Hamburg 12/8/1993

                               NeXT Application
                        Adaptive Resonance Theorie (ART)

     Neural network models based on the ART developped by
     Carpenter and Grossberg have the ability of stable unsupervised
     learning.

     In this package you will find a NeTXStep application (tested for 

     release 3.0)

         ART.app and its source code (*.m, *.h, *.nib)

     simulating the basic ART-2 network for recognition and
     classification of analog patterns. Additionally you find a modifi-
     cation of the network, which allows distributed classification
     by of superposition of orthogonal pattern components.

     A PostScript file ART.ps contains a detailed description in German
     language of my ART-2 modification, its results and a rough 

     introduction to the usage of ART.app.

     This document is the outcome of my project work at the institute
     of Technical Computer Science VI at TU Hamburg-Harburg, Germany.

     For fundamental understanding of Grossberg's Adaptive Resonance 

     Theory and derived neural networks please have a look at

     (1) Carpenter, G.A., Grossberg, S.: A massively parallel
         architecture for a self-organizing neural pattern recognition 

         machine, Computer Vision, Graphics and Image Processing,
         Academic Press, Inc., 1987

     (2) Carpenter, G.A., Grossberg, S.: ART-2: self-organization of 

         stable category recognition codes for analog input patterns,
         Applied Optics, Vol.26, 1987

      For questions and suggestions please contact


         e-Mail: ti6cmt@tick.ti6.tu-harburg.de


                           Christian Mueller-Tomfelde