MetaCURE: Meta Reinforcement Learning with
               Empowerment-Driven Exploration

  Jin Zhang * 1 Jianhao Wang * 1 Hao Hu 1 Tong Chen 1 Yingfeng Chen 2 Changjie Fan 2 Chongjie Zhang 1

             Abstract                with sparse rewards remains challenging, as task-relevant in-
                                 formation is scarce in such settings, and efficient exploration
   Meta reinforcement learning (meta-RL) extracts
                                 is required to quickly find  ...