Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Real AI

Author: Ashwin P. Phatak

Date: 10:08:25 10/06/01

Go up one level in this thread


Hi.Thanks to all for your interest.
Someone mentioned that a program which learned from its mistakes would be close
to AI,and todays programs do that.I have to disagree.In my opinion,it is just a
primitive memory function which is a bit better than my TV,which when switched
on goes to the channel last viewed - very useful(!),but hardly intelligent.A
really intelligent program would understand the "nature" of its mistake
positionally - it would in fact "ask" an expert if required to point out the
mistake.

Writing such a program is going to be really difficult.I believe a good way to
begin is first to modify the evaluation function to mimic human thought.That is
going to be my first step.The learning and other features would follow much
later.

Regards,
Ashwin.



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.