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. We introduce an $\ell_2$ 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 $6.22\%$ more sparse. When compared only against the experiment-wise best performing benchmark method on synthetic data, our approach produces solutions that are on average $3.10\%$ more sparse. On real world ECG data, for a given $\ell_2$ reconstruction error our approach produces solutions that are on average $9.95\%$ more sparse than benchmark methods ($3.88\%$ more sparse if only compared against the best performing benchmark), while for a given sparsity level our approach produces solutions that have on average $10.77\%$ lower reconstruction error than benchmark methods ($1.42\%$ 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.

A compressed sensing approach for efficient ensemble learning

Joint Sparse Learning for Classification Ensemble.

Convex ensemble learning with sparsity and diversity

Compressed Sensing: A Discrete Optimization Approach

Sparse Ensembles Using Weighted Combination Methods Based on Linear Programming

Achieving More with Less: A Tensor-Optimization-Powered Ensemble Method

Towards Inference Efficient Deep Ensemble Learning

A novel classifier ensemble method with sparsity and diversity

Ensemble Pruning Via Constrained Eigen-Optimization

Learning a Compressed Sensing Measurement Matrix via Gradient Unrolling

End-to-End Ensemble Learning by Exploiting the Correlation Between Individuals and Weights.

Learning Sparse Structured Ensembles with Stochastic Gradient MCMC Sampling and Network Pruning.

Distributed Compressive Sensing: A Deep Learning Approach

A Compressed Sensing Ensemble Classifier with Application to Human Detection

Dynamically building diversified classifier pruning ensembles via canonical correlation analysis

Optimized Compressed Sensing for Communication Efficient Federated Learning.

A Double Penalty Model for Ensemble Learning

An Efficient Algorithm for Compressed Sensing Image Reconstruction

Classifier Ensemble Using A Heuristic Learning With Sparsity And Diversity