Le Vendredi 01 Décembre 2006 00:20, Darren Cook a écrit : > >>> I believe that MC will be the only way to write a GO program in the > >>> near future leaving the other stuff in the dust ... > > > > ... > > I am certain it is for 19x19. Just look at the KGS games of Mogo on > > 19x19. I played one game against it, and won. I got the feeling it was > > slightly easier to beat than GNU Go, but that may be because I am used > > to the way Monte-Carlo programs play. I predict that in one year or two, > > classical programs will be far behind MC programs on 19x19. Maybe it > > will take less than one year. > > Are there any details, or publications, on what Mogo is doing at 19x19? There will be a technical report shortly (if everything works fine). There is no all the very recent ideas (yet to be written...), but the main ideas are there.
Most of them are not specific to 19x19, some are. > I'd thought consensus opinion here was that monte carlo scaled to 19x19 > badly. Yes monte carlo scales badly, be the results are improving very quickly! To give an idea of the scale (at least for MoGo), 70k simulations/move (with the best parameters) against gnugo 3.6/level 8 gives 89% in 9x9, 68% in 13x13, 32% in 19x19. So well, we are still quite far from gnugo (and it is only 3.6, newest versions are certainly better). However, 2.5 months ago, MoGo in 19x19 were close to 0% against gnugo 3.6/level 8 in 19x19. So I think MC programs can fill the gap quite quickly. There is also an issue of time, because in 19x19, we are in the fast increasing part of the strength/time curve. I mean that in 9x9, we don't see so much improvement with more time, but in 19x19, with the time settings of next KGS tournament (28 minutes) (and the kgs games of MoGo are with this setting), MoGo can hardly do 70k sims/move at the beginning and has an average closer to 30k sims/move. With 30k sims/move, the winning rate against gnugo becomes only 16%! With games of 50 minutes, the level would be much closer to the one of gnugo. Also, there are a lot to improvements to do in MC in a quite short term, so I share the point of view of Rémi, Don and some others when saying that MC programs will fill the gap with classical programs in 19x19. And this can be soon. Now, it is the work of the "classical approach" developers to make us wrong :). Sylvain _______________________________________________ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
