I hope DeepMind publish similar per-move value network assessments for each of the games (and even better would be policy network "heat maps")
James On Mon, Mar 14, 2016 at 6:05 PM, Paweł Morawiecki < pawel.morawie...@gmail.com> wrote: > Hi, > > It looks like this http://zhuanlan.zhihu.com/yuandong/20639694 > is an analysis of Facebook AI of Game 1 and 2. > > I just looked at it very briefly but, interestingly, in Game 2 Darkforest > agrees with pro commentators who said that Lee Sedol #70 (S15) was slow and > it would be better to chase black's group in the centre. Indeed Darkforest > suggested F10. > > Details:http://www.yuandong-tian.com/win_rate2.txt > > Regards, > Pawel > > _______________________________________________ > Computer-go mailing list > Computer-go@computer-go.org > http://computer-go.org/mailman/listinfo/computer-go >
_______________________________________________ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go