Presentation 2022-03-09
Learning coefficients of deep linear neural networks
Ryuji Ito, Miki Aoyagi, Seikai Ashizawa, Yuki Tsukamoto, Daisuke Kaji,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Recently, the learning theory has been analyzed based on resolution of learning machine singularities in algebraic geometry. In particular, the asymptotic form of the Bayesian generalization error, the average stochastic complexity (the free energy) and so on have been obtained using these method. In this paper, we clarify the learning coefficients ofdeep linear neural networks.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) resolution of singularities / learning coefficients / singular models / linear neural networks
Paper # IBISML2021-45
Date of Issue 2022-03-01 (IBISML)

Conference Information
Committee IBISML
Conference Date 2022/3/8(2days)
Place (in Japanese) (See Japanese page)
Place (in English) Online
Topics (in Japanese) (See Japanese page)
Topics (in English) Machine Learning, etc.
Chair Ichiro Takeuchi(Nagoya Inst. of Tech.)
Vice Chair Masashi Sugiyama(Univ. of Tokyo)
Secretary Masashi Sugiyama(Univ. of Tokyo)
Assistant Tomoharu Iwata(NTT) / Atsuyoshi Nakamura(Hokkaido Univ.)

Paper Information
Registration To Technical Committee on Infomation-Based Induction Sciences and Machine Learning
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Learning coefficients of deep linear neural networks
Sub Title (in English)
Keyword(1) resolution of singularities
Keyword(2) learning coefficients
Keyword(3) singular models
Keyword(4) linear neural networks
1st Author's Name Ryuji Ito
1st Author's Affiliation Nihon University(Nihon Univ.)
2nd Author's Name Miki Aoyagi
2nd Author's Affiliation Nihon University(Nihon Univ.)
3rd Author's Name Seikai Ashizawa
3rd Author's Affiliation Nihon University(Nihon Univ.)
4th Author's Name Yuki Tsukamoto
4th Author's Affiliation Nihon University(Nihon Univ.)
5th Author's Name Daisuke Kaji
5th Author's Affiliation Denso Corporation(Denso)
Date 2022-03-09
Paper # IBISML2021-45
Volume (vol) vol.121
Number (no) IBISML-419
Page pp.pp.46-52(IBISML),
#Pages 7
Date of Issue 2022-03-01 (IBISML)