BIOMEDICAL IMAGING GROUP (BIG)
Laboratoire d'imagerie biomédicale (LIB)

Not All ℓp-Norms Are Compatible with Sparse Stochastic Processes

A. Amini, M. Unser

Signal Processing with Adaptive Sparse Structured Representations (SPARS'13), Lausanne VD, Swiss Confederation, July 8-11, 2013.

We adopt the framework of sparse stochastic processes of [1] and investigate the sparse/compressible priors obtained by linearly measuring the processes. We show such priors are necessarily infinitely divisible. This property is satisfied by many priors used in statistical learning such as Gaussian, Laplace, and a wide range of fat-tailed distributions such as Student's-t and α-stable laws. However, it excludes some popular priors used in compressed sensing including all distributions that decay like exp(−O(|x|p)) for 1 < p < 2. This fact can be considered as evidence against the usage of ℓp-norms for 1 < p < 2 in regularization techniques involving sparse priors.

Reference

  1. M. Unser, P. Tafti, Q. Sun, "A Unified Formulation of Gaussian vs. Sparse Stochastic Processes—Part I: Continuous-Domain Theory," arXiv:1108.6150v1 [cs.IT] 

@INPROCEEDINGS(http://bigwww.epfl.ch/publications/amini1304.html,
AUTHOR="Amini, A. and Unser, M.",
TITLE="Not All $\ell_{p}$-Norms Are Compatible with Sparse Stochastic
	Processes",
BOOKTITLE="Signal Processing with Adaptive Sparse Structured
	Representations ({SPARS'13})",
YEAR="2013",
editor="",
volume="",
series="",
pages="",
address="Lausanne VD, Swiss Confederation",
month="July 8-11,",
organization="",
publisher="",
note="")
© 2013 FNSNF. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from FNSNF. This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each author's copyright. In most cases, these works may not be reposted without the explicit permission of the copyright holder.