Learning Representations in Reinforcement Learning: an Information Bottleneck Approach

Yingjun Pei, Xinwen Hou