Presentation 1996/3/18
Bayesian Estimation of Parameters of Various Regularizers in the Learning of Neural Networks
Kazuhiro Yoshida, Masumi Ishikawa,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) A regularizer is often used in the learning of neural networks to avoid over-generalization or over-fitting. It is effective, because it penalizes large connection weights. In most cases, a regularizer of the sum of squared connection weights, called a Gaussian regularizer, is used. It corresponds to the one proposed by Plant and others. One of th difficulties in the use of regularizers is that the determibation of a regularization parameter. The present paper proposes the use of various types of regularizers such as the sum of the absolute values of connection weight, called a Laplace regularizer, which has been used in structural learning with fogetting.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Bayesian estimation / regularization / neural network
Paper # NC-95-122
Date of Issue

Conference Information
Committee NC
Conference Date 1996/3/18(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) Bayesian Estimation of Parameters of Various Regularizers in the Learning of Neural Networks
Sub Title (in English)
Keyword(1) Bayesian estimation
Keyword(2) regularization
Keyword(3) neural network
1st Author's Name Kazuhiro Yoshida
1st Author's Affiliation Faculty of Computer Science and Systems Engineering Kyushu Institute of Technology()
2nd Author's Name Masumi Ishikawa
2nd Author's Affiliation Faculty of Computer Science and Systems Engineering Kyushu Institute of Technology
Date 1996/3/18
Paper # NC-95-122
Volume (vol) vol.95
Number (no) 598
Page pp.pp.-
#Pages 8
Date of Issue