Re: computer-go: Go AI

["Rod Waldhoff" <r_waldhoff@xxxxxxxxxxx>]

There have been quite a few machine learning approaches to computer go, and
many described in the literature, try:

http://www.usgo.org/computer/literat.html
http://home.t-online.de/home/markus.enzenberger/compgo_biblio.html

Not very many machine learning approaches to go have been successful.  I
think part of the reason for this is that often go is used as a 'sample
problem' for trying out new machine learning techniques, so the focus is on
the learning algorithm rather than go.

----- Original Message -----
From: Jeff Massung <jmassung@xxxxxxxxxxxxxxxxx>
To: <computer-go@xxxxxxxxxxxxxxxxx>
Sent: Sunday, January 02, 2000 1:09 PM
Subject: Re: computer-go: Go AI


> I know that some have, but they have been rather unsuccessful, simply
> because of the number of patterns that are possible, and because it has to
> understand which patterns are more important than others, and how they
> affect the board as a whole.
>
> Jeff
> ----- Original Message -----
> From: Daniel Eaton <danieljameseaton@xxxxxxxxxxxxxxxxx>
> To: <computer-go@xxxxxxxxxxxxxxxxx>
> Sent: Thursday, December 30, 1999 11:35 AM
> Subject: computer-go: Go AI
>
>
> > Hey, does anyone know if any go ai programs have taken the approach of
> > machine learning... neural networks and such?
> > Thanks, Daniel Eaton.
> > ______________________________________________________
> > Get Your Private, Free Email at http://www.hotmail.com
> >
>
>