Presentation | 1996/3/19 A Study on Relations between Internal Representation and Generalization Ability of Multi Layered Neural Networks for Function Approximation Problems Eiji Watanabe, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | This report discusses the relation between internal representations and generalization ability of multi layered neural networks for function approximation problems. Here, the over learning problem, which makes the generalization ability poor, is divided into the two problems; the excessive degrees of freedom and the non-uniqueness of weights. These problems are carefully discussed from the view point of internal representation; weights and outputs or hidden units. First, the excessive degrees of freedom problem is discussed by introducing the entropy for weights. Next, the non-uniqueness of weights problem is also discussed by introducing the principal components analysis method. Finally, the relations between internal representations and generalization ability of multi layered neural networks are concretely discussed by numerical results. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | Multi layered neural network / function approximation problem / internal representation / generalization ability / over learning |
Paper # | NC95-164 |
Date of Issue |
Conference Information | |
Committee | NC |
---|---|
Conference Date | 1996/3/19(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 Study on Relations between Internal Representation and Generalization Ability of Multi Layered Neural Networks for Function Approximation Problems |
Sub Title (in English) | |
Keyword(1) | Multi layered neural network |
Keyword(2) | function approximation problem |
Keyword(3) | internal representation |
Keyword(4) | generalization ability |
Keyword(5) | over learning |
1st Author's Name | Eiji Watanabe |
1st Author's Affiliation | Department of Information Processing Engineering Faculty of Engineering, Fukuyama University() |
Date | 1996/3/19 |
Paper # | NC95-164 |
Volume (vol) | vol.95 |
Number (no) | 599 |
Page | pp.pp.- |
#Pages | 8 |
Date of Issue |