Presentation 2007-01-19
Application to Prediction Problem of Parallelized Neuron Networks in the Aircraft
Shunsuke KOBAYAKAWA, Hirokazu YOKOI,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) The hierarchical back-propagation networks over two outputs has total calculations using calculation results which are got independently by each output in calculations to change the weight and threshold of last middle layer at learning. This calculation process is a origin of cause of mutual interference to the learning for each output. And the independent learning for each output is carried out only in the output layer. Therefore this type is very disadvantageous to learn a nonlinear map of multi-output. We propose general parallelized neuron networks as a method for solving this problem and the application to the aircraft length system motion prediction system.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Parallelized / Neuron networks / Learning / Aircraft / Prediction
Paper # SANE2006-126
Date of Issue

Conference Information
Committee SANE
Conference Date 2007/1/12(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 Space, Aeronautical and Navigational Electronics (SANE)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Application to Prediction Problem of Parallelized Neuron Networks in the Aircraft
Sub Title (in English)
Keyword(1) Parallelized
Keyword(2) Neuron networks
Keyword(3) Learning
Keyword(4) Aircraft
Keyword(5) Prediction
1st Author's Name Shunsuke KOBAYAKAWA
1st Author's Affiliation Graduate School of Life Science and Systems Engineering, Kyushu Institute of Technology()
2nd Author's Name Hirokazu YOKOI
2nd Author's Affiliation Graduate School of Life Science and Systems Engineering, Kyushu Institute of Technology
Date 2007-01-19
Paper # SANE2006-126
Volume (vol) vol.106
Number (no) 471
Page pp.pp.-
#Pages 3
Date of Issue