Dual Preference Learning for Multi-Agent Reinforcement Learning
TL;DR
Dual Preference Learning addresses reward function challenges in multi-agent reinforcement learning by learning from preferences instead of manual engineering, overcoming issues with pairwise trajectory comparisons.
Dual Preference Learning for Multi-Agent Reinforcement Learning
Sehyeok Kang; Minu Kim; Jihwan Oh; Se-Young Yun
https://doi.org/10.1109/ACCESS.2025.3645778
Volume 14
Designing effective reward functions is fundamental challenging in reinforcement learning, especially in complex multi-agent systems with intricate credit assignment. Preference-based reinforcement learning (PbRL) offers an alternative to manual reward engineering by learning from preferences. However, the prevalent approach in PbRL, which involves pairwise trajectory comparisons, encounters diffi...