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 ...


雷达卡




京公网安备 11010802022788号







