Author: Komputer Korner
Date: 13:13:30 01/05/99
Go up one level in this thread
On January 04, 1999 at 15:35:44, Jay Scott wrote: > >On January 04, 1999 at 15:10:07, James Robertson wrote: > >>On January 04, 1999 at 14:21:31, Jay Scott wrote: >> >>>One idea about why it works so well in backgammon is that the >>>in backgammon the dice force a self-play program to explore all the >>>important parts of the game space. In chess, a learning program can make >>>the same mistakes over and over and never correct them because it doesn't >>>know how to punish them. >> >>Well; they know how to punish them, but they don't know why. All they know is >>that they won the game at the end. The problem with chess learning is that every >>aspect learned must be put in by the program by the programmer. e.g. if the >>programmer puts piece/square table learning in, the program learns piece/square >>tables. If they put doubled pawn awareness in, the program fiddles with the >>doubled pawn scoring. And so on... > >It depends, of course. If the programmer puts in the basic position info >(this set of squares has white pawns...) and allows the program to combine it >arbitrarily, then in principle the program can learn anything. For example, >self-play learning would make sense for a genetic programming system, which >could do this. It happens that nobody has gotten a genetic programming >system to successfully learn anything this difficult yet, but there's >no obstacle in principle. It's those darn practical details! :-) > > Jay Read the articles in the ICCA journal about temporal difference learning for chess programs. It does work. -- Komputer Korner
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.