Newsgroups: comp.ai
Path: cantaloupe.srv.cs.cmu.edu!bb3.andrew.cmu.edu!nntp.sei.cmu.edu!cis.ohio-state.edu!math.ohio-state.edu!howland.reston.ans.net!spool.mu.edu!umn.edu!news
From: "Alan J. Robinson" <robin073@maroon.tc.umn.edu>
Subject: Time's up!
Message-ID: <42933.robin073@maroon.tc.umn.edu>
X-Minuet-Version: Minuet1.0_Beta_18A
Sender: news@news.cis.umn.edu (Usenet News Administration)
Nntp-Posting-Host: dialup-4-78.gw.umn.edu
X-Popmail-Charset: English
Organization: University of Minnesota, Twin Cities
Date: Tue, 5 Sep 1995 14:00:46 GMT
Lines: 76

Over the years many people have expressed dissatisfaction with the AI 
research agenda and with progress in AI research - many of the Usenet 
posts express these sentiments.  In the past few years I have spent a 
lot of time in scientific areas outside AI, and have become aware of 
some trends which I believe will revolutionary impact in many areas 
of science and technology, including AI.  The purpose of my post is 
to point out these trends and how they might affect AI researchers - 
I hope that the information might be of some value in preparing 
future education and research plans.

The 1990s were declared the Decade of the Brain by the U.S. Congress, 
reflecting the rapid and increasing pace of developments in pure and 
applied neuroscience research.  These developments have now reached 
the point where much of the information which educated people and 
even most scientific researchers still hold about the behavioral and 
brain sciences is now obsolete.  This includes material still being 
taught in the leading psychology departments and medical schools, and 
by notable scientists.  Nearly all the definitive research has been 
published since 1985, and it's significance is still not widely 
appreciated or understood.

The fundamental importance of neuroscience research has been 
recognized by the appointment of Dr. Floyd Bloom, Head of the 
Neuropharmacology Dept. at Scripps Institute in La Jolla, CA, as 
Editor-in-Chief of Science.  The previous Editor-in-Chief, Dr. 
Daniel E. Koshland Jr, Professor of Biochemistry at UC Berkeley, also 
recognized its importance.  Neuroscience is complemented by
molecular biology, and the very top scientists in these combined 
research efforts are Crick at Salk, Edelman at Scripps, and Watson 
at Cold Spring Harbor.

It now seems highly probable that the psychiatric disorders will 
be the first major aspect of brain functioning which will have a 
comprehensive theory.  The chapter "Diseases of Consciousness" 
in Edelman's book "The Remembered Present" shows why these 
disorders are relevant to AI.  The psychiatric disorders also 
involve destabilization of homeostatic control loops in the 
brain with resulting chaotic oscillations, of direct interest 
to neural net researchers.

Biology is a data driven science, and excessive theoretical 
speculation and philosophizing is discouraged - something for AI 
researchers to keep in mind.  My message title refers to the time for 
reserach developments to become commercialized.  In medicine, which 
is the slowest to change, it can take 50 years between the time that 
a phenomenon is first observed in the laboratory and the corresponding 
routine, wide-spread clinical use ofthe resulting 
technology.

Both AI and computing started around 1945, give or take a couple of 
years - the logical basis of neural action was published I believe in 
1943.  Since then there have been many generations of commercialized 
computer technology.  Fifty years later, AI is still not in routine 
commercial use, a sure indication that the research is on the 
wrong track.  (To be fair, the problem turned out to be much more 
difficult than anticipated, and useful neuroscience data has only 
recently become available.)

With the neuroscience research program accelerating like a juggernaut, 
it can only be a matter of time before its impact is felt in related 
fields such as AI.  A quick re-education in neuroscience, molecular 
biology, and biopsychology would appear to be the very minimum 
response for AI and neural net researchers.  The Behavioral and Brain 
Sciences journal is a good starting point to obtain a feel for the 
interdisciplinary nature of the enterprise, but the reader should take 
into consideration the fact that even many of the commentators on 
the articles are themselves not up-to-date outside their own 
specialties.

Please E-mail me if you have any questions - but please note, this is 
just my own interpretation of events - I don't have a direct line to 
Crick or Watson or the other luminaries.

                            Alan J. Robinson
Golden Hind International - Computer Graphics, Artificial Intelligence
                       robin073@maroon.tc.umn.edu
