Multi-Agent Reinforcement Learning from Human Feedback: Data Coverage and Algorithmic Techniques
Lists
Only you see this
Abstract
We initiate the study of Multi-Agent Reinforcement Learning from Human Feedback (MARLHF), exploring both theoretical foundations and empirical validations. We define the task as identifying Nash equilibrium from a preference-only offline dataset in general-sum games, a problem marked by the challenge of sparse feedback signals. Our theory establishes the upper complexity bounds for Nash Equilibrium in effective MARLHF, demonstrating that single-policy coverage is inadequate and highlighting the importance of unilateral dataset coverage. These theoretical insights are verified through comprehensive experiments. To enhance the practical performance, we further introduce two algorithmic techniques. (1) We propose a Mean Squared Error (MSE) regularization along the time axis to achieve a more uniform reward distribution and improve reward learning outcomes. (2) We propose an extra penalty based on dataset distribution to incorporate pessimism, enhancing stability and effectiveness during training. Our findings underscore the multifaceted approach required for MARLHF, paving the way for effective preference-based multi-agent systems.
BibTeX
@misc{zhang2025multi,
author = {Natalia Zhang and Xinqi Wang and Qiwen Cui and Runlong Zhou and Sham M. Kakade and Simon Shaolei Du},
title = {{Multi-Agent Reinforcement Learning from Human Feedback: Data Coverage and Algorithmic Techniques}},
year = {2025},
url = {https://openreview.net/forum?id=4vPC6Aj6N7},
howpublished = {OpenReview},
note = {Submitted to ICLR 2025},
}