A3C

Deep Reinforcement Learning (John Schulman, OpenAI)

Lex Fridman

An asynchronous implementation of policy gradient methods that achieves very good results.