DYNAMIC PROPERTIES OF A NEURAL NETWORK MODEL OF SPATIAL MEMORY

V. A. Lyakhovetskiy, A. S. Potapov


Read the full article 

Abstract

 

A model for movements sequences memorizing (spatial positions) based on the hetero-associative neural network is considered. An estimating criterion for correctness of memorized data is proposed as the number of iterations necessary for transition into some steady state. This criterion is in good agreement with psychophysiological facts and can be utilized for unsupervised training of the neural network model. Interpretation of different error types from the position of nonlinear dynamics is given.


Keywords: hetero-associative neural networks, working storage, repeating errors.
Copyright 2001-2017 ©
Scientific and Technical Journal
of Information Technologies, Mechanics and Optics.
All rights reserved.

Яндекс.Метрика