Abstract:We study the Compressed Sensing (CS) problem, which is the problem of finding the most sparse vector that satisfies a set of linear measurements up to some numerical tolerance. CS is a central problem in Statistics, Operations Research and Machine Learning which arises in applications such as signal processing, data compression, image reconstruction, and multi-label learning. We introduce an regularized formulation of CS which we reformulate as a mixed integer second order cone program. We derive a second order cone relaxation of this problem and show that under mild conditions on the regularization parameter, the resulting relaxation is equivalent to the well studied basis pursuit denoising problem. We present a semidefinite relaxation that strengthens the second order cone relaxation and develop a custom branch-and-bound algorithm that leverages our second order cone relaxation to solve small-scale instances of CS to certifiable optimality. When compared against solutions produced by three state of the art benchmark methods on synthetic data, our numerical results show that our approach produces solutions that are on average more sparse. When compared only against the experiment-wise best performing benchmark method on synthetic data, our approach produces solutions that are on average more sparse. On real world ECG data, for a given reconstruction error our approach produces solutions that are on average more sparse than benchmark methods ( more sparse if only compared against the best performing benchmark), while for a given sparsity level our approach produces solutions that have on average lower reconstruction error than benchmark methods ( lower error if only compared against the best performing benchmark). When used as a component of a multi-label classification algorithm, our approach achieves greater classification accuracy than benchmark compressed sensing methods. This improved accuracy comes at the cost of an increase in computation time by several orders of magnitude. Thus, for applications where runtime is not of critical importance, leveraging integer optimization can yield sparser and lower error solutions to CS than existing benchmarks.

Computational Complexity versus Statistical Performance on Sparse Recovery Problems

Computational complexity versus statistical performance on sparse recovery problems

Sparsity optimized compressed sensing image recovery

Weighted LASSO for Sparse Recovery with Statistical Prior Support Information

Compressive Sensing with Sparse Measurement Matrices

Near Optimal Compressed Sensing of a Class of Sparse Low-Rank Matrices via Sparse Power Factorization

Estimating Unknown Sparsity in Compressed Sensing

Sharpness and well-conditioning of nonsmooth convex formulations in statistical signal recovery

Sparse Recovery Beyond Compressed Sensing: Separable Nonlinear Inverse Problems

Robustness of Sparse Recovery Via &Lt;inline-Formula> &Lt;tex-Math Notation="latex">$f$ &Lt;/tex-Math></inline-formula>-minimization: A Topological Viewpoint

Compressed sensing: a discrete optimization approach

Compressed sensing with l0-norm: statistical physics analysis and algorithms for signal recovery

Sparsity Estimation in Image Compressive Sensing

On unified view of nullspace-type conditions for recoveries associated with general sparsity structures

Compressed Sensing: How sharp is the Restricted Isometry Property

Compressive Sensing of Digital Sparse Signals.

Sparsification of Matrices and Compressed Sensing

On the Fundamental Limits of Recovering Tree Sparse Vectors from Noisy Linear Measurements

Robust Sparse Recovery with Sparse Bernoulli matrices via Expanders