Fast design algorithm for competitive learning

Wen Jyi Hwang, Faa Jeng Lin, Yi Chong Zeng

Research output: Contribution to journalArticle

8 Citations (Scopus)

Abstract

The authors present a new, fast computation algorithm which significantly reduces the training time of the competitive learning (CL) algorithm without sacrificing its performance. The training process is performed in the wavelet domain to identify the winning neural units with a small number of multiplications. Simulation results show that in many cases the algorithm requires < 10% of the computational time required by traditional CL algorithms.

Original languageEnglish
Pages (from-to)1469-1471
Number of pages3
JournalElectronics Letters
Volume33
Issue number17
Publication statusPublished - 1997 Aug 14
Externally publishedYes

Fingerprint

Learning algorithms

Keywords

  • Learning systems
  • Neural networks
  • Wavelet transforms

ASJC Scopus subject areas

  • Electrical and Electronic Engineering

Cite this

Hwang, W. J., Lin, F. J., & Zeng, Y. C. (1997). Fast design algorithm for competitive learning. Electronics Letters, 33(17), 1469-1471.

Fast design algorithm for competitive learning. / Hwang, Wen Jyi; Lin, Faa Jeng; Zeng, Yi Chong.

In: Electronics Letters, Vol. 33, No. 17, 14.08.1997, p. 1469-1471.

Research output: Contribution to journalArticle

Hwang, WJ, Lin, FJ & Zeng, YC 1997, 'Fast design algorithm for competitive learning', Electronics Letters, vol. 33, no. 17, pp. 1469-1471.
Hwang WJ, Lin FJ, Zeng YC. Fast design algorithm for competitive learning. Electronics Letters. 1997 Aug 14;33(17):1469-1471.
Hwang, Wen Jyi ; Lin, Faa Jeng ; Zeng, Yi Chong. / Fast design algorithm for competitive learning. In: Electronics Letters. 1997 ; Vol. 33, No. 17. pp. 1469-1471.
@article{fc8b45db5645422db2708f963cf886b0,
title = "Fast design algorithm for competitive learning",
abstract = "The authors present a new, fast computation algorithm which significantly reduces the training time of the competitive learning (CL) algorithm without sacrificing its performance. The training process is performed in the wavelet domain to identify the winning neural units with a small number of multiplications. Simulation results show that in many cases the algorithm requires < 10{\%} of the computational time required by traditional CL algorithms.",
keywords = "Learning systems, Neural networks, Wavelet transforms",
author = "Hwang, {Wen Jyi} and Lin, {Faa Jeng} and Zeng, {Yi Chong}",
year = "1997",
month = "8",
day = "14",
language = "English",
volume = "33",
pages = "1469--1471",
journal = "Electronics Letters",
issn = "0013-5194",
publisher = "Institution of Engineering and Technology",
number = "17",

}

TY - JOUR

T1 - Fast design algorithm for competitive learning

AU - Hwang, Wen Jyi

AU - Lin, Faa Jeng

AU - Zeng, Yi Chong

PY - 1997/8/14

Y1 - 1997/8/14

N2 - The authors present a new, fast computation algorithm which significantly reduces the training time of the competitive learning (CL) algorithm without sacrificing its performance. The training process is performed in the wavelet domain to identify the winning neural units with a small number of multiplications. Simulation results show that in many cases the algorithm requires < 10% of the computational time required by traditional CL algorithms.

AB - The authors present a new, fast computation algorithm which significantly reduces the training time of the competitive learning (CL) algorithm without sacrificing its performance. The training process is performed in the wavelet domain to identify the winning neural units with a small number of multiplications. Simulation results show that in many cases the algorithm requires < 10% of the computational time required by traditional CL algorithms.

KW - Learning systems

KW - Neural networks

KW - Wavelet transforms

UR - http://www.scopus.com/inward/record.url?scp=0031199294&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0031199294&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:0031199294

VL - 33

SP - 1469

EP - 1471

JO - Electronics Letters

JF - Electronics Letters

SN - 0013-5194

IS - 17

ER -