Improved Corruption Robust Algorithms for Episodic Reinforcement Learning

                   Yifang Chen 1 Simon S. Du 1 Kevin Jamieson 1

              Abstract              stage according to the underlying transition function.
   We study episodic reinforcement learning under      The majority of the literature in learning in MDPs studies
   unknown adversarial corruptions in both the re-      stationary environments, where the underlying unknown
   wards and the transition probabilities of  ...