On Polar Polytopes and the Recovery of Sparse Representations

DOI10.1109/TIT.2007.903129zbMATH Open1325.94038arXivcs/0510032MaRDI QIDQ3548904FDOQ3548904

Authors: Mark D. Plumbley 

Publication date: 21 December 2008

Published in: IEEE Transactions on Information Theory (Search for Journal in Brave)

Abstract: Suppose we have a signal y which we wish to represent using a linear combination of a number of basis atoms a_i, y=sum_i x_i a_i = Ax. The problem of finding the minimum L0 norm representation for y is a hard problem. The Basis Pursuit (BP) approach proposes to find the minimum L1 norm representation instead, which corresponds to a linear program (LP) that can be solved using modern LP techniques, and several recent authors have given conditions for the BP (minimum L1 norm) and sparse (minimum L0 solutions) representations to be identical. In this paper, we explore this sparse representation problem} using the geometry of convex polytopes, as recently introduced into the field by Donoho. By considering the dual LP we find that the so-called polar polytope P of the centrally-symmetric polytope P whose vertices are the atom pairs +-a_i is particularly helpful in providing us with geometrical insight into optimality conditions given by Fuchs and Tropp for non-unit-norm atom sets. In exploring this geometry we are able to tighten some of these earlier results, showing for example that the Fuchs condition is both necessary and sufficient for L1-unique-optimality, and that there are situations where Orthogonal Matching Pursuit (OMP) can eventually find all L1-unique-optimal solutions with m nonzeros even if ERC fails for m, if allowed to run for more than m steps.

Full work available at URL: https://arxiv.org/abs/cs/0510032

Recommendations

• Independent Component Analysis and Blind Signal Separation
• Sparse representations in unions of bases
• On Sparse Representations in Arbitrary Redundant Bases
• Optimally sparse representation in general (nonorthogonal) dictionaries via ℓ ¹ minimization
• On sparse representation in pairs of bases

Mathematics Subject Classification ID

Pattern recognition, speech recognition (68T10) Applications of mathematical programming (90C90) Signal theory (characterization, reconstruction, filtering, etc.) (94A12) Convex sets in (n) dimensions (including convex hypersurfaces) (52A20)

Cited In (12)

• Computing and analyzing recoverable supports for sparse reconstruction
• Independent Component Analysis and Blind Signal Separation
• Optimization methods for synthetic aperture radar imaging
• DASSO: Connections Between the Dantzig Selector and Lasso
• Title not available (Why is that?)
• Optimal dual certificates for noise robustness bounds in compressive sensing
• Optimally sparse representations of cartoon-like cylindrical data
• 1-bit compressive sensing: reformulation and RRSP-based sign recovery theory
• Testable uniqueness conditions for empirical assessment of undersampling levels in total variation-regularized X-ray CT
• On Theorem 10 in “On Polar Polytopes and the Recovery of Sparse Representations” [Sep 07 3188-3195]
• Covering point-sets with parallel hyperplanes and sparse signal recovery
• Weak stability of \(\ell_1\)-minimization methods in sparse data reconstruction

This page was built for publication: On Polar Polytopes and the Recovery of Sparse Representations

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3548904)