Presentation 2001/11/9
A Distributed Learning Algorithm for Multi-Layered Neural Networks and Its Applications to Pattern Classification Problems
Eiji Watanabe,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In this report, a distributed learning algorithm of multi-layered neural networks(NNs)is newly proposed using a PC cluster. In the proposed distributed learning algorithm, the learning data is distributed among the group of NNs based on the idea in the clustering method. Then the proposed learning algorithm has the following functions : (i)the decentralization of the learning load in each NN, (ii)the improvement of the learning ability in the group of NNs. Next, a neural network for the integration of the response of each NN is introduced for the purpose of the improvement of the generalization ability. Finally, the proposed learning algorithm is applied to the pattern classification problems and it is shown that the proposed learning algorithm can distribute the learning load according to the learning ability of each NN.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Multi-layered neural network / PC cluster / Module structure / Distributed learning algorithm / Information integration
Paper # NC 2001-63
Date of Issue

Conference Information
Committee NC
Conference Date 2001/11/9(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 Distributed Learning Algorithm for Multi-Layered Neural Networks and Its Applications to Pattern Classification Problems
Sub Title (in English)
Keyword(1) Multi-layered neural network
Keyword(2) PC cluster
Keyword(3) Module structure
Keyword(4) Distributed learning algorithm
Keyword(5) Information integration
1st Author's Name Eiji Watanabe
1st Author's Affiliation Department of Information Science and Systems Engineering Faculty of Science and Engineering, Konan University()
Date 2001/11/9
Paper # NC 2001-63
Volume (vol) vol.101
Number (no) 432
Page pp.pp.-
#Pages 8
Date of Issue