New method to prune the neural network

Weishui Wan, Kotaro Hirasawa, Takayuki Furuzuki, Chunzhi Jin

Research output: Contribution to journalArticle

Abstract

Using backpropagation algorithm (BP) to train neural networks is a widely adopted practice in both theory and practical applications. But its distributed weight representation, that is the weight matrix of final network after training by using BP are usually not sparsified, and prohibits its use in the rule discovery of inherent functional relations between the input and output data, so in this aspect some kinds of structure optimization are needed to improve its poor performance. In this paper with this in mind a new method to prune neural networks is proposed based on some statistical quantities of neural networks. Comparing with the other known pruning methods such as structural learning with forgetting (SLF) and RPROP algorithm, the proposed method can attain comparable or even better results over these methods without evident increase of the computational load. Detailed simulations using the Iris data sets exhibit our above assertion.

Original languageEnglish
Pages (from-to)43-49
Number of pages7
JournalResearch Reports on Information Science and Electrical Engineering of Kyushu University
Volume5
Issue number1
Publication statusPublished - 2000 Mar
Externally publishedYes

Fingerprint

Backpropagation algorithms
Neural networks

ASJC Scopus subject areas

  • Hardware and Architecture
  • Engineering (miscellaneous)
  • Electrical and Electronic Engineering

Cite this

New method to prune the neural network. / Wan, Weishui; Hirasawa, Kotaro; Furuzuki, Takayuki; Jin, Chunzhi.

In: Research Reports on Information Science and Electrical Engineering of Kyushu University, Vol. 5, No. 1, 03.2000, p. 43-49.

Research output: Contribution to journalArticle

@article{802206c68003498ea94331b5c3d8914e,
title = "New method to prune the neural network",
abstract = "Using backpropagation algorithm (BP) to train neural networks is a widely adopted practice in both theory and practical applications. But its distributed weight representation, that is the weight matrix of final network after training by using BP are usually not sparsified, and prohibits its use in the rule discovery of inherent functional relations between the input and output data, so in this aspect some kinds of structure optimization are needed to improve its poor performance. In this paper with this in mind a new method to prune neural networks is proposed based on some statistical quantities of neural networks. Comparing with the other known pruning methods such as structural learning with forgetting (SLF) and RPROP algorithm, the proposed method can attain comparable or even better results over these methods without evident increase of the computational load. Detailed simulations using the Iris data sets exhibit our above assertion.",
author = "Weishui Wan and Kotaro Hirasawa and Takayuki Furuzuki and Chunzhi Jin",
year = "2000",
month = "3",
language = "English",
volume = "5",
pages = "43--49",
journal = "Research Reports on Information Science and Electrical Engineering of Kyushu University",
issn = "1342-3819",
publisher = "Kyushu University, Faculty of Science",
number = "1",

}

TY - JOUR

T1 - New method to prune the neural network

AU - Wan, Weishui

AU - Hirasawa, Kotaro

AU - Furuzuki, Takayuki

AU - Jin, Chunzhi

PY - 2000/3

Y1 - 2000/3

N2 - Using backpropagation algorithm (BP) to train neural networks is a widely adopted practice in both theory and practical applications. But its distributed weight representation, that is the weight matrix of final network after training by using BP are usually not sparsified, and prohibits its use in the rule discovery of inherent functional relations between the input and output data, so in this aspect some kinds of structure optimization are needed to improve its poor performance. In this paper with this in mind a new method to prune neural networks is proposed based on some statistical quantities of neural networks. Comparing with the other known pruning methods such as structural learning with forgetting (SLF) and RPROP algorithm, the proposed method can attain comparable or even better results over these methods without evident increase of the computational load. Detailed simulations using the Iris data sets exhibit our above assertion.

AB - Using backpropagation algorithm (BP) to train neural networks is a widely adopted practice in both theory and practical applications. But its distributed weight representation, that is the weight matrix of final network after training by using BP are usually not sparsified, and prohibits its use in the rule discovery of inherent functional relations between the input and output data, so in this aspect some kinds of structure optimization are needed to improve its poor performance. In this paper with this in mind a new method to prune neural networks is proposed based on some statistical quantities of neural networks. Comparing with the other known pruning methods such as structural learning with forgetting (SLF) and RPROP algorithm, the proposed method can attain comparable or even better results over these methods without evident increase of the computational load. Detailed simulations using the Iris data sets exhibit our above assertion.

UR - http://www.scopus.com/inward/record.url?scp=0033750072&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0033750072&partnerID=8YFLogxK

M3 - Article

VL - 5

SP - 43

EP - 49

JO - Research Reports on Information Science and Electrical Engineering of Kyushu University

JF - Research Reports on Information Science and Electrical Engineering of Kyushu University

SN - 1342-3819

IS - 1

ER -