Newsgroups: comp.ai.neural-nets
Path: cantaloupe.srv.cs.cmu.edu!rochester!udel!gatech!news.mathworks.com!newshost.marcam.com!uunet!in1.uu.net!world!crump
From: crump@world.std.com (David Crump)
Subject: Any Nets that don't require initial learning?
Message-ID: <DGEBIB.4Mp@world.std.com>
Organization: The World Public Access UNIX, Brookline, MA
Date: Fri, 13 Oct 1995 16:37:23 GMT
Lines: 14

I am getting quite interested in Neural Networks these days and have been 
trying to read anything I can get my hands on. I have one question 
however: All of my reading states that the net requires some kind of 
initial training in order for it to work. Even in the "competitive" 
learning algorithm, some initial training must take place.

Is anyone familiar with networks that do _not_ require training and can 
learn on their own? Obviously, the brain does not have any training per 
se during its initial development.

Thanks a lot,
David Crump
crump@world.std.com
"Don't pat yourself on the back...PUSH yourself on the back!"
