Fetching the paper…

Kinematic State Abstraction and Provably Efficient Rich-Observation Reinforcement Learning · Around