DocumentCode
2618366
Title
Optimal associative mappings in recurrent networks
Author
Yang, Jian ; Dumont, Guy A.
Author_Institution
Dept. of Electr. Eng., British Columbia Univ., Vancouver, BC, Canada
fYear
1991
fDate
18-21 Nov 1991
Firstpage
48
Abstract
Optimal associative mappings are suggested in the Hopfield model. Orthogonal vector space and optimal distributed storage are provided by the extended Hopfield network and the iterative storage procedure, respectively. Subspace is implemented to establish optimal memory structure. The implementation of the novel version of the Hopfield network with this storage technique markedly improves the network performances. This is demonstrated through an application to pattern recognition of acoustic emission signals
Keywords
content-addressable storage; neural nets; pattern recognition; Hopfield model; acoustic emission signals; content addressable storage; iterative storage procedure; neural nets; optimal associative mappings; optimal distributed storage; optimal memory structure; pattern recognition; recurrent networks; Associative memory; Encoding; Feature extraction; Intelligent networks; Neurons; Vectors;
fLanguage
English
Publisher
ieee
Conference_Titel
Neural Networks, 1991. 1991 IEEE International Joint Conference on
Print_ISBN
0-7803-0227-3
Type
conf
DOI
10.1109/IJCNN.1991.170380
Filename
170380
Link To Document