Presentation 2000/12/1
A Learning Scheme of Multi-layered Neural Networks with Error Forward Propagation and Linear Multiple Regression.
Kazuyuki NAGASAWA, Naohiro FUKUMURA, Yoji UNO,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Many learning schemes have been proposed to acquire neural inverse models of controlled objects. Most schemes use "Back Propagation" rule to update the weights of multi-layered neural networks. "Back Propagation" rule requires teaching signals for the output layers in advance. Here we propose "Forward Propagation" scheme to acquire neural inverse models. In this scheme, error signals between desired and realized trajectories are forward propagated and are applied to estimate teaching signals for the middle and output layers. This rule corresponds to Newton-like method. Linear multiple regression, which is faster than usual steepest descent method, is available to update the weights.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Back Propagation / multi-layered neural network / Newton-like method / linear multiple regression
Paper # NC2000-78
Date of Issue

Conference Information
Committee NC
Conference Date 2000/12/1(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Neurocomputing (NC)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) A Learning Scheme of Multi-layered Neural Networks with Error Forward Propagation and Linear Multiple Regression.
Sub Title (in English)
Keyword(1) Back Propagation
Keyword(2) multi-layered neural network
Keyword(3) Newton-like method
Keyword(4) linear multiple regression
1st Author's Name Kazuyuki NAGASAWA
1st Author's Affiliation Department of Information and Computer Sciences, Toyohashi University of Technology()
2nd Author's Name Naohiro FUKUMURA
2nd Author's Affiliation Department of Information and Computer Sciences, Toyohashi University of Technology
3rd Author's Name Yoji UNO
3rd Author's Affiliation Department of Information and Computer Sciences, Toyohashi University of Technology
Date 2000/12/1
Paper # NC2000-78
Volume (vol) vol.100
Number (no) 490
Page pp.pp.-
#Pages 8
Date of Issue