Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: A Practical, Implementable Self Learning Chess Computer

Author: Mike Adams

Date: 17:20:15 10/18/00

Go up one level in this thread


On October 17, 2000 at 13:46:02, Bruce Moreland wrote:

>On October 16, 2000 at 06:12:20, Graham Laight wrote:
>
>>Following discussions in the "recognising chess patterns" thread last week, I
>>believe we have reached the stage where it possible, with today's PCs, to create
>>a self learning chess computer. Before, I have always had doubts about whether
>>it could be done with current technology - now these doubts are gone.
>>
>>Disappointingly, the discussion thread died out - possibly because people didn't
>>realise what was in the thread.
>>
>>So here's the deal. I will think through the outline design of such a system,
>>and write it up in my (hopefully!) clear style for everyone's perusal if, and
>>ONLY if, at least 5 people promise to comment on the design after I have written
>>it.
>>



>>So, if you want to read it, and you're willing to comment on what I write,
>>please indicate this by responding to this message.
>>
>>Regards,
>>Graham
>
>Your proposal is the same sort of nonsense that you've been promoting for the
>past several years, and I'll give you the same sort of response I've given you
>for the past several years.
>
>Some of the steps toward your solution are important enough that if those steps
>could be completed, chess wouldn't be anywhere near the most interesting problem
>to use those tools on.  Your "practical" plan calls for enormous work and
>several major breakthroughs in AI.
>
>Several of your steps are described in a few sentences, but the amount of work
>involved to solve then would be national in scale, if they are solvable by
>humans at all in a practical time-frame.
>
>And tying together these monumental tasks is a plan that could have been devised
>by any vaguely technical person during their first few moments of thought about
>the problem.
>
>bruce


   I'm not exactly sure what he is proposing but i read in yahoo news the other
month that a self learning checkers program reached expert level at checkers.
This is not nearly as strong as a non learning checkers program with traditional
algorithms can get.  Morph on icc is a self learning chess program but as anyone
who fingers it can see it has not gotten much beyond 1000 rating.  I dont know
really what methods of learning it uses but i understand its emphasis is on
patter recognition not search. Personaly i dont think you can do well at chess
without search.  None the less morph is an academic project at ucsc and from
what i saw of the morph website quite a few papers have been published.  Mike
Adams



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.