Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Genetic Algorithms applied to King-Pawn endgames

Author: Pat King

Date: 19:17:54 09/11/00

Go up one level in this thread


On September 11, 2000 at 16:00:27, Dave Gomboc wrote:

>On September 11, 2000 at 14:59:02, Pat King wrote:
>
>>On September 10, 2000 at 19:32:32, Richard A. Fowell (fowell@netcom.com) wrote:
>>
>>>I just ran across this paper on the subject topic.
>>>Somewhat academic, since a tiny tablebase handles
>>>this ending well, but I thought it might be interesting
>>>for those wondering about applications of GA to chess.
>>>
>>>http://www.cs.utah.edu/~wyman/classes/ML_proj/paper.html
>>>
>>>Richard A. Fowell
>>
>>It's good to know of some success out there. I tried, several years ago, to
>>weight my evaluation via GA. The program was just too painfully slow. I have not
>>given up, however, and hope to include GA techniques in Zotron some day, so that
>>no two players face quite the same opponent. I think it will lead to interesting
>>opponents, but not neccessarily better ones. So far, the learning rate seems to
>>favor hand tweaking (even such a shaky hand as mine).
>>
>>Pat King
>
>From reading the article, I'd hesitate before calling the technique a success.
For chess, no. For machine learning, yes. That's why I'm not aiming for more
than "interesting" at this point
>
>Dave
Pat



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.