Thank you Sylvain for conducting these experiments. We have had some very
enlightening results posted here recently in my opinion. I have to admit,
I'm surprised at how well the program seems to scale. Fortunately, I didn't
make a bet. :)
Taking for granted that these results indeed show what they seem to, and
combining them with the success of Monte-Carlo methods on 7x7 and 9x9 boards,
I'll have to change my opinion about the future of computer go quite radically.
It now seems believable to me that computer go will go the way of
computer chess,
and within the next decade or so as well. Or maybe Chrilly will make a monster
go machine even before that.
Could somebody comment please on the likely usefulness of massively parallel
machines to UCT-like algorithms.
Thanks again.
Tom.
At 21:12 10/04/2007, you wrote:
Hello,
2007/4/6, Tom Cooper
<<mailto:[EMAIL PROTECTED]>[EMAIL PROTECTED]>:
My guess is that the complexity of achieving a fixed standard of play
(eg 1 dan) using a global alpha-beta or MC search is an exponential
function of the board size.
(...)
To some extent, this is testable today by finding how a global search
program's strength scales with board size and with thinking
time
I have experiments of MoGo's playing strength against a fixed player
(Gnugo 3.7.10 level 8) on different board sizes and different thinking times.
Of course, to meet your statement we have here to assume that the
level of gnugo level 8 is a constant with the board size.
The results are that in order to keep the same winning rate, you
have to increase the number of simulations by something a little
larger than linear in the board area. From 9x9 to 13x13, you need
something like 3 times more simulations for the same winning rate.
Same thing from 13x13 to 19x19. As the time of one simulation is
linear in the board area, to keep the same level you have to give a
time which increases as power ~2.5 of the board area. So between 9x9
and 19x19, you have to give 32x more time per move for the "same
play level" (always defined as winning rate against gnugo).
This is far from being exponential. (maybe if it was exponential,
there would be little interest to work on 9x9 go).
Sylvain
_______________________________________________
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/
_______________________________________________
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/
