Multi-Branch Structure and its Localized Property in Layered Neural Networks

Takashi Yamashita, Kotaro Hirasawa, Takayuki Furuzuki

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

Neural networks (NNs) can solve only a simple problem if the network size is too small, on the other hand, if the network size increases, it costs a lot in terms of memory space and calculation time. So, we have studied how to construct the network structure with high performances and low costs in space and time. A solution is a multi-branch structure. Conventional NNs use the single-branch for the connections, while the multi-branch structure has multi-branches between nodes. In this paper, a new method which enables the multi-branch NNs to have localized property is proposed. It is well known that RBF networks have localized property that makes it possible to approximate functions faster than sigmoidal NNs. By using the multi-branch structure having localized property of RBF networks, NNs could obtain high performances keeping the lower costs in space and time. Simulation results of function approximations and a classification problem illustrated the effectiveness of multi-branch NNs.

Original languageEnglish
Pages (from-to)941-947
Number of pages7
JournalIEEJ Transactions on Electronics, Information and Systems
Volume125
Issue number6
DOIs
Publication statusPublished - 2005 Jan 1

Fingerprint

Neural networks
Radial basis function networks
Costs
Data storage equipment

Keywords

  • Localized property
  • Multi-branch structure
  • Neural networks

ASJC Scopus subject areas

  • Electrical and Electronic Engineering

Cite this

Multi-Branch Structure and its Localized Property in Layered Neural Networks. / Yamashita, Takashi; Hirasawa, Kotaro; Furuzuki, Takayuki.

In: IEEJ Transactions on Electronics, Information and Systems, Vol. 125, No. 6, 01.01.2005, p. 941-947.

Research output: Contribution to journalArticle

@article{21bfa281dade44bbbadfc0ca8fed2544,
title = "Multi-Branch Structure and its Localized Property in Layered Neural Networks",
abstract = "Neural networks (NNs) can solve only a simple problem if the network size is too small, on the other hand, if the network size increases, it costs a lot in terms of memory space and calculation time. So, we have studied how to construct the network structure with high performances and low costs in space and time. A solution is a multi-branch structure. Conventional NNs use the single-branch for the connections, while the multi-branch structure has multi-branches between nodes. In this paper, a new method which enables the multi-branch NNs to have localized property is proposed. It is well known that RBF networks have localized property that makes it possible to approximate functions faster than sigmoidal NNs. By using the multi-branch structure having localized property of RBF networks, NNs could obtain high performances keeping the lower costs in space and time. Simulation results of function approximations and a classification problem illustrated the effectiveness of multi-branch NNs.",
keywords = "Localized property, Multi-branch structure, Neural networks",
author = "Takashi Yamashita and Kotaro Hirasawa and Takayuki Furuzuki",
year = "2005",
month = "1",
day = "1",
doi = "10.1541/ieejeiss.125.941",
language = "English",
volume = "125",
pages = "941--947",
journal = "IEEJ Transactions on Electronics, Information and Systems",
issn = "0385-4221",
publisher = "The Institute of Electrical Engineers of Japan",
number = "6",

}

TY - JOUR

T1 - Multi-Branch Structure and its Localized Property in Layered Neural Networks

AU - Yamashita, Takashi

AU - Hirasawa, Kotaro

AU - Furuzuki, Takayuki

PY - 2005/1/1

Y1 - 2005/1/1

N2 - Neural networks (NNs) can solve only a simple problem if the network size is too small, on the other hand, if the network size increases, it costs a lot in terms of memory space and calculation time. So, we have studied how to construct the network structure with high performances and low costs in space and time. A solution is a multi-branch structure. Conventional NNs use the single-branch for the connections, while the multi-branch structure has multi-branches between nodes. In this paper, a new method which enables the multi-branch NNs to have localized property is proposed. It is well known that RBF networks have localized property that makes it possible to approximate functions faster than sigmoidal NNs. By using the multi-branch structure having localized property of RBF networks, NNs could obtain high performances keeping the lower costs in space and time. Simulation results of function approximations and a classification problem illustrated the effectiveness of multi-branch NNs.

AB - Neural networks (NNs) can solve only a simple problem if the network size is too small, on the other hand, if the network size increases, it costs a lot in terms of memory space and calculation time. So, we have studied how to construct the network structure with high performances and low costs in space and time. A solution is a multi-branch structure. Conventional NNs use the single-branch for the connections, while the multi-branch structure has multi-branches between nodes. In this paper, a new method which enables the multi-branch NNs to have localized property is proposed. It is well known that RBF networks have localized property that makes it possible to approximate functions faster than sigmoidal NNs. By using the multi-branch structure having localized property of RBF networks, NNs could obtain high performances keeping the lower costs in space and time. Simulation results of function approximations and a classification problem illustrated the effectiveness of multi-branch NNs.

KW - Localized property

KW - Multi-branch structure

KW - Neural networks

UR - http://www.scopus.com/inward/record.url?scp=62249189598&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=62249189598&partnerID=8YFLogxK

U2 - 10.1541/ieejeiss.125.941

DO - 10.1541/ieejeiss.125.941

M3 - Article

AN - SCOPUS:62249189598

VL - 125

SP - 941

EP - 947

JO - IEEJ Transactions on Electronics, Information and Systems

JF - IEEJ Transactions on Electronics, Information and Systems

SN - 0385-4221

IS - 6

ER -