An incremental network construction algorithm for approximating discontinuous functions

Author

Lee, Hyukjoon ; Mehrotra, Kishan ; Mohan, Chilukuri K. ; Ranka, Sanjay

Author_Institution

Sch. of Comput. & Inf. Sci., Syracuse Univ., NY, USA

Volume

4

fYear

1994

fDate

27 Jun-2 Jul 1994

Firstpage

2191

Abstract

Traditional neural network training techniques do not work well on problems with many discontinuities, such as those that arise in multicomputer communication cost modeling. We develop a new algorithm to solve this problem. This algorithm incrementally adds modules to the network, successively expanding the `window´ in the data space where the current module works well. The need for a new module is automatically recognized by the system. This algorithm performs very well on problems with many discontinuities, and requires fewer computations than traditional backpropagation

Keywords

approximation theory; feedforward neural nets; function approximation; functional analysis; learning (artificial intelligence); attentive modular construction and training algorithm; data space; discontinuous function approximation; incremental network construction algorithm; learning algorithm; modules; neural network; Approximation algorithms; Backpropagation algorithms; Computational Intelligence Society; Computer networks; Cost function; Feedforward neural networks; Information science; Multi-layer neural network; Neural networks; Surface reconstruction;

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.1994.374556

Filename

374556