen, Pradeep Seshadri, H. S. Udaykumar, Stephen Baek

机构 * School of Data Science, University of Virginia, United States（数据科学学院，弗吉尼亚大学）； Department of Mechanical Engineering, University of Iowa, United States（机械工程系，爱荷华大学）； Department of Mechanical and Aerospace Engineering, University of Virginia, United States（机械与航空航天工程系，弗吉尼亚大学）

专题命中模型式强化学习：latent dynamics(abstract)；分类 cs.LG；dynamics model(abstract)

Comments 32 pages, 19 figures, complementary results added, restructured Introduction section

2509.05732 2025-09-09 cs.LG cs.AI 56%

Simulation Priors for Data-Efficient Deep Learning

Lenart Treven, Bhavya Sukhija, Jonas Rothfuss, Stelian Coros, Florian Dörfler, Andreas Krause

机构 * ETH Zürich（苏黎世联邦理工学院）

专题命中模型式强化学习：model-based reinforcement learning(abstract)；分类 cs.AI、cs.LG

2508.08144 2025-08-12 cs.RO cs.AI cs.SY eess.SY 56%

COMponent-Aware Pruning for Accelerated Control Tasks in Latent Space Models

Ganesh Sundaram, Jonas Ulmen, Amjad Haider, Daniel Görges

机构 * Department of Electrical and Computer Engineering, RPTU University Kaiserslautern-Landau（电子与计算机工程系，RPTU大学）

专题命中模型式强化学习：model-based reinforcement learning(abstract)；分类 cs.AI、cs.RO

Comments Submitted in: The 2026 IEEE/SICE International Symposium on System Integration (SII 2026)

2410.09486 2025-08-01 cs.LG cs.RO 56%

ActSafe: Active Exploration with Safety Constraints for Reinforcement Learning

Yarden As, Bhavya Sukhija, Lenart Treven, Carmelo Sferrazza, Stelian Coros, Andreas Krause

专题命中模型式强化学习：model-based RL(abstract)；分类 cs.LG、cs.RO

2507.06381 2025-07-10 cs.LG cs.AI math.DS q-bio.NC 56%

KPFlow: An Operator Perspective on Dynamic Collapse Under Gradient Descent Training of Recurrent Networks

James Hazelden, Laura Driscoll, Eli Shlizerman, Eric Shea-Brown

专题命中模型式强化学习：latent dynamics(abstract)；分类 cs.AI、cs.LG

2505.19002 2025-05-27 cs.LG cs.AI stat.ML 56%

Semi-pessimistic Reinforcement Learning

Jin Zhu, Xin Zhou, Jiaang Yao, Gholamali Aminian, Omar Rivasplata, Simon Little, Lexin Li, Chengchun Shi

机构 * London School of Economics and Political Science（伦敦经济政治学院）； University of California at Berkeley（加州大学伯克利分校）； The Alan Turing Institute（艾伦·图灵研究所）； University of Manchester（曼彻斯特大学）

专题命中模型式强化学习：model-based RL(abstract)；分类 cs.AI、cs.LG

2502.04593 2025-05-26 cs.LG cs.AI cs.NE stat.ML 56%

The Alpha-Alternator: Dynamic Adaptation To Varying Noise Levels In Sequences Using The Vendi Score For Improved Robustness and Performance

Mohammad Reza Rezaei, Adji Bousso Dieng

专题命中模型式强化学习：latent dynamics(abstract)；分类 cs.AI、cs.LG

Comments The codebase will be made available upon publication. This paper is dedicated to Patrice Lumumba

2504.21326 2025-05-01 cs.LG cs.AI 56%

Q-function Decomposition with Intervention Semantics with Factored Action Spaces

Junkyu Lee, Tian Gao, Elliot Nelson, Miao Liu, Debarun Bhattacharjya, Songtao Lu

机构 * IBM T. J. Watson Research Center（IBM 沃森研究中心）； The Chinese University of Hong Kong（香港中文大学）

专题命中模型式强化学习：model-based reinforcement learning(abstract)；分类 cs.AI、cs.LG

Comments AISTATS 2025

2503.19212 2025-03-26 cs.LG cs.AI cs.SY eess.SY 56%

Continual Reinforcement Learning for HVAC Systems Control: Integrating Hypernetworks and Transfer Learning

Gautham Udayakumar Bekal, Ahmed Ghareeb, Ashish Pujari

专题命中模型式强化学习：model-based reinforcement learning(abstract)；分类 cs.AI、cs.LG

2503.13842 2025-03-19 cs.LG cs.AI stat.ML 56%

Counterfactual experience augmented off-policy reinforcement learning

Sunbowen Lee, Yicheng Gong, Chao Deng

专题命中模型式强化学习：model-based reinforcement learning(abstract)；分类 cs.AI、cs.LG

Comments Accepted by Neurocomputing, https://doi.org/10.1016/j.neucom.2025.130017

2410.10253 2025-03-10 cs.LG cs.AI cs.NE 56%

Feedback Favors the Generalization of Neural ODEs

Jindou Jia, Zihan Yang, Meng Wang, Kexin Guo, Jianfei Yang, Xiang Yu, Lei Guo

专题命中模型式强化学习：latent dynamics(abstract)；分类 cs.AI、cs.LG

Comments 27 pages, 23 figures

2501.09781 2025-03-06 cs.CV 56%

VideoWorld: Exploring Knowledge Learning from Unlabeled Videos

Zhongwei Ren, Yunchao Wei, Xun Guo, Yao Zhao, Bingyi Kang, Jiashi Feng, Xiaojie Jin

专题命中模型式强化学习：latent dynamics(abstract)；分类 cs.CV；dynamics model(abstract)

Comments Code and models are released at: https://maverickren.github.io/VideoWorld.github.io/

2312.01544 2025-03-03 cs.LG cs.AI cs.SY eess.SY 56%

KEEC: Koopman Embedded Equivariant Control

Xiaoyuan Cheng, Yiming Yang, Xiaohang Tang, Wei Jiang, Yukun Hu

专题命中模型式强化学习：latent dynamics(abstract)；分类 cs.AI、cs.LG

2412.10400 2025-02-25 cs.CL cs.AI cs.LG 56%

Reinforcement Learning Enhanced LLMs: A Survey

Shuhe Wang, Shengyu Zhang, Jie Zhang, Runyi Hu, Xiaoya Li, Tianwei Zhang, Jiwei Li, Fei Wu, Guoyin Wang, Eduard Hovy

专题命中模型式强化学习：model-based RL(abstract)；分类 cs.AI、cs.LG

2502.12182 2025-02-19 physics.plasm-ph cs.AI cs.LG 56%

Towards Transparent and Accurate Plasma State Monitoring at JET

Andrin Bürli, Alessandro Pau, Thomas Koller, Olivier Sauter, JET Contributors

专题命中模型式强化学习：latent dynamics(abstract)；分类 cs.AI、cs.LG

2502.03550 2025-02-07 cs.LG cs.RO 56%

TD-M(PC)$^2$: Improving Temporal Difference MPC Through Policy Constraint

Haotian Lin, Pengcheng Wang, Jeff Schneider, Guanya Shi

专题命中模型式强化学习：model-based reinforcement learning(abstract)；分类 cs.LG、cs.RO

2405.12001 2025-02-04 cs.LG cs.AI 56%

Scrutinize What We Ignore: Reining In Task Representation Shift Of Context-Based Offline Meta Reinforcement Learning

Hai Zhang, Boyuan Zheng, Tianying Ji, Jinhang Liu, Anqi Guo, Junqiao Zhao, Lanqing Li

专题命中模型式强化学习：model-based RL(abstract)；分类 cs.AI、cs.LG

Comments Accept at ICLR 2025

2501.16142 2025-01-28 cs.LG cs.AI 56%

Towards General-Purpose Model-Free Reinforcement Learning

Scott Fujimoto, Pierluca D'Oro, Amy Zhang, Yuandong Tian, Michael Rabbat

专题命中模型式强化学习：model-based RL(abstract)；分类 cs.AI、cs.LG

Comments ICLR 2025

2501.15928 2025-01-28 cs.NI cs.AI cs.LG 56%

Generative AI for Lyapunov Optimization Theory in UAV-based Low-Altitude Economy Networking

Zhang Liu, Dusit Niyato, Jiacheng Wang, Geng Sun, Lianfen Huang, Zhibin Gao, Xianbin Wang

专题命中模型式强化学习：model-based reinforcement learning(abstract)；分类 cs.AI、cs.LG

Comments 8 pages, 5 figures, magazine paper

1. 模型式强化学习 1126 篇

Driving Reaction Trajectories via Latent Flow Matching

CMR: Contractive Mapping Embeddings for Robust Humanoid Locomotion on Unstructured Terrains

Generalized Information Gathering Under Dynamics Uncertainty

Scale-Consistent State-Space Dynamics via Fractal of Stationary Transformations

FM-EAC: Feature Model-based Enhanced Actor-Critic for Multi-Task Control in Dynamic Environments

PIMRL: Physics-Informed Multi-Scale Recurrent Learning for Burst-Sampled Spatiotemporal Dynamics

Adversarial Diffusion for Robust Reinforcement Learning

Generative Modeling of Clinical Time Series via Latent Stochastic Differential Equations

Stable-by-Design Neural Network-Based LPV State-Space Models for System Identification

FlightKooba: A Fast Interpretable FTP Model

Physics-informed Neural-operator Predictive Control for Drag Reduction in Turbulent Flows

Latent Representation Learning of Multi-scale Thermophysics: Application to Dynamics in Shocked Porous Energetic Material

Simulation Priors for Data-Efficient Deep Learning

COMponent-Aware Pruning for Accelerated Control Tasks in Latent Space Models

ActSafe: Active Exploration with Safety Constraints for Reinforcement Learning

KPFlow: An Operator Perspective on Dynamic Collapse Under Gradient Descent Training of Recurrent Networks

Semi-pessimistic Reinforcement Learning

The Alpha-Alternator: Dynamic Adaptation To Varying Noise Levels In Sequences Using The Vendi Score For Improved Robustness and Performance

Q-function Decomposition with Intervention Semantics with Factored Action Spaces

Continual Reinforcement Learning for HVAC Systems Control: Integrating Hypernetworks and Transfer Learning

Counterfactual experience augmented off-policy reinforcement learning

Feedback Favors the Generalization of Neural ODEs

VideoWorld: Exploring Knowledge Learning from Unlabeled Videos

KEEC: Koopman Embedded Equivariant Control

Reinforcement Learning Enhanced LLMs: A Survey

Towards Transparent and Accurate Plasma State Monitoring at JET

TD-M(PC)$^2$: Improving Temporal Difference MPC Through Policy Constraint

Scrutinize What We Ignore: Reining In Task Representation Shift Of Context-Based Offline Meta Reinforcement Learning

Towards General-Purpose Model-Free Reinforcement Learning

Generative AI for Lyapunov Optimization Theory in UAV-based Low-Altitude Economy Networking