Sparse and Smooth Signal Estimation: Convexification of L0 Formulations
share
Signal estimation problems with smoothness and sparsity priors can be naturally modeled as quadratic optimization with ℓ_0-"norm" constraints. Since such problems are non-convex and hard-to-solve, the standard approach is, instead, to tackle their convex surrogates based on ℓ_1-norm relaxations. In this paper, we propose new iterative conic quadratic relaxations that exploit not only the ℓ_0-"norm" terms, but also the fitness and smoothness functions. The iterative convexification approach substantially closes the gap between the ℓ_0-"norm" and its ℓ_1 surrogate. Experiments using an off-the-shelf conic quadratic solver on synthetic as well as real datasets indicate that the proposed iterative convex relaxations lead to significantly better estimators than ℓ_1-norm while preserving the computational efficiency. In addition, the parameters of the model and the resulting estimators are easily interpretable.
READ FULL TEXT