Phase transition and noise sensitivity of ℓp-minimization for 0 ≤ p ≤ 1

Haolei Weng, Le Zheng, Arian Maleki, Xiaodong Wang

科研成果: 书/报告/会议事项章节会议稿件同行评审

3 引用 (Scopus)

摘要

Recovering a sparse vector x0 ⋯ ℝN from its noisy linear observations, y ⋯ ℝn with y = Ax0 + w, has been the central problem of compressed sensing. One of the classes of recovery algorithms that has attracted attention is the class of ℓp-regularized least squares (LPLS) that seeks the minimum of 1/2 ||y - Ax||22 + λ||x||pp for p ⋯ [0, 1]. In this paper we employ the Replica method1 from statistical physics to analyze the global minima of LPLS. Our paper reveals several surprising asymptotic properties of LPLS: (i) The phase transition curve of LPLS is the same for every 0 ≤ p < 1. These phase transition curves are much higher than the phase transition curve for p = 1. (ii) The phase transition curve of LPLS for every value of 0 ≤ p ≤ 1 depends only on the sparsity level and does not change with the distribution of the non-zero coefficients of x0. (iii) Despite the equality of the phase transition curves, different values of p show different performances once a small amount of measurement noise, w, is added.

源语言英语
主期刊名Proceedings - ISIT 2016; 2016 IEEE International Symposium on Information Theory
出版商Institute of Electrical and Electronics Engineers Inc.
675-679
页数5
ISBN(电子版)9781509018062
DOI
出版状态已出版 - 10 8月 2016
已对外发布
活动2016 IEEE International Symposium on Information Theory, ISIT 2016 - Barcelona, 西班牙
期限: 10 7月 201615 7月 2016

出版系列

姓名IEEE International Symposium on Information Theory - Proceedings
2016-August
ISSN(印刷版)2157-8095

会议

会议2016 IEEE International Symposium on Information Theory, ISIT 2016
国家/地区西班牙
Barcelona
时期10/07/1615/07/16

指纹

探究 'Phase transition and noise sensitivity of ℓp-minimization for 0 ≤ p ≤ 1' 的科研主题。它们共同构成独一无二的指纹。

引用此