Author: William H Rogers
Date: 10:29:26 11/09/00
Go up one level in this thread
On November 09, 2000 at 12:28:32, Bob Durrett wrote: >On November 09, 2000 at 10:31:21, William H Rogers wrote: > >>Most of the chess programs written today only "learn" by improving their opening >>books. > ><snip> > >Gosh! This sounds really dreary! > >Maybe we need a completely new generation of chess engines which incorporate >some serious middlegame learning abilities, other than just tweaking some >parameters. > >There have been many "artificial intelligence" threads here. Have none of these >threads had any impact on engine design? > There are several programs with what you might call a hi IQ, that is they are rated as masters or grandmasters depending on what kind of hardware they run on. If you could put down on paper what kinds of decisions that a program could make and how it could make it in the mid-game or anywhere else in the program, then let us all know, as it would be a great next step in computer chess programming. I am sure that every serious chess programmer is racking his or her brains to come up with some better evaluations, after all, that is the real heart of a program as the industry is concerned right now. However, some are just waiting for faster machines.?? Bill >Does anybody know whether or not "high I.Q." chess-playing programs are "on the >drawing board"? At least in the dreams and fantasies of the programmers? > >I hope there is hope of a big breakthrough in software design. We need a high >I.Q. program, don't we?
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.