You are viewing a single comment's thread from:
RE: Reinforcement learning and Atari games - addition to ML series part 4
Mainly in the "easy tasks" - where reward comes quite soon after an action - the RL agents are really better than human players.