Author: Christophe Theron
Date: 10:10:08 12/27/01
Go up one level in this thread
On December 27, 2001 at 02:13:34, Tom Kerrigan wrote:
>On December 26, 2001 at 17:47:12, Christophe Theron wrote:
>
>>I don't think so, but I think at some point the only way to improve will be to
>>incorporate a way for the program to learn without the programmer, to remember
>>its experience and improve on it, and to adapt its play to its opponent.
>
>I don't see this as likely because of the numbers involved.
>
>For a computer to recognize features, it has to loop over them. And there might
>as well be an infinite number of features possible on a chess board, so that
>loop is going to take a while.
>
>The human brain is sloppy and bad at this task, so maybe there's some way to do
>sloppy and bad learning that does better than what we have now, but I wouldn't
>know how to go about that.
That's the problem, I do not know either.
But the reason we know that there is another way of playing chess is because
human chess players have a "NPS" that is a very small fraction of todays
computers NPS. The best human players only look at 1/100000 of the nodes a
computer looks at in order to play a move of comparable quality.
So the "very slow NPS" approach is workable, because there are computing
entities using it successfully.
The reason why this approach has always failed is because a human programmer
cannot write such a program directly. That is the first approach I have used, so
I know a little bit about the problems it creates.
My guess is that the "very slow NPS" approach needs to know about a lot of
particular cases (rules used very seldomly). So the number of rules in such a
program, the interactions between them, and the fact that they have to be
adjusted after every game, makes it totally unpractical for the current ways of
writing programs.
The program must be able to manage its set of rules by itself, maybe with the
help of a human player ("teacher"?).
There is no proof that the very slow NPS approach is superior to our current way
of programming computers to play chess. It might very well be that the VSNPS
approach actually requires more computing power than the fast NPS one.
But it is an interesting thing to explore, because it would certainly tell us a
little bit about how our brains work.
Christophe
This page took 0.01 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.