Necessary moment conditions for exact reconstruction via basis pursuit

TL;DR

This paper demonstrates that certain 'spiky' measurement vectors in compressed sensing can cause basis pursuit to fail in exact reconstruction, highlighting limitations of convex relaxation methods under specific conditions.

Contribution

It constructs a specific random vector example showing the failure of basis pursuit with 'spiky' vectors, emphasizing the limitations of convex relaxation in compressed sensing.

Findings

Basis pursuit fails for certain 'spiky' measurement vectors

Constructed vectors satisfy weak small ball property and moment conditions

Shows alternative algorithms like -minimization may succeed

Abstract

Let $X=(x_1,...,x_n)$ be a random vector that satisfies a weak small ball property and whose coordinates $x_i$ satisfy that $\|x_i\|_{L_p} \lesssim \sqrt{p} \|x_i\|_{L_2}$ for $p \sim \log n$. In \cite{LM_compressed}, it was shown that $N$ independent copies of $X$ can be used as measurement vectors in Compressed Sensing (using the basis pursuit algorithm) to reconstruct any $d$-sparse vector with the optimal number of measurements $N\gtrsim d \log\big(e n/d\big)$. In this note we show that the result is almost optimal. We construct a random vector $X$ with iid, mean-zero, variance one coordinates that satisfies the same weak small ball property and whose coordinates satisfy that $\|x_i\|_{L_p} \lesssim \sqrt{p} \|x_i\|_{L_2}$ for $p \sim (\log n)/(\log N)$, but the basis pursuit algorithm fails to recover even $1$-sparse vectors. The construction shows that `spiky' measurement vectors may lead to a poor performance by the basis pursuit algorithm, but on the other hand may still perform in an optimal way if one chooses a different reconstruction algorithm (like $\ell_0$-minimization). This exhibits the fact that the convex relaxation of $\ell_0$-minimization comes at a significant cost when using `spiky' measurement vectors.