A Hybrid global learning algorithm based on global search and least squares techniques for backpropagation networks

Chi-Tat Leung; Tommy W. S. Chow

doi:10.1109/ICNN.1997.614187

A Hybrid global learning algorithm based on global search and least squares techniques for backpropagation networks

Chi-Tat Leung, Tommy W. S. Chow

Department of Electrical Engineering

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

16 Citations (Scopus)

Abstract

A hybrid learning algorithm for backpropagation network based on global search and least squares methods is presented to speed up the speed of convergence. The proposed algorithm comprises global search and least squares parts. The global search part trains a backpropagation network over a reduced weight space. The remained weights are calculated in accordance with linear least squares method. Two problems of nonlinear function approximation and modified XOR are applied to demonstrate the fast global search performance of the proposed algorithm. The results indicate that the proposed algorithm enables the learning process to significantly speed up by at most 4670 % in terms of iterations and do not trap in local minima. © 1997 IEEE

Original language	English
Title of host publication	The 1997 IEEE International Conference on Neural Networks Proceedings
Publisher	IEEE
Pages	1890-1895
Volume	3
ISBN (Print)	0-7803-4122-8
DOIs	https://doi.org/10.1109/ICNN.1997.614187
Publication status	Published - Jun 1997
Event	1997 IEEE International Conference on Neural Networks (ICNN'97) - Westin Galleria Hotel, Houston, Texas, United States Duration: 9 Jun 1997 → 12 Jun 1997

Conference

Conference	1997 IEEE International Conference on Neural Networks (ICNN'97)
Abbreviated title	ICNN'97
Place	United States
City	Houston, Texas
Period	9/06/97 → 12/06/97

Access Output

10.1109/ICNN.1997.614187

Other Access Options

Check CityUHK Library

Permanent Link

Right click to copy link

Cite this

@inproceedings{9cf898639ef744e9aec6aa6e796aee4c,

title = "A Hybrid global learning algorithm based on global search and least squares techniques for backpropagation networks",

abstract = "A hybrid learning algorithm for backpropagation network based on global search and least squares methods is presented to speed up the speed of convergence. The proposed algorithm comprises global search and least squares parts. The global search part trains a backpropagation network over a reduced weight space. The remained weights are calculated in accordance with linear least squares method. Two problems of nonlinear function approximation and modified XOR are applied to demonstrate the fast global search performance of the proposed algorithm. The results indicate that the proposed algorithm enables the learning process to significantly speed up by at most 4670 % in terms of iterations and do not trap in local minima. {\textcopyright} 1997 IEEE",

author = "Chi-Tat Leung and Chow, {Tommy W. S.}",

year = "1997",

month = jun,

doi = "10.1109/ICNN.1997.614187",

language = "English",

isbn = "0-7803-4122-8",

volume = "3",

pages = "1890--1895",

booktitle = "The 1997 IEEE International Conference on Neural Networks Proceedings",

publisher = "IEEE",

address = "United States",

note = "1997 IEEE International Conference on Neural Networks (ICNN'97), ICNN'97 ; Conference date: 09-06-1997 Through 12-06-1997",

}

Leung, C-T & Chow, TWS 1997, A Hybrid global learning algorithm based on global search and least squares techniques for backpropagation networks. in The 1997 IEEE International Conference on Neural Networks Proceedings . vol. 3, IEEE, pp. 1890-1895, 1997 IEEE International Conference on Neural Networks (ICNN'97), Houston, Texas, United States, 9/06/97. https://doi.org/10.1109/ICNN.1997.614187

A Hybrid global learning algorithm based on global search and least squares techniques for backpropagation networks. / Leung, Chi-Tat; Chow, Tommy W. S.
The 1997 IEEE International Conference on Neural Networks Proceedings . Vol. 3 IEEE, 1997. p. 1890-1895.

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

TY - GEN

T1 - A Hybrid global learning algorithm based on global search and least squares techniques for backpropagation networks

AU - Leung, Chi-Tat

AU - Chow, Tommy W. S.

PY - 1997/6

Y1 - 1997/6

N2 - A hybrid learning algorithm for backpropagation network based on global search and least squares methods is presented to speed up the speed of convergence. The proposed algorithm comprises global search and least squares parts. The global search part trains a backpropagation network over a reduced weight space. The remained weights are calculated in accordance with linear least squares method. Two problems of nonlinear function approximation and modified XOR are applied to demonstrate the fast global search performance of the proposed algorithm. The results indicate that the proposed algorithm enables the learning process to significantly speed up by at most 4670 % in terms of iterations and do not trap in local minima. © 1997 IEEE

AB - A hybrid learning algorithm for backpropagation network based on global search and least squares methods is presented to speed up the speed of convergence. The proposed algorithm comprises global search and least squares parts. The global search part trains a backpropagation network over a reduced weight space. The remained weights are calculated in accordance with linear least squares method. Two problems of nonlinear function approximation and modified XOR are applied to demonstrate the fast global search performance of the proposed algorithm. The results indicate that the proposed algorithm enables the learning process to significantly speed up by at most 4670 % in terms of iterations and do not trap in local minima. © 1997 IEEE

UR - http://www.scopus.com/inward/record.url?scp=0030658717&partnerID=8YFLogxK

UR - https://www.scopus.com/record/pubmetrics.uri?eid=2-s2.0-0030658717&origin=recordpage

U2 - 10.1109/ICNN.1997.614187

DO - 10.1109/ICNN.1997.614187

M3 - RGC 32 - Refereed conference paper (with host publication)

SN - 0-7803-4122-8

VL - 3

SP - 1890

EP - 1895

BT - The 1997 IEEE International Conference on Neural Networks Proceedings

PB - IEEE

T2 - 1997 IEEE International Conference on Neural Networks (ICNN'97)

Y2 - 9 June 1997 through 12 June 1997

ER -

A Hybrid global learning algorithm based on global search and least squares techniques for backpropagation networks

Abstract

Conference

Access Output

Other Access Options

Permanent Link

Other Files and Links

Fingerprint

Cite this