Abstract:Motivated by recent results in the statistical physics of spin glasses, we study the recovery of a sparse vector $\mathbf{x}_0\in \mathbb{S}^{n-1}$, $\|\mathbf{x}_0\|_{\ell_0} = k<n$, from $m$ quadratic measurements of the form $ (1-\lambda)\langle \mathbf{A}_i, \mathbf{x}_0\mathbf{x}_0^T \rangle + \lambda \langle\mathbf{c}_i,\mathbf{x}_0 \rangle $ where $\mathbf{A}_{i}, \mathbf{c}_{i}$ have i.i.d Gaussian entries. This can be related to a constrained version of the 2-spin Hamiltonian with external field for which it was recently shown (in the absence of any structural constraint and in the asymptotic regime) that the geometry of the energy landscape becomes trivial above a certain threshold $\lambda > \lambda_c\in (0,1)$. Building on this idea we study the evolution of the so-called square root bottleneck for $\lambda\in [0,1]$ in the setting of the sparse rank one matrix recovery/sensing problem. We show that recovery of the vector $\mathbf{x}_0$ can be guaranteed as soon as $m\gtrsim k^2 (1-\lambda)^2/\lambda^2$, $\lambda \gtrsim k^{-1/2}$ provided that this vector satisfies a sufficiently strong incoherence condition, thus retrieving the linear regime for an external field $(1-\lambda)/\lambda \lesssim k^{-1/2}$. Our proof relies on an interpolation between the linear and quadratic settings, as well as on standard convex geometry arguments.

Recovery of sparsest signals via $\ell^q$-minimization

On the Schatten P-Quasi-norm Minimization for Low-Rank Matrix Recovery

Block Sparse Recovery Via Mixed L 2 / L 1 Minimization

Signal Recovery with Partially Known Signal Support

Sparse Signal Recovery by Minimization Under Restricted Isometry Property

Stability and robustness of the l2/lq-minimization for block sparse recovery.

Sparse Signal Recovery By L(Q) Minimization Under Restricted Isometry Property

Sparse Estimation Via ℓ_q Optimization Method in High-Dimensional Linear Regression

Phase Retrieval from Quantized Measurements Via Approximate Message Passing

Sparse Recovery with Coherent Tight Frame via Analysis Dantzig Selector and Analysis LASSO

Optimal RIP Bounds for Sparse Signals Recovery Via ℓ Minimization

Sorted Minimization for Sparse Signal Recovery

The Dantzig Selector: Sparse Signals Recovery via l_p-q Minimization

Sparse Estimation Via Lower-Order Penalty Optimization Methods in High-Dimensional Linear Regression.

Sorted L1/L2 Minimization for Sparse Signal Recovery

Subspace Weighted L(2,1) Minimization For Sparse Signal Recovery

Decentralized Jointly Sparse Optimization by Reweighted $\ell_{q}$ Minimization

A piecewise linear programming algorithm for sparse signal reconstruction

Restricted -Isometry Properties of Nonconvex Matrix Recovery

In the recovery of sparse vectors from quadratic measurements, the presence of linear terms breaks the square root bottleneck

Stochastic Recovery Of Sparse Signals From Random Measurements