Development of a supervisory training rule for multilayered feedforward neural network using local linearization and analytic optimal solution

Chun Ho Jeon, Yu Jin Cheon, Su Whan Sung, Changkyu Lee, Changkyoo Yoo, Dae Ryook Yang

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

Abstract

A new supervisory training rule for the multilayered feedforward neural network (FNN) using local linearization and analytic optimal solution is proposed. The cause of the nonlinearity of the neural network in the training is pinpointed and the nonlinearity is removed by a local linearization. And, the optimal solution of the linearized FNN minimizing the objective function for the training is analytically derived. The proposed training rule shows the shortest training time among the previous approaches. The superiority of the proposed approach is demonstrated by applying the proposed training rule to the modeling of the pH process.

Original languageEnglish
Title of host publicationICCAS-SICE 2009 - ICROS-SICE International Joint Conference 2009, Proceedings
Pages3697-3701
Number of pages5
Publication statusPublished - 2009 Dec 1
EventICROS-SICE International Joint Conference 2009, ICCAS-SICE 2009 - Fukuoka, Japan
Duration: 2009 Aug 182009 Aug 21

Other

OtherICROS-SICE International Joint Conference 2009, ICCAS-SICE 2009
CountryJapan
CityFukuoka
Period09/8/1809/8/21

Fingerprint

Feedforward neural networks
Linearization
Neural networks

Keywords

  • Linearization
  • Neural network
  • Optimal solution
  • Training rule

ASJC Scopus subject areas

  • Information Systems
  • Control and Systems Engineering
  • Industrial and Manufacturing Engineering

Cite this

Jeon, C. H., Cheon, Y. J., Sung, S. W., Lee, C., Yoo, C., & Yang, D. R. (2009). Development of a supervisory training rule for multilayered feedforward neural network using local linearization and analytic optimal solution. In ICCAS-SICE 2009 - ICROS-SICE International Joint Conference 2009, Proceedings (pp. 3697-3701). [5334761]

Development of a supervisory training rule for multilayered feedforward neural network using local linearization and analytic optimal solution. / Jeon, Chun Ho; Cheon, Yu Jin; Sung, Su Whan; Lee, Changkyu; Yoo, Changkyoo; Yang, Dae Ryook.

ICCAS-SICE 2009 - ICROS-SICE International Joint Conference 2009, Proceedings. 2009. p. 3697-3701 5334761.

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

Jeon, CH, Cheon, YJ, Sung, SW, Lee, C, Yoo, C & Yang, DR 2009, Development of a supervisory training rule for multilayered feedforward neural network using local linearization and analytic optimal solution. in ICCAS-SICE 2009 - ICROS-SICE International Joint Conference 2009, Proceedings., 5334761, pp. 3697-3701, ICROS-SICE International Joint Conference 2009, ICCAS-SICE 2009, Fukuoka, Japan, 09/8/18.
Jeon CH, Cheon YJ, Sung SW, Lee C, Yoo C, Yang DR. Development of a supervisory training rule for multilayered feedforward neural network using local linearization and analytic optimal solution. In ICCAS-SICE 2009 - ICROS-SICE International Joint Conference 2009, Proceedings. 2009. p. 3697-3701. 5334761
Jeon, Chun Ho ; Cheon, Yu Jin ; Sung, Su Whan ; Lee, Changkyu ; Yoo, Changkyoo ; Yang, Dae Ryook. / Development of a supervisory training rule for multilayered feedforward neural network using local linearization and analytic optimal solution. ICCAS-SICE 2009 - ICROS-SICE International Joint Conference 2009, Proceedings. 2009. pp. 3697-3701
@inproceedings{10a8b4aa07a841518dcd41a0f523ee49,
title = "Development of a supervisory training rule for multilayered feedforward neural network using local linearization and analytic optimal solution",
abstract = "A new supervisory training rule for the multilayered feedforward neural network (FNN) using local linearization and analytic optimal solution is proposed. The cause of the nonlinearity of the neural network in the training is pinpointed and the nonlinearity is removed by a local linearization. And, the optimal solution of the linearized FNN minimizing the objective function for the training is analytically derived. The proposed training rule shows the shortest training time among the previous approaches. The superiority of the proposed approach is demonstrated by applying the proposed training rule to the modeling of the pH process.",
keywords = "Linearization, Neural network, Optimal solution, Training rule",
author = "Jeon, {Chun Ho} and Cheon, {Yu Jin} and Sung, {Su Whan} and Changkyu Lee and Changkyoo Yoo and Yang, {Dae Ryook}",
year = "2009",
month = "12",
day = "1",
language = "English",
isbn = "9784907764333",
pages = "3697--3701",
booktitle = "ICCAS-SICE 2009 - ICROS-SICE International Joint Conference 2009, Proceedings",

}

TY - GEN

T1 - Development of a supervisory training rule for multilayered feedforward neural network using local linearization and analytic optimal solution

AU - Jeon, Chun Ho

AU - Cheon, Yu Jin

AU - Sung, Su Whan

AU - Lee, Changkyu

AU - Yoo, Changkyoo

AU - Yang, Dae Ryook

PY - 2009/12/1

Y1 - 2009/12/1

N2 - A new supervisory training rule for the multilayered feedforward neural network (FNN) using local linearization and analytic optimal solution is proposed. The cause of the nonlinearity of the neural network in the training is pinpointed and the nonlinearity is removed by a local linearization. And, the optimal solution of the linearized FNN minimizing the objective function for the training is analytically derived. The proposed training rule shows the shortest training time among the previous approaches. The superiority of the proposed approach is demonstrated by applying the proposed training rule to the modeling of the pH process.

AB - A new supervisory training rule for the multilayered feedforward neural network (FNN) using local linearization and analytic optimal solution is proposed. The cause of the nonlinearity of the neural network in the training is pinpointed and the nonlinearity is removed by a local linearization. And, the optimal solution of the linearized FNN minimizing the objective function for the training is analytically derived. The proposed training rule shows the shortest training time among the previous approaches. The superiority of the proposed approach is demonstrated by applying the proposed training rule to the modeling of the pH process.

KW - Linearization

KW - Neural network

KW - Optimal solution

KW - Training rule

UR - http://www.scopus.com/inward/record.url?scp=77951100839&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=77951100839&partnerID=8YFLogxK

M3 - Conference contribution

SN - 9784907764333

SP - 3697

EP - 3701

BT - ICCAS-SICE 2009 - ICROS-SICE International Joint Conference 2009, Proceedings

ER -