Score: 0

Video-Enhanced Offline Reinforcement Learning: A Model-Based Approach

Published: May 10, 2025 | arXiv ID: 2505.06482v2

By: Minting Pan , Yitao Zheng , Jiajian Li and more

Potential Business Impact:

Teaches robots to learn from watching videos.

Business Areas:
Image Recognition Data and Analytics, Software

Offline reinforcement learning (RL) enables policy optimization using static datasets, avoiding the risks and costs of extensive real-world exploration. However, it struggles with suboptimal offline behaviors and inaccurate value estimation due to the lack of environmental interaction. We present Video-Enhanced Offline RL (VeoRL), a model-based method that constructs an interactive world model from diverse, unlabeled video data readily available online. Leveraging model-based behavior guidance, our approach transfers commonsense knowledge of control policy and physical dynamics from natural videos to the RL agent within the target domain. VeoRL achieves substantial performance gains (over 100% in some cases) across visual control tasks in robotic manipulation, autonomous driving, and open-world video games.

Country of Origin
🇨🇳 China

Page Count
17 pages

Category
Computer Science:
Machine Learning (CS)