A structural learning by adding independent noises to hidden units

Author

Kurita, T. ; Asoh, H. ; Umeyama, S. ; Akaho, S. ; Hosomi, A.

Author_Institution

Math. Inf. Section, Electrotech. Lab., Ibaraki, Japan

Volume

1

fYear

1994

fDate

27 Jun-2 Jul 1994

Firstpage

275

Abstract

The paper demonstrates that a skeletal structure of a network emerges when independent noises are added to the inputs of the hidden units of multilayer perceptron during the learning by error backpropagation. By analyzing the average behavior of the error backpropagation algorithm to such noises, it is shown that the weights from the hidden units to the output units tend to get smaller and the outputs of the hidden units tend to be 0 or 1. Such tendency have been demonstrated by experiments of learning of pattern classification problem

Keywords

backpropagation; multilayer perceptrons; pattern classification; average behavior; error backpropagation; hidden units; independent noises; multilayer perceptron; pattern classification problem; skeletal structure; structural learning; weights; Algorithm design and analysis; Backpropagation algorithms; Gaussian noise; Informatics; Linearity; Multilayer perceptrons; Noise figure; Noise measurement; Pattern classification;

fLanguage

English

Publisher

ieee

Conference_Titel

Neural Networks, 1994. IEEE World Congress on Computational Intelligence., 1994 IEEE International Conference on

Conference_Location

Orlando, FL

Print_ISBN

0-7803-1901-X

Type

conf

DOI

10.1109/ICNN.1994.374174

Filename

374174