-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 I know Hsu from my computer chess days. Of course most of us probably realize he was on the Deep Blue team.
I think he is betting on null move proving - but I'm real skeptical that it will be effective in Computer Go. It will indeed reduce the tree significantly, but this comes at a qualitative price that is not so bad in Chess but is a lot in Go. I would like to see that kind of hardware (which was described in the article) applied to 19x19 GO using UCT combined with more research on UCT, that could produce something that plays in the Dan range, although I'm very skeptical it will be in the "high Dan" range, certainly not the stronger than any human player. Having said that, one must be very careful about making predictions, lest one ends up with egg on their face. Computer chess started with wild optimism, then extreme pessimism then big surprise! I clearly remember the days when it was ludicrous to imagine chess computers ever touching even the weak masters. It was just not possible and a simple calculation on a paper napkin proved that it could never happen. Woops! In those days you were considered an idealistic fool if you were optimistic but it turns out the optimistic ones were the most realistic and the ones who were objectively "reasonable" and pragmatic were the fools. I never did scaling experiments on 19x19 with UCT, but I have no reason to believe it would not show the same amazing scaling curve. Lazarus is so weak at 19x19 that it would take a few doublings of speed to get competitive with reasonably good programs. Has anyone else done scaling experiments with 19x19 and UCT? - - Don Ian Osgood wrote: > Greetings, > > I noticed that the following link was recently added to the Computer Go > Wikipedia article. > > http://www.spectrum.ieee.org/oct07/5552 Cracking Go, by Feng-hsiung Hsu, > IEEE Spectrum magazine, October 2007. > > He claims it should be possible to build a Go machine stronger than any > human player. > > Ian > _______________________________________________ > computer-go mailing list > computer-go@computer-go.org > http://www.computer-go.org/mailman/listinfo/computer-go/ > -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.6 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFHAmfWDsOllbwnSikRAgyFAJ9m4ZFXFzkDXC+S9PAki/bQeOIRVwCgq4Ns cUdQ17VrpLPOAO1ipyu9QXc= =aQ37 -----END PGP SIGNATURE----- _______________________________________________ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
