Stability and dynamics of simple electronic neural networks with added inertia

We examine systems of one and two nonlinear threshold elements, or neurons, of the kind used in electronic neural networks. When the neuron couplings are of an inertial nature it is found that the dynamics can be complex, in contrast to the simpler behavior displayed when they are of the standard re...

Full description

Saved in:
Bibliographic Details
Published in:Physica. D 1986-12, Vol.23 (1), p.464-469
Main Authors: Babcock, K.L., Westervelt, R.M.
Format: Article
Language:English
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:We examine systems of one and two nonlinear threshold elements, or neurons, of the kind used in electronic neural networks. When the neuron couplings are of an inertial nature it is found that the dynamics can be complex, in contrast to the simpler behavior displayed when they are of the standard resistor-capacitor variety. For various values of the neuron gain and the quality factor of the couplings we find ringing about the stationary points, instability and spontaneous oscillation, intertwined basins of attraction, and chaotic response to a harmonic drive. Furthermore, the collective transient response of a network of multiple neurons may be underdamped even when the neurons or connections are overdamped. These results imply that care should be exercised when adding inertia to neural couplings to enhance performance of optimizing networks.
ISSN:0167-2789
1872-8022
DOI:10.1016/0167-2789(86)90152-1