Global feedforward neural network learning for classification and regression

Kar Ann Toh, Juwei Lu, Wei Yun Yau

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Citations (Scopus)

Abstract

This paper addresses the issues of global optimality and training of a Feedforward Neural Network (FNN) error funtion incoporating the weight decay regularizer. A network with a single hidden-layer and a single output-unit is considered. Explicit vector and matrix canonical forms for the Jacobian and Hessian of the network are presented. Convexity analysis is then performed utilizing the known canonical structure of the Hessian. Next, global optimality characterization of the FNN error function is attempted utilizing the results of convex characterization and a convex monotonic transformation. Based on this global optimality characterization, an iterative algorithm is proposed for global FNN learning. Numerical experiments with benchmark examples show better convergence of our network learning as compared to many existing methods in the literature. The network is also shown to gener­alize well for a face recognition problem.

Original languageEnglish
Title of host publicationEnergy Minimization Methods in Computer Vision and Pattern Recognition - 3rd International Workshop, EMMCVPR 2001, Proceedings
EditorsAnil K. Jain, Mario Figueiredo, Josiane Zerubia
PublisherSpringer Verlag
Pages407-422
Number of pages16
ISBN (Print)3540425233, 9783540425236
DOIs
Publication statusPublished - 2001 Jan 1
Event3rd International Workshop on Energy Minimization Methods in Computer Vision and Pattern Recognition, EMMCVPR 2001 - Sophia Antipolis, France
Duration: 2001 Sep 32001 Sep 5

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume2134
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other3rd International Workshop on Energy Minimization Methods in Computer Vision and Pattern Recognition, EMMCVPR 2001
CountryFrance
CitySophia Antipolis
Period01/9/301/9/5

    Fingerprint

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Toh, K. A., Lu, J., & Yau, W. Y. (2001). Global feedforward neural network learning for classification and regression. In A. K. Jain, M. Figueiredo, & J. Zerubia (Eds.), Energy Minimization Methods in Computer Vision and Pattern Recognition - 3rd International Workshop, EMMCVPR 2001, Proceedings (pp. 407-422). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 2134). Springer Verlag. https://doi.org/10.1007/3-540-44745-8_27