Neural network training using stochastic PSO

Xin Chen; Yangmin Li

Neural network training using stochastic PSO

Research output: Chapter in book / Conference proceeding › Conference article published in proceeding or book › Academic research › peer-review

2 Citations (Scopus)

Abstract

Particle swarm optimization is widely applied for training neural network. Since in many applications the number of weights of NN is huge, when PSO algorithms are applied for NN training, the dimension of search space is so large that PSOs always converge prematurely. In this paper an improved stochastic PSO (SPSO) is presented, to which a random velocity is added to improve particles' exploration ability. Since SPSO explores much thoroughly to collect information of solution space, it is able to find the global best solution with high opportunity. Hence SPSO is suitable for optimization about high dimension problems, especially for NN training.

Original language	English
Title of host publication	Neural Information Processing - 13th International Conference, ICONIP 2006, Proceedings
Publisher	Springer Verlag
Pages	1051-1060
Number of pages	10
ISBN (Print)	3540464816, 9783540464815
Publication status	Published - 1 Jan 2006
Externally published	Yes
Event	13th International Conference on Neural Information Processing, ICONIP 2006 - Hong Kong, Hong Kong Duration: 3 Oct 2006 → 6 Oct 2006

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	4233 LNCS - II
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	13th International Conference on Neural Information Processing, ICONIP 2006
Country/Territory	Hong Kong
City	Hong Kong
Period	3/10/06 → 6/10/06

ASJC Scopus subject areas

General Computer Science
Theoretical Computer Science

Cite this

@inproceedings{65c27a1972bf4550ad6ded7eb96c2ee6,

title = "Neural network training using stochastic PSO",

abstract = "Particle swarm optimization is widely applied for training neural network. Since in many applications the number of weights of NN is huge, when PSO algorithms are applied for NN training, the dimension of search space is so large that PSOs always converge prematurely. In this paper an improved stochastic PSO (SPSO) is presented, to which a random velocity is added to improve particles' exploration ability. Since SPSO explores much thoroughly to collect information of solution space, it is able to find the global best solution with high opportunity. Hence SPSO is suitable for optimization about high dimension problems, especially for NN training.",

author = "Xin Chen and Yangmin Li",

year = "2006",

month = jan,

day = "1",

language = "English",

isbn = "3540464816",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "1051--1060",

booktitle = "Neural Information Processing - 13th International Conference, ICONIP 2006, Proceedings",

address = "Germany",

note = "13th International Conference on Neural Information Processing, ICONIP 2006 ; Conference date: 03-10-2006 Through 06-10-2006",

}

Chen, X & Li, Y 2006, Neural network training using stochastic PSO. in Neural Information Processing - 13th International Conference, ICONIP 2006, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 4233 LNCS - II, Springer Verlag, pp. 1051-1060, 13th International Conference on Neural Information Processing, ICONIP 2006, Hong Kong, Hong Kong, 3/10/06.

Neural network training using stochastic PSO. / Chen, Xin; Li, Yangmin.
Neural Information Processing - 13th International Conference, ICONIP 2006, Proceedings. Springer Verlag, 2006. p. 1051-1060 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4233 LNCS - II).

Research output: Chapter in book / Conference proceeding › Conference article published in proceeding or book › Academic research › peer-review

TY - GEN

T1 - Neural network training using stochastic PSO

AU - Chen, Xin

AU - Li, Yangmin

PY - 2006/1/1

Y1 - 2006/1/1

N2 - Particle swarm optimization is widely applied for training neural network. Since in many applications the number of weights of NN is huge, when PSO algorithms are applied for NN training, the dimension of search space is so large that PSOs always converge prematurely. In this paper an improved stochastic PSO (SPSO) is presented, to which a random velocity is added to improve particles' exploration ability. Since SPSO explores much thoroughly to collect information of solution space, it is able to find the global best solution with high opportunity. Hence SPSO is suitable for optimization about high dimension problems, especially for NN training.

AB - Particle swarm optimization is widely applied for training neural network. Since in many applications the number of weights of NN is huge, when PSO algorithms are applied for NN training, the dimension of search space is so large that PSOs always converge prematurely. In this paper an improved stochastic PSO (SPSO) is presented, to which a random velocity is added to improve particles' exploration ability. Since SPSO explores much thoroughly to collect information of solution space, it is able to find the global best solution with high opportunity. Hence SPSO is suitable for optimization about high dimension problems, especially for NN training.

UR - http://www.scopus.com/inward/record.url?scp=33750736050&partnerID=8YFLogxK

M3 - Conference article published in proceeding or book

SN - 3540464816

SN - 9783540464815

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 1051

EP - 1060

BT - Neural Information Processing - 13th International Conference, ICONIP 2006, Proceedings

PB - Springer Verlag

T2 - 13th International Conference on Neural Information Processing, ICONIP 2006

Y2 - 3 October 2006 through 6 October 2006

ER -

Neural network training using stochastic PSO

Abstract

Publication series

Conference

ASJC Scopus subject areas

Other files and links

Fingerprint

Cite this