Presentation 2012-11-08
Revocery algorithm in compressed sensing based on maximim a posteriori estimation
Koujin TAKEDA, Yoshiyuki KABASHIMA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) We propose an idea for how to derive sparse data recovery algorithm with smaller computational cost in compressed sensing. It is known that the cost of the standard l_1 recovery algorithm based on linear programming is O(N^3). We show that this cost can be reduced to O(N^2)(for sparse matrix O(N)) by introducing the maximization procedure of posterior. Furthermore, by this method we can reconstruct original data with as small number of observations as possible in principle. We also discuss the relationship between belief-propagation based recovery algorithm in preceding work and our idea.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) compressed sensing / data compression / sparsity / statistical mechanics / maximum a posteriori estimation
Paper # IBISML2012-76
Date of Issue

Conference Information
Committee IBISML
Conference Date 2012/10/31(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) Revocery algorithm in compressed sensing based on maximim a posteriori estimation
Sub Title (in English)
Keyword(1) compressed sensing
Keyword(2) data compression
Keyword(3) sparsity
Keyword(4) statistical mechanics
Keyword(5) maximum a posteriori estimation
1st Author's Name Koujin TAKEDA
1st Author's Affiliation Department of Computational Intelligence and Systems Science, Tokyo Institute of Technology()
2nd Author's Name Yoshiyuki KABASHIMA
2nd Author's Affiliation Department of Computational Intelligence and Systems Science, Tokyo Institute of Technology
Date 2012-11-08
Paper # IBISML2012-76
Volume (vol) vol.112
Number (no) 279
Page pp.pp.-
#Pages 6
Date of Issue