Properties of learning in ART1

The authors consider the ART1 neural network architecture. Useful properties of ART1, associated with the learning of an arbitrary list of binary input patterns, are examined. These properties reveal some of the good characteristics of the ART1 neural network architecture when it is used as a tool for the learning of recognition categories. In particular, it was found that if ART1 is repeatedly presented with an arbitrary list of binary input patterns, learning self-stabilizes in at most m list presentations, where m corresponds to the number of distinct size patterns in the input list