In this paper the process of forgetting, termed 'engram decay', in artificial neural networks is examined for three classes of networks: the self organizing map, fuzzy logic adaptive resonance theory and, maximally connected backpropagation networks. All networks were trained to categorize three varieties of iris. How iris categories are forgotten is shown to be strongly related to the distribution of iris categories in feature space.
IEEE International Conference on Neural Networks (ICNN-95), Perth, Australia, 27 November-01 December 1995, pp. 669-674
Copyright © 1995 IEEE. The published version is reproduced in accordance with the copyright policy of the publisher. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.