From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.ecf!news-server.csri.toronto.edu!bonnie.concordia.ca!uunet!cs.utexas.edu!swrinde!gatech!pitt!geb Mon Dec 16 11:01:50 EST 1991
Article 2114 of comp.ai.philosophy:
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.ecf!news-server.csri.toronto.edu!bonnie.concordia.ca!uunet!cs.utexas.edu!swrinde!gatech!pitt!geb
>From: geb@dsl.pitt.edu (gordon e. banks)
Newsgroups: comp.ai.philosophy
Subject: Re: From neurons to computation: how?
Message-ID: <12684@pitt.UUCP>
Date: 14 Dec 91 13:45:16 GMT
References: <1991Dec11.023152.14901@smsc.sony.com> <12664@pitt.UUCP> <1991Dec11.220512.22087@smsc.sony.com>
Sender: news@cs.pitt.edu
Organization: Decision Systems Laboratory, Univ. of Pittsburgh, PA.
Lines: 39

In article <1991Dec11.220512.22087@smsc.sony.com> markc@smsc.sony.com (Mark Corscadden) writes:
>In article <12664@pitt.UUCP> geb@dsl.pitt.edu (gordon e. banks) writes:
>>The slug brains were mapped by Eric Kandel and coworkers of Columbia
>>University.  I believe they do have all the firing times, thresholds,
>>and synaptic connections.  There are 27 neurons in this slug brain,
>>if I remember correctly.
>
>Please accept my apology for my previous remark, and thank you for the
>pointer.  If this slug brain does indeed only contains 27 neurons then
>it seems at least possible that the complete functioning is known.  If
>I find the work by Eric Kandel I'll post a summary, if there is interest
>in this topic here.  *I* think it's extremely interesting, but it seems
>to be getting away from the point of comp.ai.philosophy (sorry).
>
Weemba says he thought Kandel's slugs had many more neurons, but
I'm sure I recall the 27 from some study.  There is a fellow at
CMU working on some slug or worm, perhaps it is in his talk that
I picked up the number 27.

I don't think it is getting away from comp.ai.philosophy at all.
A biologically oriented approach is quite warranted, I think.
Some people figure they are going to go out and design a human
brain at first shot, but since human brains slowly evolved from
very primitive brains, it makes a lot of sense to me to think that
if we learn to simulate and understand these simple brains, we can
progress from step to step in complexity and have a much better chance
of building a real artificial intelligence.  The problem, of course,
is that most people interested in computers know little about biology
and vice versa.  So we get attempts to build a brain using logic,
which is at the top pinnacle of evolution and tells us little about
the real hardware.  Top down is fine, but bottom up is needed too.



-- 
----------------------------------------------------------------------------
Gordon Banks  N3JXP      | "I have given you an argument; I am not obliged
geb@cadre.dsl.pitt.edu   |  to supply you with an understanding." -S.Johnson
----------------------------------------------------------------------------


