Evolutionary artificial neural networks

Author

Brown, A.D. ; Card, H.C.

Author_Institution

Dept. of Electr. & Comput. Eng., Manitoba Univ., Winnipeg, Man., Canada

Volume

1

fYear

1997

fDate

25-28 May 1997

Firstpage

313

Abstract

We present experiments which show that a genetic algorithm (GA) can effectively search for a set of local feature detectors, which can be used by higher neural network layers to perform an image classification task. Three different methods of encoding hidden unit weights into the GA are presented, including one which coevolves all the feature detectors in a single chromosome, and two which promote the cooperation of feature detectors by encoding them in their own chromosome. The fitness function measures the classification percentage and confidence of the networks on validation data in order to encourage generalization

Keywords

encoding; feature extraction; feedforward neural nets; generalisation (artificial intelligence); genetic algorithms; image classification; chromosome; encoding; evolutionary neural networks; feature detectors; fitness function; generalization; genetic algorithm; hidden unit weights; image classification; multilayer neural nets; Artificial neural networks; Biological cells; Computer vision; Concatenated codes; Detectors; Encoding; Genetic algorithms; Image classification; Neural networks; Stochastic processes;

fLanguage

English

Publisher

ieee

Conference_Titel

Electrical and Computer Engineering, 1997. Engineering Innovation: Voyage of Discovery. IEEE 1997 Canadian Conference on

Conference_Location

St. Johns, Nfld.

ISSN

0840-7789

Print_ISBN

0-7803-3716-6

Type

conf

DOI

10.1109/CCECE.1997.614852

Filename

614852