What is the Largest Sparsity Pattern that Can Be Recovered by 1-Norm Minimization?

TL;DR

This paper characterizes the largest sparsity patterns recoverable by 1-norm minimization using a mathematical framework involving maximum abstract simplicial complexes, with applications to graph incidence matrices and partial Fourier transforms.

Contribution

It introduces a novel framework based on maximum abstract simplicial complexes to determine the largest recoverable sparsity patterns in 1-norm minimization, extending understanding to specific matrix types.

Findings

Largest recoverable patterns are described by maximum abstract simplicial complexes.

Recovery for incidence matrices depends on the graph's simple cycles.

Polynomial-time certification of sparse recovery for certain matrices.

Abstract

Much of the existing literature in sparse recovery is concerned with the following question: given a sparsity pattern and a corresponding regularizer, derive conditions on the dictionary under which exact recovery is possible. In this paper, we study the opposite question: given a dictionary and the 1-norm regularizer, find the largest sparsity pattern that can be recovered. We show that such a pattern is described by a mathematical object called a "maximum abstract simplicial complex", and provide two different characterizations of this object: one based on extreme points and the other based on vectors of minimal support. In addition, we show how this new framework is useful in the study of sparse recovery problems when the dictionary takes the form of a graph incidence matrix or a partial discrete Fourier transform. In case of incidence matrices, we show that the largest sparsity pattern that can be recovered is determined by the set of simple cycles of the graph. As a byproduct, we show that standard sparse recovery can be certified in polynomial time, although this is known to be NP-hard for general matrices. In the case of the partial discrete Fourier transform, our characterization of the largest sparsity pattern that can be recovered requires the unknown signal to be real and its dimension to be a prime number.