https://drive.google.com/file/d/0BxXI_RttTZAhTUpqUFdEZ3BXNFE/view
game of Pong is a MDP.
终于一睹AK真容了，很有想法，很幽默
http://karpathy.github.io/
Deep RL Bootcamp Lecture 4B Policy Gradients Revisited
原文：https://www.cnblogs.com/ecoflex/p/8976042.html