A neural computation for canonical representations of nonlinear functions

Author

Huang, Qiu ; Liu, Ruey-Wen

Author_Institution

Dept. of Electr. & Comput. Eng., Notre Dame Univ., IN, USA

fYear

1990

fDate

1-3 May 1990

Firstpage

2504

Abstract

An efficient constructive method to compute a canonical representation of any continuous nonlinear function with a neural network is presented. This neural network consists of only one hidden layer and the number of nodes in the hidden layer is estimated. A particular initial condition is chosen so that the backpropagation algorithm converges to a solution with an error between the neural network realization and the given function less than a given δ>0. Two simulation examples are used to demonstrate the method, and a simple application of this method to the design of a negative device is given

Keywords

convergence of numerical methods; curve fitting; iterative methods; neural nets; any continuous nonlinear function; backpropagation algorithm convergence; canonical representations of nonlinear functions; constructive method; negative resistance devices design; neural computation; neural network; number of nodes; one hidden layer; simulation examples; Backpropagation algorithms; Circuits; Computer networks; Equations; Fitting; Manufacturing; Multilayer perceptrons; Neural networks; Neurons; Reliability engineering;

fLanguage

English

Publisher

ieee

Conference_Titel

Circuits and Systems, 1990., IEEE International Symposium on

Conference_Location

New Orleans, LA

Type

conf

DOI

10.1109/ISCAS.1990.112519

Filename

112519