Supervised learning by adaptive resonance neural networks

Author(s): Carpenter, G.A. | Grossberg, S. | Markuzon, N. | Reynolds, J.H. | Rosen, D.B. |

Year: 1993

Citation: In M. Marinaro & G. Scarpetta (Eds.), Structure: From Physics to General Systems, 2. Festschrift volume in honor of the 70th birthday of Professor Eduardo R. Caianiello. Singapore: World Scientific Publishing Co., 36-47.

Abstract: ARTMAP is a class of neural network architectures that perform incremental supervised learning of recognition categories and multidimensional maps in response to input vectors presented in arbitrary order. The first ARTMAP system (Carpenter, Grossberg, and Reynolds, 1991) was used to classify binary vectors. This article describes a more general ARTMAP system that learns to classify analog as well as binary vectors (Carpenter, Grossberg, Markuzon, REynodls, and Rosen, 1992). This generalization is accomplished by replacing the ART 1 modules (Carpenter and Grossberg, 1987) of the binary ARTMAP system with Fuzzy ART modules (Carpenter, Grossberg, and Rosen, 1991). Where ART 1 dynamics are described in terms of set-theoretic operations, Fuzzy ART dynamics are described in terms of fuzzy set-theoretic operations (Zadeh, 1965). Hence the new system is called Fuzzy ARTMAP. Also introduced is an ARTMAP voting strategy.

Topics: Machine Learning, Models: Fuzzy ARTMAP,

PDF download

Cross References