This was a horribly written paper. It includes all the popular buzzwords but the authors ability to convey the meaning of these concepts was utterly lacking. Don't expect to learn anything at all from this one. OP's article is far better than this.
I am the author of Deep Reinforcement Learning: An Overview.
I fully understand my overview is far from perfect,
and I warmly welcome (constructive and responsible) comments and criticisms.
My overview was published on arXiv on Jan 25, 2017, with major updates on July 15, 2017, much earlier than the brief survey of DRL, submitted to arXiv on Aug 19 2017 (manuscript to the special issue due on March 1, 2017).
The brief survey did not cite my overview.
There were some discussions about it on Twitter a couple days ago when the brief survey appeared online. I thought it was friendly handled. I realized that it was not, when I saw this comment.
As a result, 1) I registered an account on YCombinator, with my real name, and reply now; 2) in the meanwhile, I sent a formal request to the authors and the editors of the brief survey to cite my overview, and make comments on it, positively or negatively.
