Newsgroups: comp.ai.neural-nets
Path: cantaloupe.srv.cs.cmu.edu!bb3.andrew.cmu.edu!news.sei.cmu.edu!cis.ohio-state.edu!magnus.acs.ohio-state.edu!math.ohio-state.edu!caen!saimiri.primate.wisc.edu!ames!tulane!cs.cs.uno.edu!news
From: dnulu@mandelbrot.math.uno.edu (Deepak Nulu)
Subject: summary of unsupervised networks
Message-ID: <1994Nov15.204257.15500@cs.uno.edu>
Keywords: unsupervised networks, unsupervised learning
Sender: news@cs.uno.edu
Organization: University of New Orleans (Computer Science)
Date: Tue, 15 Nov 1994 20:42:57 GMT
Lines: 119

hello neural netters,

this is a summary of the responses i got for the following post:

> i would like to know the 'names' of various neural nets that
> make use of unsupervised learning. the only nets i have come
> across (or seem to recollect) are:
>
> Self Organizing Maps (SOM)
> Adaptive Resonance Theory (ART)

i received only two responses; thanks to dr. albert nigrin and dr.  
jonathan marshall for their responses. this is the response from
dr. nigrin:

-----------------------
%You might look at my SONNET networks which generalize the ART networks to
%embedded patterns and operation on ever changing temporal patterns.   
These
%networks were described in:
%
%Albert Nigrin. 1993. Neural networks for Pattern Recognition.  The MIT  
Press,
%Cambridge MA.
%
%Michael Page. 1993.  Modelling aspects of music perception using
%self-organizing neural networks.  Unpublished doctoral dissertation,  
University
%of Wales.
%
%Simon Roberts and Mike Greenhough. 1994.  The detection of rhythmic  
repetition
%using a self-organizing neural network.  International Computer Music
%Conference.   Also available from the authors at spxscr@thor.cf.ac.uk
%
%There are also some SONNET papers that are ftp'able from the machine:
%palestrina.acc-lab.american.edu  (the numeric address is 147.9.201.30) in  
the
%ftp directory: pub/neural_nets.
%
%You might also look at the EXIN networks described in:
%
%Jonathan A. Marshall. 1992. Development of perceptual context-sensitivity  
in
%unsupervised neural networks: Parsing, grouping and segmentation.
%In Proceedings of the International Joint Conference On Neural Networks,
%volume 3, pp. 315--320, Baltimore, Md.
%
%
%Good Luck
%Al
-----------------------

the following is the response that i got from dr. marshall, with a few
deletions made by me. the paper he mailed me is titled
"adaptive perceptual pattern recognition by self-organizing neural  
networks:
context, uncertainity, multiplicity and scale."

-----------------------
%The paper I sent you has been accepted for publication in Neural  
Networks,
%and it will appear in 1995.
%
%The one that Al Nigrin mentioned (from the 1992 Proc of the IJCNN) is an
%older, incomplete version.
%
%--Jonathan
%
%  _____
% /     \   Jonathan A. Marshall                         
marshall@cs.unc.edu
% -------   Dept. of Computer Science, Sitterson Hall
% | | | |   CB 3175, University of North Carolina       Office  
919-962-1887
% | | | |   Chapel Hill, NC 27599-3175, U.S.A.             Fax  
919-962-1799
% =======
-----------------------

i managed to find a few more unsupervised learning networks. three of them
can be found in

d.e. rumelhart, j.l. mcclelland, and the pdp research group, "parallel
distributed processing. explorations in the microstructure of cognition
volume 1: foundations", the mit press, 1988.

chap 5.	feature discovery by competitive learning - d.e. rumelhart &
d. zipser
chap 6.	information processing in dynamical systems: foundations of
harmony theory - p.smolensky
chap 7.	learning and relearning in boltzmann machines - d.e. rumelhart,
g.e. hinton, r.j. williams

the neocognitron developed by kunihiko fukushima is also an unsupervised
network:

fukushima, k., "Neocognitron: a self-organizing neural network model for a
mechanism of pattern recognition unaffected by shift in position",  
biological cybernetics 36, 193--202, springer verlag, 1980.

if anybody has comments, or additions, please post to this newsgroup. and  
here is a parting question:

can the auto-associative learning task be considered as unsupervised  
learning? if so, is the hopfield network considered to be an unsupervised  
network?

thanx.

deepak

--
=====================================================================
Deepak Nulu                     |       email:  dnulu@math.uno.edu
Graduate Student                |               dxnee@uno.edu
Dept. of Electrical Engg.       |
Dept. of Mathematics            |       phone:  (504) 283-4153
Univ. of New Orleans            |
