Share Email Print
cover

Proceedings Paper

Equivalent models of neural networks and their effective optoelectronic implementations based on matrix multivalued elements
Author(s): Vladimir G. Krasilenko; Anatoly K. Bogukhvalskiy; Andrey T. Magas
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

The theory and equivalental models of neural networks based on equivalence operation (non-equivalence) of continuous and multivalued neural logic are considered. Their connection with metric of metric-address spaces are shown. Normalized equivalencies of vectors with multilevel components are determined. Equivalental models for simple network with weighted correlation coefficients, for network with adapted weighing and double weighing are suggested. It is shown, that the network model with double weighing (adapted and correlation coefficients) being most generalized can also conduct the recalculation process of networks to two-step algorithms without calculation of connections matrix. Equivalental models require calculations based on vector- matrix procedures with equivalence operation and can be realized on vector-matrix equivalentors with space and time integration. The apparatus implementations of models with productivity of 108 divided by 109 connections/sec and neuron number 256 and more are suggested.

Paper Details

Date Published: 28 February 1997
PDF: 10 pages
Proc. SPIE 3055, International Conference on Optical Storage, Imaging, and Transmission of Information, (28 February 1997); doi: 10.1117/12.267699
Show Author Affiliations
Vladimir G. Krasilenko, Collective scientific-industrial venture: Injector (Ukraine)
Anatoly K. Bogukhvalskiy, Collective scientific-industrial venture: Injector (Ukraine)
Andrey T. Magas, Collective scientific-industrial venture: Injector (Ukraine)


Published in SPIE Proceedings Vol. 3055:
International Conference on Optical Storage, Imaging, and Transmission of Information
Viacheslav V. Petrov; Sergei V. Svechnikov, Editor(s)

© SPIE. Terms of Use
Back to Top