DocumentCode
2850992
Title
Extreme conditions for one step convergence of the Hopfield neural network
Author
C. Gomide, F.
fYear
1989
fDate
14-17 Nov 1989
Firstpage
220
Abstract
The authors analyze the best and worst conditions of equilibrium for the simplified version of the Hopfield neural network. This analysis can elucidate how the network is able to recognize the learned patterns, and how more complex and detailed analysis can be carried out in a system-theoretic framework. It is shown that, in the worst case, the equilibrium is not guaranteed for a stored pattern
Keywords
learning systems; neural nets; pattern recognition; Hopfield neural network; artificial intelligence; extreme conditions; learned patterns; one step convergence; pattern recognition; Convergence; Difference equations; Hamming distance; Hopfield neural networks; Matrices; Neural networks; Nonlinear equations; Pattern analysis; Pattern recognition; Testing;
fLanguage
English
Publisher
ieee
Conference_Titel
Systems, Man and Cybernetics, 1989. Conference Proceedings., IEEE International Conference on
Conference_Location
Cambridge, MA
Type
conf
DOI
10.1109/ICSMC.1989.71284
Filename
71284
Link To Document