Hi, There

一份简要的简历

LinkedIn | GitHub | dockers | Mail me

Ci Song / 宋辞

Well, it is now more obvious that

• the most difficult part of RL is to design the components, somehow, the algorithm part is not that important as I used to think of

• through the work on the rubik's cube applying RL, I found that Q-learning is absolutely the first option you should try when the situation you are dealing with is limited. Q-learning is easy, fast, and reasonable for explanation