A memory optimal BFGS neural network training algorithm

Author

McLoone, Sein F. ; Asirvadam, Vijanth S. ; Irwin, George W.

Author_Institution

Intelligent Syst. & Control Group, Queen´´s Univ., Belfast, UK

Volume

1

fYear

2002

fDate

6/24/1905 12:00:00 AM

Firstpage

513

Lastpage

518

Abstract

This paper considers the implementation of a novel memory optimal neural network training algorithm which maximises performance in relation to available memory. Mathematically, it is similar to the full memory BFGS training when there are no constraints on memory and to the variable memory (VM) BFGS when memory is limited. However, it requires less computations per iteration than VM and uses a much better strategy for discarding old curvature information when memory is limited

Keywords

computational complexity; content-addressable storage; feedforward neural nets; learning (artificial intelligence); optimisation; BFGS neural network; computational complexity; feedforward neural networks; full memory; learning algorithm; multilayer perceptrons; optimization; variable memory; Control systems; Cost function; Feedforward neural networks; Intelligent control; Intelligent structures; Intelligent systems; Memory management; Multilayer perceptrons; Neural networks; Virtual manufacturing;

fLanguage

English

Publisher

ieee

Conference_Titel

Neural Networks, 2002. IJCNN '02. Proceedings of the 2002 International Joint Conference on

Conference_Location

Honolulu, HI

ISSN

1098-7576

Print_ISBN

0-7803-7278-6

Type

conf

DOI

10.1109/IJCNN.2002.1005525

Filename

1005525