[Computer-go] 25x25 experiment

Aja Huang ajahuang at google.com
Wed Apr 29 08:57:35 PDT 2015


On Wed, Apr 29, 2015 at 4:35 PM, Detlef Schmicker <ds2 at physik.de> wrote:
>
> In our way of integrating the net into MC Christopher Clark
> <http://arxiv.org/find/cs/1/au:+Clark_C/0/1/0/all/0/1>, Amos Storkey
> <http://arxiv.org/find/cs/1/au:+Storkey_A/0/1/0/all/0/1> is significantly
> (>60ELO) stronger than your net (I did a net similar to yours (not all
> features, but last move features), with a little  above 50% prediction
> rate), therefore I am using the net without last move feature.
>

Interesting, but I don't really understand your experiments. Our 12-layer
CNN alone is about KGS 2k, and when integrating it into MCTS our program
using very simple playouts is as strong as Fuego (about KGS 1.5d to 2d).
Are you sure their CNN is stronger?

Aja
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://computer-go.org/pipermail/computer-go/attachments/20150429/34bcdf8e/attachment.html>


More information about the Computer-go mailing list