Newsgroups: comp.ai.philosophy
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!yeshua.marcam.com!charnel.ecst.csuchico.edu!olivea!uunet!brunix!news
From: Dave Blei <cs141041@cslab5c>
Subject: Re: RACE and IQ 
In-Reply-To: <38ib3q$ovf@nntp.Stanford.EDU> 
Content-Type: TEXT/PLAIN; charset=US-ASCII
Message-ID: <Pine.SOL.3.90.941026002730.9817D-100000@cslab5c>
Sender: news@cs.brown.edu
Organization: Brown University Department of Computer Science
References: <Cy79r6.I67@murdoch.acc.Virginia.EDU> <38ib3q$ovf@nntp.Stanford.EDU> 
Mime-Version: 1.0
Date: Wed, 26 Oct 1994 04:54:36 GMT
Lines: 28



On Tue, 25 Oct 1994 rherardi@leland.Stanford.EDU wrote:

> most AI software cannot possibly be described as intelligent 
> since the problem solving domains of AI software programs are invariably 
> restricted (even the game of chess is a micro-universe). In other words, AI 
> software programs do not solve unfamiliar problem types. 
> 
> Unlike humans, AI programs typically cannot solve unfamiliar problems even if 
> the time pressure is taken away. 
> 

But which human can solve an unfamiliar problem without a restricted 
domain?  I don't think I'd be able to win or play a game if the rules, 
goal, and language were completely unknown to me.  And even this domain 
is not unrestricted since the concepts of 'game' and 'win' and 'play' are 
known.  If I were to hand you a blue cube and a red ball and say "foo," 
(an unrestricted domain) what would you do?

Perhaps, we are able to abstract problems to a higher level than current 
AI programs but this does not mean that they are not intelligent.  It 
only means that they are less intelligent.

David Blei
Brown University.


