Presentation 2015-07-18
Feature extraction based on generation of Bayesian Network
Kaneharu Nishino, Mary Inaba,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Networks used in Deep Learning generally have feedforward architectures, and they can not use top-down information for recognition. However in brains, there are not only feedforward connections from the lower level to the higher but also feedback connections from higher to lower. As the reason for the feedbacks, some studies propose that the information processing model of brains is based on a Bayesian Network. In this paper, we propose Bayesian AutoEncoder (BAE) to construct Bayesian Networks. The networks constructed by BAE can execute inference using bottom-up and top-down information. We confirmed that BAE can construct small networks and extract features.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Batesian Network / feature extraction
Paper # NC2015-13
Date of Issue 2015-07-11 (NC)

Conference Information
Committee MBE / NC
Conference Date 2015/7/18(1days)
Place (in Japanese) (See Japanese page)
Place (in English) The University of Tokushima
Topics (in Japanese) (See Japanese page)
Topics (in English) Me, general
Chair Tetsuo Kobayashi(Kyoto Univ.) / Toshimichi Saito(Hosei Univ.)
Vice Chair Yutaka Fukuoka(Kogakuin Univ.) / Shigeo Sato(Tohoku Univ.)
Secretary Yutaka Fukuoka(akita noken) / Shigeo Sato(Kogakuin Univ.)
Assistant Takenori Oida(Kyoto Univ.) / Ryota Horie(Shibaura Inst. of Tech.) / Hiroyuki Kanbara(Tokyo Inst. of Tech.) / Hisanao Akima(Tohoku Univ.)

Paper Information
Registration To Technical Committee on ME and Bio Cybernetics / Technical Committee on Neurocomputing
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Feature extraction based on generation of Bayesian Network
Sub Title (in English)
Keyword(1) Batesian Network
Keyword(2) feature extraction
1st Author's Name Kaneharu Nishino
1st Author's Affiliation The University of Tokyo(Univ. Tokyo)
2nd Author's Name Mary Inaba
2nd Author's Affiliation The University of Tokyo(Univ. Tokyo)
Date 2015-07-18
Paper # NC2015-13
Volume (vol) vol.115
Number (no) NC-148
Page pp.pp.7-11(NC),
#Pages 5
Date of Issue 2015-07-11 (NC)