Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!cornell!batcomputer!rpi!rpics!hiebeler
From: hiebeler@rpics (Dave Hiebeler)
Newsgroups: comp.ai.neural-nets
Subject: Re: bottom-up (was Re: NN Question)
Message-ID: <814@rpi.edu>
Date: 3 Mar 89 05:23:29 GMT
References: <32125@gt-cmmsr.GATECH.EDU> <10143@nsc.nsc.com>
Sender: usenet@rpi.edu
Reply-To: hiebeler@turing.cs.rpi.edu (Dave Hiebeler)
Distribution: usa
Organization: RPI CS Dept.
Lines: 23

In article <10143@nsc.nsc.com> andrew@nsc.nsc.com (andrew) writes:
>That's why, to avoid the "AI trap", it's maybe best to start bottom-up,
>rather than the heretofore conventional psychological/serial-symbolic
>approach of top-down (macroscopic) behavioural analysis. 

  I'm all for the bottom-up approach to things; my interest is
in modeling physical phenomena (and also some other phenomena,
sometimes referred to as "artificial life") using cellular
automata.  You can't get much more bottom-up than cellular automata;
one of the basic principles is that you take a whole lot of very
simple things at a very low-level with fairly simple interactions,
and observe the resultant macroscopic behavior, which is sometimes
quite surprising.  I don't much like the top-down approach that
many people try to use with AI and some other areas; I don't think
it will accomplish as much as the bottom-up method.  This is, of
course, just a personal opinion.
  I realize this is not directly related to neural-nets; I just
couldn't resist saying something.
----
Dave Hiebeler      Internet: hiebeler@cs.rpi.edu  (preferred address)
R.D. Box 225A                hiebeler%cs.rpi.edu@itsgw.rpi.edu
Chatham, NY 12037    Bitnet: userfrzk@rpitsmts.bitnet
  "xue zai xao"     "Off we go, into the wilds you ponder..."