Presentation 2014-11-17
Statistical Mechanical Analysis of Lossy Compression by Overcomplete Basis
Yoshinori NAKANISHI-OHNO, Tomoyuki OBUCHI, Masato OKADA, Yoshiyuki KABASHIMA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Information processing using the sparsity of information has been actively studied such as compressed sensing. In this study, we discuss sparse information processing from the viewpoint of lossy compression. The choice of a basis is important for lossy compression. Overcomplete bases are expected to be useful to compress information because they include more candidates to express it. We analyze the rate-distortion relation of overcomplete basis compression in a statistical mechanical manner and show they outperform a naive random projection approach. We also discuss some algorithms to realize the rate-distortion relation of overcomplete basis compression.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) overcomplete basis compression / rate distortion relation / replica method
Paper # IBISML2014-50
Date of Issue

Conference Information
Committee IBISML
Conference Date 2014/11/10(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 Information-Based Induction Sciences and Machine Learning (IBISML)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Statistical Mechanical Analysis of Lossy Compression by Overcomplete Basis
Sub Title (in English)
Keyword(1) overcomplete basis compression
Keyword(2) rate distortion relation
Keyword(3) replica method
1st Author's Name Yoshinori NAKANISHI-OHNO
1st Author's Affiliation Graduate School of Frontier Sciences, The University of Tokyo:Japan Society for the Promotion of Science()
2nd Author's Name Tomoyuki OBUCHI
2nd Author's Affiliation Interdisciplinary Graduate School of Science and Engineering, Tokyo Institute of Technology
3rd Author's Name Masato OKADA
3rd Author's Affiliation Graduate School of Frontier Sciences, The University of Tokyo
4th Author's Name Yoshiyuki KABASHIMA
4th Author's Affiliation Interdisciplinary Graduate School of Science and Engineering, Tokyo Institute of Technology
Date 2014-11-17
Paper # IBISML2014-50
Volume (vol) vol.114
Number (no) 306
Page pp.pp.-
#Pages 8
Date of Issue