DISTRIBUTED ARTMAP - A NEURAL-NETWORK FOR FAST DISTRIBUTED SUPERVISEDLEARNING

Distributed coding at the hidden layer of a multi-layer perceptron (ML P) endows the network with memory compression and noise tolerance capa bilities. However, an MLP typically requires slow off-line learning to avoid catastrophic forgetting in an open input environment. An adapti ve resonance theory (ART) model is designed to guarantee stable memori es even with fast on-line learning. However, ART stability typically r equires winner-take-all coding, which may cause category proliferation in a noisy input environment. Distributed ARTMAP (dARTMAP) seeks to c ombine the computational advantages of MLP and ART systems in a real-t ime neural network for supervised learning. An implementation algorith m here describes one class of dARTMAP networks. This system incorporat es elements of the unsupervised dART model, as well as new features, i ncluding a content-addressable memory (CAM) rule for improved contrast control at the coding field. A dARTMAP system reduces to fuzzy ARTMAP when coding is winner-take-all. Simulations show that dARTMAP retains fuzzy ARTMAP accuracy while significantly improving memory compressio n. (C) 1998 Elsevier Science Ltd. All rights reserved.