1 paper
Ting Xiao, Jiakun Zheng, Rushuai Yang +4
Unsupervised Reinforcement Learning (RL) aims to discover diverse behaviors that can accelerate the learning of downstream tasks. Previous methods typically focus on entropy-based…