Quantizability and learning complexity in multilayer neural networks

Author

Fu, LiMin

Author_Institution

Dept. of Comput. & Inf. Sci., Florida Univ., Gainesville, FL, USA

Volume

28

Issue

2

fYear

1998

fDate

5/1/1998 12:00:00 AM

Firstpage

295

Lastpage

299

Abstract

The relationship between quantizability and learning complexity in multilayer neural networks is examined. In a special neural network architecture that calculates node activations according to the certainty factor (CF) model of expert systems, the analysis based upon quantizability leads to lower and also better estimates for generalization dimensionality and sample complexity than those suggested by the multilayer perceptron model. This analysis is further supported by empirical simulation results

Keywords

expert systems; feedforward neural nets; generalisation (artificial intelligence); learning (artificial intelligence); multilayer perceptrons; neural net architecture; quantisation (signal); simulation; certainty factor model; empirical simulation; expert systems; generalization dimensionality; learning complexity; multilayer neural networks; neural network architecture; node activations; quantizability; Analytical models; Computer networks; Degradation; Expert systems; Intelligent networks; Machine learning; Multi-layer neural network; Multilayer perceptrons; Neural networks; Quantization;

fLanguage

English

Journal_Title

Systems, Man, and Cybernetics, Part C: Applications and Reviews, IEEE Transactions on

Publisher

ieee

ISSN

1094-6977

Type

jour

DOI

10.1109/5326.669575

Filename

669575