A calculation of neural network statistics: the probability distribution of average cell firing in a feedforward neural network

Author

Levy, William B.

Author_Institution

Dept. of Neurosurg., Virginia Univ., Charlottesville, VA, USA

Volume

3

fYear

1992

fDate

7-11 Jun 1992

Firstpage

341

Abstract

The author describes a relationship between the statistics of the input environment and the distribution of average postsynaptic excitation for a feedforward network governed by a particular associative modification rule. In some circumstances, which are illustrated, this calculation implies a normal distribution. Only summary statistics of the pairwise correlations of the input environment are necessary to produce an accurate estimate of the average cell firing distribution. An application of this estimate to the design of a preprocessing network is described

Keywords

feedforward neural nets; probability; average cell firing; average postsynaptic excitation; feedforward neural network; neural network statistics; pairwise correlations; preprocessing network; probability distribution; Equations; Feedforward neural networks; Higher order statistics; Information processing; Intelligent networks; Neural networks; Neurons; Probability distribution; Statistical distributions; Vectors;

fLanguage

English

Publisher

ieee

Conference_Titel

Neural Networks, 1992. IJCNN., International Joint Conference on

Conference_Location

Baltimore, MD

Print_ISBN

0-7803-0559-0

Type

conf

DOI

10.1109/IJCNN.1992.227151

Filename

227151