Mastering Atari Games with Limited Data

Weirui Ye, Shaohuai Liu, Thanard Kurutach +2

Reinforcement learning has achieved great success in many applications. However, sample efficiency remains a key challenge, with prominent methods requiring millions (or even billi…