-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi Steve,
I don't fully understand what you are saying here. steve uurtamo wrote: > I think that there's an apples/oranges thing going on here. > >> My hunch, however, is that they won't play a >> significant role in creating a machine that can top the best human >> players in the 19-by-19 game. > > i agree with this statement. My hunch is that won't happen for a few decades. If I'm right, then it's too far to predict which technique will do it. Probably something new. Again, it's more interesting to talk about which technique will become the rage and produce the best programs in the short term. I think that is UCT and it's happening now. UCT is the most promising for 19x19 progress that we have now. >> And MC programs are more scalable that traditional programs. That >> seems like some evidence that it can or will. Especially given that >> the current techniques are still so young. > > i do not agree with this statement. I think this statement is more or less true. Didn't you see the scalability data for 19x19? In fact didn't you help me produce it? > "top the best human players in a 19x19 game" is quite a bit different than > "at the level of the strongest traditional programs". "at the level of", or > "near the level of", or "slightly better than" just means (perhaps) that the > wheel has been re-invented. it could mean more than that, but there surely > doesn't seem to be much evidence for that at this point. This is a hard paragraph to understand. I wish you had included the statement that you are referencing here and who said it. In what sense could this be the wheel re-invented? It appears to be the case that 19x19 UCT MC programs are "better than" the traditional program now. At least from the email from Rémi. I'm not sure what you imply by lack of evidence, but the evidence is getting pretty strong in favor of UCT type programs being superior if that's what you mean. It isn't even a question at smaller boards. > "scalable" doesn't mean linear, and it also doesn't give an asymptotic growth > function or a constant. if anyone anywhere could give a good estimate for how > many cpus it would take, with any particular algorithm, to beat a professional > player, and if the number were feasible, there's no reason not to start > building > such a machine. There is almost a constant ELO improvement for each factor of speed increase such as a doubling. This may not apply to every implementation - it's possible to apply non scalable improvements that may have to be removed later. I think Mogo did this earlier with a scheme to reduce the work in earlier version of the 19x19 player. I figured they would have to remove this eventually because it seemed cheesy to me. And sure enough they did. I'm sure some will believe this observed scalability is short lived but I know of no reason to believe that other than superstition. Even to this day, people keep thinking this about computer chess. My friend and former partner Larry Kaufman recently told me that many were surprised that the scalability of extra CPU power in chess has remained nearly linear. For decades now it was predicted that this was "about to end" based purely on irrational hunches but this hasn't happened yet. Yes, it has gradually tapered, but it's been remarkably gradual and now computers are playing well above the best humans and computers rarely lose a game to the best humans. Even more remarkable is that there is a fight among the best computers - they vary significantly in strength and this is an indication that there is still a long way to go. There is no indication whatsoever than computers are even close to topping out in computer chess. That's why I believe a super hardware gizmo could easily be built that would be in the DAN range somewhere at 19x19, at least low Dan. I'm not so bold as to predict that it will be at top human levels any time soon though. - - Don > s. > > > > > > ____________________________________________________________________________________ > Be a better Heartthrob. Get better relationship answers from someone who > knows. Yahoo! Answers - Check it out. > http://answers.yahoo.com/dir/?link=list&sid=396545433 > _______________________________________________ > 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 iD8DBQFHDtNEDsOllbwnSikRAmy0AKDc57sD8hl+t8pWXR4izlTB2AIXaACgofqs P21m+sIN+Bx8gqWpbrx+bLI= =u+Pm -----END PGP SIGNATURE----- _______________________________________________ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
