Learning Invarting Representations for Reinforcement Learning without Reconstruction

ICML 2020