Justin Fu, Aviral Kumar, Matthew Soh +1
Q-learning methods represent a commonly used class of algorithms in reinforcement learning: they are generally efficient and simple, and can be combined readily with function appro…