Title :
Stability regions of recurrent type neural networks
Author_Institution :
Dept. of Electr. Eng., Bogazici Univ., Istanbul, Turkey
fDate :
5/21/1992 12:00:00 AM
Abstract :
The stability regions of the stable equilibria of the Hopfield type recurrent neural networks are investigated. The dimensions of these regions directly determine how much partial information is needed for an imperfect pattern to approach a desired pattern. Therefore estimation of such regions is instrumental in the analysis of analogue recurrent neural networks.
Keywords :
neural nets; stability; Hopfield type recurrent neural networks; analogue recurrent neural networks; desired pattern; imperfect pattern; partial information; stability regions; stable equilibria;
Journal_Title :
Electronics Letters
DOI :
10.1049/el:19920649
