世界模型 - arXivDaily 专题

2505.24265 2026-05-01 cs.MA 58%

R3DM: Enabling Role Discovery and Diversity Through Dynamics Models in Multi-agent Reinforcement Learning

R3DM：通过动态模型在多智能体强化学习中实现角色发现与多样性

Harsh Goel, Mohammad Omama, Behdad Chalaki, Vaishnav Tadiparthi, Ehsan Moradi Pari, Sandeep Chinchali

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

AI总结 R3DM通过动态模型最大化智能体角色、观察轨迹与预期未来行为间的互信息，提升多智能体协作效率，实验表明其在SMAC和SMACv2环境中显著提升胜率。

Comments 21 pages, To appear in the International Conference of Machine Learning (ICML 2025)

2603.29315 2026-04-01 cs.RO cs.AI 58%

IMPASTO: Integrating Model-Based Planning with Learned Dynamics Models for Robotic Oil Painting Reproduction

IMPASTO：整合基于模型的规划与学习的动力学模型用于机器人油画复现

Yingke Wang, Hao Li, Yifeng Zhu, Hong-Xing Yu, Ken Goldberg, Li Fei-Fei, Jiajun Wu, Yunzhu Li, Ruohan Zhang

机构 * Stanford University（斯坦福大学）； The University of Texas at Austin（德克萨斯大学奥斯汀分校）； University of California, Berkeley（加州大学伯克利分校）； Columbia University（哥伦比亚大学）

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

AI总结 IMPASTO通过整合学习的动力学模型与基于模型的规划，实现机器人基于目标油画图像的复现，无需人类示范或精确模拟，提升复现精度。

2409.19647 2026-03-17 cs.RO cs.AI cs.SY eess.SY 58%

Fine-Tuning Hybrid Physics-Informed Neural Networks for Vehicle Dynamics Model Estimation

针对车辆动力学模型估计的混合物理信息神经网络微调

Shiming Fang, Kaiyan Yu

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

AI总结本文提出FTHD方法，结合监督和非监督PINNs，通过微调预训练的DDM，在减少数据量的情况下提升参数估计精度，并通过EKF-FTHD增强数据鲁棒性。

Journal ref Int. J. Intell. Robot. Appl., vol. 9, no. 4, pp. 1594-1610, 2025

2502.08658 2026-01-06 cs.RO cs.AI 58%

Knowledge-data fusion dominated vehicle platoon dynamics modeling and analysis: A physics-encoded deep learning approach

基于知识-数据融合的车辆编队动力学建模与分析：一种物理编码深度学习方法

Hao Lyu, Yanyong Guo, Pan Liu, Shuo Feng, Weilin Ren, Quansheng Yue

机构 * School of Transportation, Southeast University, Nanjing, China, 211189（东南大学交通学院）； Jiangsu Key Laboratory of Urban ITS, Nanjing, China, 210096（江苏省城市智能交通重点实验室）； Jiangsu Collaborative Innovation Center of Modern Urban Traffic Technologies, Nanjing, China, 210096（江苏省现代城市交通技术协同创新中心）； Department of Automation, Beijing National Research Center for Information Science and Technology (BNRist), Tsinghua University, Beijing, China, 100084（自动化系，北京信息科学与技术国家研究中心（BNRist），清华大学）

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

AI总结本文提出PeMTFLN模型，通过物理编码深度学习方法实现车辆编队动力学建模与分析，提升预测精度和稳定性。

Journal ref Information Fusion, Vol. 126, 103622, 2026

2511.21846 2025-12-01 eess.SY cs.AI cs.LG cs.SY 58%

LILAD: Learning In-context Lyapunov-stable Adaptive Dynamics Models

LILAD: 学习上下文Lyapunov稳定自适应动力学模型

Amit Jena, Na Li, Le Xie

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

AI总结 LILAD通过上下文学习同时学习动态模型和Lyapunov函数，确保系统在分布偏移和任务外情况下的稳定性与适应性。

Comments This article has been accepted for AAAI-26 (The 40th Annual AAAI Conference on Artificial Intelligence)

2510.08556 2025-10-10 cs.RO cs.CV 58%

DexNDM: Closing the Reality Gap for Dexterous In-Hand Rotation via Joint-Wise Neural Dynamics Model

Xueyi Liu, He Wang, Li Yi

机构 * Tsinghua University（清华大学）； Peking University（北京大学）； Shanghai Qi Zhi Institute（上海启智研究所）； Galbot Project（Galbot项目）

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

Comments Project Website: https://meowuu7.github.io/DexNDM/ Video: https://youtu.be/tU2Mv8vWftU

2509.23307 2025-09-30 cs.LG cs.AI 58%

A Neural ODE Approach to Aircraft Flight Dynamics Modelling

Gabriel Jarry, Ramon Dalmau, Xavier Olive, Philippe Very

机构 * EUROCONTROL Aviation Sustainability Unit (ASU)（EUROCONTROL航空可持续性单位）； ONERA – DTIS Université de Toulouse（ONERA-DTIS图卢兹大学）

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

2409.18768 2025-02-12 cs.AI cs.LG cs.RO cs.SY eess.SY 58%

Learning from Demonstration with Implicit Nonlinear Dynamics Models

Peter David Fagan, Subramanian Ramamoorthy

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

Comments 21 pages, 9 figures

2102.11394 2024-10-08 cs.LG cs.RO cs.SY eess.SY stat.ML 58%

Explore the Context: Optimal Data Collection for Context-Conditional Dynamics Models

Jan Achterhold, Joerg Stueckler

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

Comments Accepted for publication at the 24th International Conference on Artificial Intelligence and Statistics (AISTATS) 2021, with supplementary material. Corrected version (see footnote on p. 6)

2409.02390 2024-09-05 cs.NE cs.AI cs.CV q-bio.NC 58%

Neural Dynamics Model of Visual Decision-Making: Learning from Human Experts

Jie Su, Fang Cai, Shu-Kuo Zhao, Xin-Yi Wang, Tian-Yi Qian, Da-Hui Wang, Bo Hong

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

2309.11148 2024-05-29 cs.RO cs.CV 58%

Online Calibration of a Single-Track Ground Vehicle Dynamics Model by Tight Fusion with Visual-Inertial Odometry

Haolong Li, Joerg Stueckler

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

Comments Accepted for publication in IEEE International Conference on Robotics and Automation (ICRA), 2024

2306.11941 2024-05-14 cs.LG cs.AI 58%

Efficient Dynamics Modeling in Interactive Environments with Koopman Theory

Arnab Kumar Mondal, Siba Smarak Panigrahi, Sai Rajeswar, Kaleem Siddiqi, Siamak Ravanbakhsh

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

Comments Accepted to ICLR 2024 and EWRL 2023

2403.13850 2024-03-22 cs.LG cs.AI physics.flu-dyn 58%

Spatio-Temporal Fluid Dynamics Modeling via Physical-Awareness and Parameter Diffusion Guidance

Hao Wu, Fan Xu, Yifan Duan, Ziwei Niu, Weiyan Wang, Gaofeng Lu, Kun Wang, Yuxuan Liang, Yang Wang

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

2305.10912 2023-09-26 cs.AI cs.RO 58%

A Generalist Dynamics Model for Control

Ingmar Schubert, Jingwei Zhang, Jake Bruce, Sarah Bechtle, Emilio Parisotto, Martin Riedmiller, Jost Tobias Springenberg, Arunkumar Byravan, Leonard Hasenclever, Nicolas Heess

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

2110.14700 2023-08-11 eess.SY cs.SY 58%

DDK: A Deep Koopman Approach for Dynamics Modeling and Trajectory Tracking of Autonomous Vehicles

Yongqian Xiao

专题命中模型式强化学习：latent dynamics(abstract)；dynamics model(title)

Journal ref 2023 IEEE International Conference on Robotics and Automation (ICRA)

2306.03405 2023-06-07 cs.RO cs.LG 58%

Vehicle Dynamics Modeling for Autonomous Racing Using Gaussian Processes

Jingyun Ning, Madhur Behl

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

Comments 12 pages, 6 figures, 10 tables

2305.12369 2023-05-23 cs.CV cs.AI cs.LG 58%

HIINT: Historical, Intra- and Inter- personal Dynamics Modeling with Cross-person Memory Transformer

Yubin Kim, Dong Won Lee, Paul Pu Liang, Sharifa Algohwinem, Cynthia Breazeal, Hae Won Park

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

2303.11756 2023-03-22 cs.RO cs.LG 58%

Improving Deep Dynamics Models for Autonomous Vehicles with Multimodal Latent Mapping of Surfaces

Johan Vertens, Nicolai Dorka, Tim Welschehold, Michael Thompson, Wolfram Burgard

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

2110.01894 2023-03-20 cs.LG cs.RO 58%

Combining Physics and Deep Learning to learn Continuous-Time Dynamics Models

Michael Lutter, Jan Peters

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

2202.01889 2022-06-27 cs.LG cs.AI stat.ML 58%

Generalizing to New Physical Systems via Context-Informed Dynamics Model

Matthieu Kirchmeyer, Yuan Yin, Jérémie Donà, Nicolas Baskiotis, Alain Rakotomamonjy, Patrick Gallinari

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

Comments Accepted at ICML 2022

2205.13804 2022-05-30 cs.RO cs.LG 58%

End-to-End Learning of Hybrid Inverse Dynamics Models for Precise and Compliant Impedance Control

Moritz Reuss, Niels van Duijkeren, Robert Krug, Philipp Becker, Vaisakh Shaj, Gerhard Neumann

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

Comments Accepted for publication at Robotics: Science and System XVIII (RSS), year 2022. Paper length is 13 pages (i.e. 9 pages of technical content, 1 page of the Bibliography/References and 3 pages of Appendix)

2104.13877 2021-04-29 cs.LG cs.AI stat.ML 58%

Autoregressive Dynamics Models for Offline Policy Evaluation and Optimization

Michael R. Zhang, Tom Le Paine, Ofir Nachum, Cosmin Paduraru, George Tucker, Ziyu Wang, Mohammad Norouzi

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

Comments ICLR 2021. 17 pages

2011.08750 2020-11-18 cs.RO cs.LG cs.SY eess.SY 58%

Iterative Semi-parametric Dynamics Model Learning For Autonomous Racing

Ignat Georgiev, Christoforos Chatzikomis, Timo Völkl, Joshua Smith, Michael Mistry

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

Comments Accepted at 4th Conference on Robot Learning (CoRL 2020)

2008.00456 2020-08-04 cs.RO cs.CV cs.LG 58%

Hindsight for Foresight: Unsupervised Structured Dynamics Models from Physical Interaction

Iman Nematollahi, Oier Mees, Lukas Hermann, Wolfram Burgard

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

Comments Accepted at the 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

1906.07372 2020-07-23 cs.LG cs.RO stat.ML 58%

RIDM: Reinforced Inverse Dynamics Modeling for Learning from a Single Observed Demonstration

Brahma S. Pavse, Faraz Torabi, Josiah P. Hanna, Garrett Warnell, Peter Stone

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

Comments IEEE Robotics and Automation Letters, presented at International Conference on Intelligent Robots and Systems (IROS 2020)

1806.10019 2020-03-18 cs.LG cs.AI stat.ML 58%

Adversarial Active Exploration for Inverse Dynamics Model Learning

Zhang-Wei Hong, Tsu-Jui Fu, Tzu-Yun Shann, Yi-Hsiang Chang, Chun-Yi Lee

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

Comments Published as a conference paper at CoRL 2019

1909.03749 2019-10-24 cs.LG cs.CV eess.IV stat.ML 58%

Learning Visual Dynamics Models of Rigid Objects using Relational Inductive Biases

Fabio Ferreira, Lin Shao, Tamim Asfour, Jeannette Bohg

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

Comments short paper (4 pages, two figures), accepted to NeurIPS 2019 Graph Representation Learning workshop

1711.05253 2018-04-02 cs.RO cs.LG 58%

Learning Image-Conditioned Dynamics Models for Control of Under-actuated Legged Millirobots

Anusha Nagabandi, Guangzhao Yang, Thomas Asmar, Ravi Pandya, Gregory Kahn, Sergey Levine, Ronald S. Fearing

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

2604.26836 2026-08-11 cs.LG cs.SY eess.SY 版本更新 56%

Uncertainty-Aware Predictive Safety Filters for Probabilistic Neural Network Dynamics

具有不确定性的预测安全过滤器用于概率神经网络动态

Bernd Frauenknecht, Lukas Kesper, Daniel Mayfrank, Henrik Hose, Sebastian Trimpe

机构 * Institute for Data Science in Mechanical Engineering (DSME), RWTH Aachen University（机械工程数据科学研究所（DSME），亚琛工业大学）； Institute of Climate and Energy Systems (ICE), Energy Systems Engineering (ICE-1), Forschungszentrum Jülich GmbH（气候与能源系统研究所（ICE），能源系统工程（ICE-1），焦耳研究中心有限公司）

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

AI总结本文提出了一种具有不确定性的预测安全过滤器（UPSi），通过将未来结果建模为可达集，利用概率集合（PE）神经网络动态模型提供严格的安全预测，从而在模型基于强化学习（MBRL）中提升探索安全性，同时保持与标准MBRL相当的性能。

Comments Reinforcement Learning Journal, 2026. Reinforcement Learning Conference (RLC), Montréal, August 2026

2608.06595 2026-08-10 cs.LG cs.AI 新提交 56%

Flowing Through States: Neural ODE Regularization for Reinforcement Learning

状态间的流动：用于强化学习的神经常微分方程正则化方法

Mohamed Ghanem, Bernd Finkbeiner

机构 * CISPA Helmholtz Center for Information Security（CISPA亥姆霍兹信息安全中心）； Technical University of Munich（慕尼黑工业大学）

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

AI总结该研究提出一种基于神经ODE的正则化方法，将其集成到Actor-Critic算法中，在A2C的Atari基准和PPO的网格世界环境中显著提升了强化学习智能体的性能。

1. 模型式强化学习 1126 篇

R3DM: Enabling Role Discovery and Diversity Through Dynamics Models in Multi-agent Reinforcement Learning

IMPASTO: Integrating Model-Based Planning with Learned Dynamics Models for Robotic Oil Painting Reproduction

Fine-Tuning Hybrid Physics-Informed Neural Networks for Vehicle Dynamics Model Estimation

Knowledge-data fusion dominated vehicle platoon dynamics modeling and analysis: A physics-encoded deep learning approach

LILAD: Learning In-context Lyapunov-stable Adaptive Dynamics Models

DexNDM: Closing the Reality Gap for Dexterous In-Hand Rotation via Joint-Wise Neural Dynamics Model

A Neural ODE Approach to Aircraft Flight Dynamics Modelling

Learning from Demonstration with Implicit Nonlinear Dynamics Models

Explore the Context: Optimal Data Collection for Context-Conditional Dynamics Models

Neural Dynamics Model of Visual Decision-Making: Learning from Human Experts

Online Calibration of a Single-Track Ground Vehicle Dynamics Model by Tight Fusion with Visual-Inertial Odometry

Efficient Dynamics Modeling in Interactive Environments with Koopman Theory

Spatio-Temporal Fluid Dynamics Modeling via Physical-Awareness and Parameter Diffusion Guidance

A Generalist Dynamics Model for Control

DDK: A Deep Koopman Approach for Dynamics Modeling and Trajectory Tracking of Autonomous Vehicles

Vehicle Dynamics Modeling for Autonomous Racing Using Gaussian Processes

HIINT: Historical, Intra- and Inter- personal Dynamics Modeling with Cross-person Memory Transformer

Improving Deep Dynamics Models for Autonomous Vehicles with Multimodal Latent Mapping of Surfaces

Combining Physics and Deep Learning to learn Continuous-Time Dynamics Models

Generalizing to New Physical Systems via Context-Informed Dynamics Model

End-to-End Learning of Hybrid Inverse Dynamics Models for Precise and Compliant Impedance Control

Autoregressive Dynamics Models for Offline Policy Evaluation and Optimization

Iterative Semi-parametric Dynamics Model Learning For Autonomous Racing

Hindsight for Foresight: Unsupervised Structured Dynamics Models from Physical Interaction

RIDM: Reinforced Inverse Dynamics Modeling for Learning from a Single Observed Demonstration

Adversarial Active Exploration for Inverse Dynamics Model Learning

Learning Visual Dynamics Models of Rigid Objects using Relational Inductive Biases

Learning Image-Conditioned Dynamics Models for Control of Under-actuated Legged Millirobots

Uncertainty-Aware Predictive Safety Filters for Probabilistic Neural Network Dynamics

Flowing Through States: Neural ODE Regularization for Reinforcement Learning