Path: utzoo!attcan!uunet!ogicse!milton!forbis
From: forbis@milton.u.washington.edu (Gary Forbis)
Newsgroups: comp.ai
Subject: Re: What AI is Exactly - Another Update.
Keywords: intelligence
Message-ID: <7766@milton.u.washington.edu>
Date: 19 Sep 90 05:13:01 GMT
References: <3734@gara.une.oz.au>
Organization: University of Washington, Seattle
Lines: 25

I've continued to think about the current attempt to define "intelligent 
system".  I feel like I am nit-picking.  I have taken a stance in another
conference which makes this minor point important to me right now.

In article <3734@gara.une.oz.au> pnettlet@gara.une.oz.au (Philip Nettleton) writes:
>----------------------------------------------------------------------
>			DEFINITION:
>	GENERAL REQUIREMENTS OF AN INTELLIGENT SYSTEM.
>
>a)	The system MUST be able to learn. This implies that the
>	system MUST have a memory for learning to be maintained.

Provided that the system is in a sufficiently rich environment I would
concede this point.  The problem arises when the system exists in a very
poor environment.  Is intellegence contextual?  That is could a system
be said to be intelligent when in one environment and not when in another?

The second complaint is that once the system HAS learned can it still be
said to be able to learn and is this necessary?  After learning has taken
place would the same response to the same stimulus take the system out of
the realm of intelligence? 

Is intelligence the process or the results?

--gary forbis@cac.washington.edu