Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Artificial Intelligence

Author: Pedro Irazoqui-Pastor

Date: 13:25:45 01/27/98

Go up one level in this thread


On January 27, 1998 at 15:22:07, Robert Hyatt wrote:
>Note that you missed the keyword, like most everyone else is missing it.
>This word is "artificial".  IE would you argue the point that somethine
>like "sweet 'n low" *is real sugar*?  Or is it an "artificial sugar" or
>a "sugar substitute"?  That's the issue I see.  Computers are playing
>chess.

Nobody calls Nutrasweet artificial sugar, they call it sweetener, or
sugar substitute, there is an important semantic difference here. It is
not sugar even though it may share some properties with sugar. If it was
called artificial sugar, it would be sugar, made artificially. Similarly
with AI, for it to be intelligence, wether artificial or natural it has
to be made up of far more than 1/0 operations. Otherwise it is just an
intelligence substitute, which does not invalidate it as competent in
the field of chess. This merely refines the defenition.

>If we accept the age-old premise that playing chess is intelligent, then
>computers are somehow "simulating intelligence" since the obviously are
>not
>human and don't have 'real' intelligence.  This is why it is, and always
>has been called 'artificial intelligence'...

The fact that something is and allways has been a certain way does not
mean it cannot change. For any scientific field, such as the one under
discussion, to advance, past assumptions cannot be held onto strictly by
merit of their having remained uncontested in the past. People make
mistak, they recognize them, and eventually those mistakes get
corrected. What I am suggesting is that if the word intelligence was
ever used to describe something which is purely arithmetical and
entirely unintelligent, then that it no longer be so.


>BTW, there's no guarantee that we don't operate at a binary level.
>Physics
>makes this likely at some level or another.
>

I have never seen anything in physics to suggest that we may be binary
in any way. Binary operations are the realm of the digital, we are
analog, analog and spiritual. Indeed, I hope I never do see anything
which reduces all the beautifull complexities of the human experience to
ones and zeroes.

That AI has allways held chess for computers as one of it's cornerstones
is valid so long as chess is approached from an intelligent perspective,
i.e. neural nets. Deep Blue approaches chess from an unintelligent brute
force perspective, effectively but unintelligently. This essentially
defeats the point of the exercise froma an AI perspective. Say I had to
solve an integral in calculus, I could plot my function, understand it's
meaning and calculate the area under the curve using my knowledge of
mathematics and derive a meaningfull result, or I could look the answer
up in an integral table and plug it in. Both approaches would give me
the same correct result, but if the objective was to demonstrate that I
could LEARN calculus the first would have achieved it while the second
would not. The first would be intelligent, the second would be brute
force. Note that a computer could never do the first, only the second or
a numerical approximation of the first.



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.