Title :
Decorrelating parametrical neural network
Author :
Kryzhanovsky, Boris V. ; Kryzhanovsky, Vladimir M. ; Fonarev, Anatoly B.
Author_Institution :
Inst. of Opt. Neural Technol., Russian Acad. of Sci., Moscow, Russia
fDate :
31 July-4 Aug. 2005
Abstract :
We developed a new network architecture, allowing us to increase the recognizing characteristics of the Hopfield model substantively. In addition it is effective when recognizing correlated patterns. It is shown that the storage capacity of the network increases exponentially with increase of the free parameter of the problem. The boundaries restricting the increase of the storage capacity are defined.
Keywords :
Hopfield neural nets; pattern recognition; Hopfield model; correlated patterns recognition; parametrical neural network; storage capacity; Associative memory; Character recognition; Decorrelation; Estimation theory; Neural networks; Neurons; Optical computing; Optical distortion; Pattern recognition; Random variables;
Conference_Titel :
Neural Networks, 2005. IJCNN '05. Proceedings. 2005 IEEE International Joint Conference on
Print_ISBN :
0-7803-9048-2
DOI :
10.1109/IJCNN.2005.1555993
