Effective training methods for function localization neural networks

Takafumi Sasakawa, Takayuki Furuzuki, Katsunori Isono, Kotaro Hirasawa

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

Inspired by Hebb's cell assembly theory about how the brain worked, we have developed a function localization neural network (FLNN). The main part of a FLNN is structurally the same as an ordinary feedforward neural network, but it is considered to consist of several overlapping modules, which are switched according to input patterns. A FLNN constructed in this way has been shown to have better representation ability than an ordinary neural network. However, BP training algorithm for such FLNN is very easy to get stuck at a local minimum. In this paper, we mainly discuss the methods for improving BP training of the FLNN by utilizing the structural property of the network. Two methods are proposed. Numerical simulations are used to show the effectiveness of the improved BP training methods.

Original languageEnglish
Title of host publicationIEEE International Conference on Neural Networks - Conference Proceedings
Pages4785-4790
Number of pages6
Publication statusPublished - 2006
EventInternational Joint Conference on Neural Networks 2006, IJCNN '06 - Vancouver, BC
Duration: 2006 Jul 162006 Jul 21

Other

OtherInternational Joint Conference on Neural Networks 2006, IJCNN '06
CityVancouver, BC
Period06/7/1606/7/21

Fingerprint

Neural networks
Feedforward neural networks
Structural properties
Brain
Computer simulation

ASJC Scopus subject areas

  • Software

Cite this

Sasakawa, T., Furuzuki, T., Isono, K., & Hirasawa, K. (2006). Effective training methods for function localization neural networks. In IEEE International Conference on Neural Networks - Conference Proceedings (pp. 4785-4790). [1716764]

Effective training methods for function localization neural networks. / Sasakawa, Takafumi; Furuzuki, Takayuki; Isono, Katsunori; Hirasawa, Kotaro.

IEEE International Conference on Neural Networks - Conference Proceedings. 2006. p. 4785-4790 1716764.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Sasakawa, T, Furuzuki, T, Isono, K & Hirasawa, K 2006, Effective training methods for function localization neural networks. in IEEE International Conference on Neural Networks - Conference Proceedings., 1716764, pp. 4785-4790, International Joint Conference on Neural Networks 2006, IJCNN '06, Vancouver, BC, 06/7/16.
Sasakawa T, Furuzuki T, Isono K, Hirasawa K. Effective training methods for function localization neural networks. In IEEE International Conference on Neural Networks - Conference Proceedings. 2006. p. 4785-4790. 1716764
Sasakawa, Takafumi ; Furuzuki, Takayuki ; Isono, Katsunori ; Hirasawa, Kotaro. / Effective training methods for function localization neural networks. IEEE International Conference on Neural Networks - Conference Proceedings. 2006. pp. 4785-4790
@inproceedings{0650463d1b224360af9780ca1bb38c7d,
title = "Effective training methods for function localization neural networks",
abstract = "Inspired by Hebb's cell assembly theory about how the brain worked, we have developed a function localization neural network (FLNN). The main part of a FLNN is structurally the same as an ordinary feedforward neural network, but it is considered to consist of several overlapping modules, which are switched according to input patterns. A FLNN constructed in this way has been shown to have better representation ability than an ordinary neural network. However, BP training algorithm for such FLNN is very easy to get stuck at a local minimum. In this paper, we mainly discuss the methods for improving BP training of the FLNN by utilizing the structural property of the network. Two methods are proposed. Numerical simulations are used to show the effectiveness of the improved BP training methods.",
author = "Takafumi Sasakawa and Takayuki Furuzuki and Katsunori Isono and Kotaro Hirasawa",
year = "2006",
language = "English",
isbn = "0780394909",
pages = "4785--4790",
booktitle = "IEEE International Conference on Neural Networks - Conference Proceedings",

}

TY - GEN

T1 - Effective training methods for function localization neural networks

AU - Sasakawa, Takafumi

AU - Furuzuki, Takayuki

AU - Isono, Katsunori

AU - Hirasawa, Kotaro

PY - 2006

Y1 - 2006

N2 - Inspired by Hebb's cell assembly theory about how the brain worked, we have developed a function localization neural network (FLNN). The main part of a FLNN is structurally the same as an ordinary feedforward neural network, but it is considered to consist of several overlapping modules, which are switched according to input patterns. A FLNN constructed in this way has been shown to have better representation ability than an ordinary neural network. However, BP training algorithm for such FLNN is very easy to get stuck at a local minimum. In this paper, we mainly discuss the methods for improving BP training of the FLNN by utilizing the structural property of the network. Two methods are proposed. Numerical simulations are used to show the effectiveness of the improved BP training methods.

AB - Inspired by Hebb's cell assembly theory about how the brain worked, we have developed a function localization neural network (FLNN). The main part of a FLNN is structurally the same as an ordinary feedforward neural network, but it is considered to consist of several overlapping modules, which are switched according to input patterns. A FLNN constructed in this way has been shown to have better representation ability than an ordinary neural network. However, BP training algorithm for such FLNN is very easy to get stuck at a local minimum. In this paper, we mainly discuss the methods for improving BP training of the FLNN by utilizing the structural property of the network. Two methods are proposed. Numerical simulations are used to show the effectiveness of the improved BP training methods.

UR - http://www.scopus.com/inward/record.url?scp=40649128098&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=40649128098&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:40649128098

SN - 0780394909

SN - 9780780394902

SP - 4785

EP - 4790

BT - IEEE International Conference on Neural Networks - Conference Proceedings

ER -