Title :
Generalized adaptive bidirectional associative memory
Author :
Schürmann, Bernd
Author_Institution :
Siemens AG, Munich, West Germany
Abstract :
A model of an artificial neural network for pattern association that is dynamically stable yet employs a weight matrix with asymmetric connections is discussed. A Lyapunov function for the system is explicitly specified. Its behavior as a function of time during the learning phase indicates the quality of the network´s performance regarding pattern association. An illustrative numerical example demonstrates the virtues of this model in comparison with others. The crucial role of the constants entering the artificial neural systems is emphasized.<>
Keywords :
Lyapunov methods; adaptive systems; content-addressable storage; neural nets; pattern recognition; Lyapunov function; artificial neural network; asymmetric connections; generalized adaptive bidirectional associative memory; learning phase; pattern association; weight matrix; Adaptive systems; Associative memories; Lyapunov methods; Neural networks; Pattern recognition;
Conference_Titel :
Neural Networks, 1989. IJCNN., International Joint Conference on
Conference_Location :
Washington, DC, USA
DOI :
10.1109/IJCNN.1989.118564
