世界模型

Mind the Model, Not the Agent: The Primacy Bias in Model-based RL

Zhongjian Qiao, Jiafei Lyu, Xiu Li

专题命中 模型式强化学习 ：model-based RL(title)；world model(abstract)；world model(abstract)；model-based reinforcement learning(abstract)

Comments Accepted by European Conference on Artificial Intelligence (ECAI) 2024

Prototypical context-aware dynamics generalization for high-dimensional model-based reinforcement learning

Junjie Wang, Yao Mu, Dong Li, Qichao Zhang, Dongbin Zhao, Yuzheng Zhuang, Ping Luo, Bin Wang, Jianye Hao

专题命中 模型式强化学习 ：model-based reinforcement learning(title)；world model(abstract)；world model(abstract)；model-based RL(abstract)

Scalable Multi-Agent Model-Based Reinforcement Learning

Vladimir Egorov, Aleksei Shpilman

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

Comments AAMAS'2022, cite https://dl.acm.org/doi/abs/10.5555/3535850.3535894

DreamerPro: Reconstruction-Free Model-Based Reinforcement Learning with Prototypical Representations

Fei Deng, Ingook Jang, Sungjin Ahn

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

Fractional Transfer Learning for Deep Model-Based Reinforcement Learning

Remo Sasso, Matthia Sabatelli, Marco A. Wiering

专题命中 模型式强化学习 ：model-based reinforcement learning(title)；world model(abstract)；world model(abstract)；model-based RL(abstract)

Comments 21 pages, 8 figures, 7 tables

Unlocking Temporal Generalization in Hamiltonian Video Dynamics Models

解锁哈密顿视频动力学模型中的时间泛化

Eli Laird, Corey Clark

机构 * Department of Computer Science, Southern Methodist University（南卫理公会大学计算机科学系） 纠错

专题命中 模型式强化学习 ：world model(abstract)；world models(abstract)；world model(abstract)；world models(abstract)

AI总结 研究世界模型在可变时间分辨率下预测动力学的问题，利用哈密顿生成网络（HGN），指出其在非保守环境中时间泛化失效的问题及原因，通过针对性修复实现稳定动力学预测，推荐连续时间视频生成中时间泛化的策略。

Comments To appear in the 1st Workshop on Physics-Aware Video Generation and Restoration at the 28th International Conference on Pattern Recognition

DynaWeb: Model-Based Reinforcement Learning of Web Agents

DynaWeb：基于模型的Web代理强化学习

Hang Ding, Peidong Liu, Junqiao Wang, Ziwei Ji, Meng Cao, Rongzhao Zhang, Lynn Ai, Eric Yang, Tianyu Shi, Lei Yu

机构 * Shanghai Jiao Tong University（上海交通大学） ； Sichuan University（四川大学） ； Hong Kong University of Science and Technology（香港科学与技术大学） ； McGill University（麦吉尔大学） ； Shanghai AI Lab（上海人工智能实验室） ； Gradient ； University of Toronto（多伦多大学） ； Mila - Quebec AI Institute（魁北克AI研究所） 纠错

专题命中 模型式强化学习 ：model-based reinforcement learning(title,abstract)；world model(abstract)；world model(abstract)；分类 cs.AI

AI总结 本文提出DynaWeb框架，通过模拟环境提升Web代理的强化学习效率，实验表明其在WebArena和WebVoyager基准上显著提升性能。

SkyDreamer: Interpretable End-to-End Vision-Based Drone Racing with Model-Based Reinforcement Learning

Aderik Verraest, Stavrow Bahnam, Robin Ferede, Guido de Croon, Christophe De Wagter

机构 * Faculty of Aerospace Engineering, Delft University of Technology（航空航天工程学院，代尔夫特理工大学） 纠错

专题命中 模型式强化学习 ：model-based reinforcement learning(title,abstract)；world model(abstract)；world model(abstract)；分类 cs.RO

Informed POMDP: Leveraging Additional Information in Model-Based RL

Gaspard Lambrechts, Adrien Bolland, Damien Ernst

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

Comments In Reinforcement Learning Conference, 2024. 10 pages, 22 pages total, 10 figures

Journal ref Reinforcement Learning Journal, 2024

INFOrmation Prioritization through EmPOWERment in Visual Model-Based RL

Homanga Bharadhwaj, Mohammad Babaeizadeh, Dumitru Erhan, Sergey Levine

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

Comments Published in International Conference on Learning Representations (ICLR 2022)

Biased Dreams: Limitations to Epistemic Uncertainty Quantification in Latent Dynamics Models

有偏的梦境：潜在空间模型中知识不确定性量化限制

Julia Berger, Bernd Frauenknecht, Sebastian Trimpe, Bastian Leibe

专题命中 模型式强化学习 ：latent dynamics(title,abstract)；dynamics model(title,abstract)；model-based reinforcement learning(abstract)；分类 cs.LG

AI总结 研究探讨了潜在空间模型中知识不确定性量化存在的偏差问题，指出潜在过渡偏向于高代表性的区域，导致环境动态差异在潜在空间中不显现，影响不确定性估计的可靠性。

Comments Reinforcement Learning Conference (RLC) 2026

Imagined Value Gradients: Model-Based Policy Optimization with Transferable Latent Dynamics Models

Arunkumar Byravan, Jost Tobias Springenberg, Abbas Abdolmaleki, Roland Hafner, Michael Neunert, Thomas Lampe, Noah Siegel, Nicolas Heess, Martin Riedmiller

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

Comments To appear at the 3rd annual Conference on Robot Learning, Osaka, Japan (CoRL 2019). 24 pages including appendix (main paper - 8 pages)

Multi-Agent Model-Based Reinforcement Learning with Joint State-Action Learned Embeddings

基于联合状态-动作学习嵌入的多智能体模型驱动强化学习

Zhizun Wang, David Meger

机构 * McGill University（麦吉尔大学） 纠错

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

AI总结 本文提出了一种基于联合状态-动作学习嵌入的多智能体模型驱动强化学习框架，通过统一表示学习与想象式回放，提升智能体在动态环境中协调能力与长期规划效率。

Comments 22 pages

Robust Model-Based Reinforcement Learning with an Adversarial Auxiliary Model

Siemen Herremans, Ali Anwar, Siegfried Mercelis

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

Comments Will be presented at the RL Safety Workshop at RLC 2024

Predictable MDP Abstraction for Unsupervised Model-Based RL

Seohong Park, Sergey Levine

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

Comments ICML 2023

Intrinsic Motivation in Model-based Reinforcement Learning: A Brief Review

Artem Latyshev, Aleksandr I. Panov

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

Comments 13 pages, 7 figures

Skill-based Model-based Reinforcement Learning

Lucy Xiaoyang Shi, Joseph J. Lim, Youngwoon Lee

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

Comments Published at the Conference on Robot Learning (CoRL) 2022. Website: https://clvrai.com/skimo

Model-based Reinforcement Learning for Decentralized Multiagent Rendezvous

Rose E. Wang, J. Chase Kew, Dennis Lee, Tsang-Wei Edward Lee, Tingnan Zhang, Brian Ichter, Jie Tan, Aleksandra Faust

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

Comments CoRL 2020. The video is available at: https://youtu.be/-ydXHUtPzWE

Dynamic Horizon Value Estimation for Model-based Reinforcement Learning

Junjie Wang, Qichao Zhang, Dongbin Zhao, Mengchen Zhao, Jianye Hao

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

Comments 9 pages, 6 figures

Climate-Dyna Deep Hedging for XVAs: Model-Based Reinforcement Learning, Residual Climate HVA, and Hedge-Instrument Discovery

面向XVAs的Climate-Dyna深度对冲：基于模型的强化学习、剩余气候HVA及对冲工具发现

Xiaozhen Wang, Francois Buet-Golfouse

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

AI总结 该研究提出Climate-Dyna深度对冲方法，通过基于模型的强化学习解决剩余气候HVA问题，在EU ETS半合成研究中有效降低了气候费用，显著减少了遗憾值并保留了大部分精确辅助增益。

Comments 15 pages, 2 figures, 1 table

Scalable Offline Model-Based RL with Action Chunks

可扩展的基于模型的强化学习与动作块

Kwanyoung Park, Seohong Park, Youngwoon Lee, Sergey Levine

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

AI总结 本文提出MAC方法，通过动作块模型和拒绝采样提升离线基于模型的强化学习在复杂长 horizon 任务中的性能。

Comments 22 pages, 7 figures

Mismatched No More: Joint Model-Policy Optimization for Model-Based RL

Benjamin Eysenbach, Alexander Khazatsky, Sergey Levine, Ruslan Salakhutdinov

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

Comments NeurIPS 2022

A Relational Intervention Approach for Unsupervised Dynamics Generalization in Model-Based Reinforcement Learning

Jixian Guo, Mingming Gong, Dacheng Tao

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

Comments ICLR2022 accepted paper

Model-based Reinforcement Learning: A Survey

Thomas M. Moerland, Joost Broekens, Aske Plaat, Catholijn M. Jonker

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

Model-Based Reinforcement Learning via Latent-Space Collocation

Oleh Rybkin, Chuning Zhu, Anusha Nagabandi, Kostas Daniilidis, Igor Mordatch, Sergey Levine

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

Comments International Conference on Machine Learning (ICML), 2021. Videos and code at https://orybkin.github.io/latco/

Benchmarking Model-Based Reinforcement Learning

Tingwu Wang, Xuchan Bao, Ignasi Clavera, Jerrick Hoang, Yeming Wen, Eric Langlois, Shunshi Zhang, Guodong Zhang, Pieter Abbeel, Jimmy Ba

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

Comments 8 main pages, 8 figures; 14 appendix pages, 25 figures

Driving Beyond Privilege: Distilling Dense-Reward Knowledge into Sparse-Reward Policies

超越特权：将密集奖励知识蒸馏到稀疏奖励策略中

Feeza Khan Khanzada, Jaerock Kwon

机构 * Department of Electrical and Computer Engineering, University of Michigan-Dearborn（电气与计算机工程系，密歇根大学迪尔伯恩分校） 纠错

专题命中 模型式强化学习 ：world model(abstract)；world model(abstract)；latent dynamics(abstract)；model-based reinforcement learning(abstract)

AI总结 本文提出通过密集奖励蒸馏学习稀疏奖励策略，提升自动驾驶在稀疏目标上的泛化能力。

The Surprising Difficulty of Search in Model-Based Reinforcement Learning

基于模型的强化学习中搜索的惊人困难

Wei-Di Chang, Mikael Henaff, Brandon Amos, Gregory Dudek, Scott Fujimoto

机构 * Meta FAIR ； McGill University（麦吉尔大学） 纠错

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

AI总结 本文研究基于模型的强化学习中的搜索问题，发现即使模型高度准确，搜索也可能损害性能，而缓解过估计偏差比提高模型或价值函数精度更重要，通过取价值函数集成的最小值可有效解决偏差并实现有效搜索。

Comments ICML 2026

Learning to Crawl: Latent Model-Based Reinforcement Learning for Soft Robotic Adaptive Locomotion

Vaughn Gzenda, Robin Chhabra

机构 * Embodied Learning and Intelligence for eXploration and Innovative soft Robotics (ELIXIR) Lab（具身学习与智能探索与创新软机器人实验室） 纠错

专题命中 模型式强化学习 ：model-based reinforcement learning(title,abstract)；latent dynamics(abstract)；分类 cs.RO；predictive model(abstract)

Model-Based Reinforcement Learning for Control of Strongly-Disturbed Unsteady Aerodynamic Flows

Zhecheng Liu, Diederik Beckers, Jeff D. Eldredge

机构 * Department of Mechanical and Aerospace Engineering（机械与航空航天工程系） ； Graduate Aerospace Laboratories（航空航天研究生实验室） 纠错

专题命中 模型式强化学习 ：model-based reinforcement learning(title,abstract)；latent dynamics(abstract)；分类 cs.LG；dynamics model(abstract)

Journal ref AIAA Journal, (Early Access) 2025