Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Addendum

Author: Rémi Coulom

Date: 05:57:46 05/23/05

Go up one level in this thread


On May 23, 2005 at 07:47:19, Christopher Conkie wrote:

>Hello Andrew,
>
>There is also this....
>
>Learning Piece-Square Values using Temporal Differences. ICCA Journal, Vol. 22,
>No. 4, pp. 223-235.
>
>A very similar title that you could probably get here by contacting ICGA at....
>
>http://www.cs.unimaas.nl/icga/
>
>Regards
>
>Christopher

It's a pity the ICGA does not make their papers freely available online.

My suggestion would be to ask the authors by mail first. It's likely they can
send a copy.

This paper is one of the most interesting I know about reinforcement learning of
evaluation functions in chess. Baxter learns from online play against a variety
of opponents. Beal and Smith use self-play. I find their results are more
convincing of the power of the method.

Rémi



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.