2020-02-17 paper笔记 Reinforcement Learning with Competitive Ensembles of Information-Constrained Primitives