How an A.I. ‘Cat-And-Mouse Game’ Generates Believable Fake Photos

Re:How an A.I. ‘Cat-And-Mouse Game’ Generates Believable Fak... by gwern on Thursday January 04, 2018 @11:46PM (#16076625)

If you're using DQN or pretty much anything in DRL, you don't have any guarantees about convergence in the first place, and using a RNN does give you the history summary you need (at least up to the minimum error achievable with that fixed-length summary, not that that is any more likely to converge than the overall DRL algo is).

[ Parent | on HN ]

Re:How an A.I. ‘Cat-And-Mouse Game’ Generates Believable Fak... by bitL on Friday January 05, 2018 @10:26AM (#16079125)

I meant that under Markovian assumption value iteration used for Bellman equation is guaranteed to converge. So it makes math people happy, even if such property doesn't hold in the real world nor in the problem they try to solve, and the "deep" in DRL is just heuristics, though surprisingly working in many cases.

[ Parent | on HN ]