Presentation 1994/3/25
Back propagation based on maximum information
Tatsuya Uno, Seiichi Koakutsu, Hironori Hirata,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) The complicated structure of neural networks organized by back propagation algorithm makes it difficult to interpret the meaning of learned structure.To overcome this defect,we define information which evaluates the degree of the organization of the network structure and propose a new version of back propagation based on maximum information in order to obtain a simple structure.Some computational experiments show that the proposed method is useful to understand the internal structure of learned problem.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) neural networks / BP algorithm / structuralization / information / maximum information
Paper # NC93-113
Date of Issue

Conference Information
Committee NC
Conference Date 1994/3/25(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) Back propagation based on maximum information
Sub Title (in English)
Keyword(1) neural networks
Keyword(2) BP algorithm
Keyword(3) structuralization
Keyword(4) information
Keyword(5) maximum information
1st Author's Name Tatsuya Uno
1st Author's Affiliation Department of Electrical and Electronics Engineering,Faculty of Engineering,Chiba University()
2nd Author's Name Seiichi Koakutsu
2nd Author's Affiliation Department of Electrical and Electronics Engineering,Faculty of Engineering,Chiba University
3rd Author's Name Hironori Hirata
3rd Author's Affiliation Department of Electrical and Electronics Engineering,Faculty of Engineering,Chiba University
Date 1994/3/25
Paper # NC93-113
Volume (vol) vol.93
Number (no) 537
Page pp.pp.-
#Pages 8
Date of Issue