Playing Atari Games with Deep Reinforcement Learning [7] - Double DQN, Prioritised Experience Replay