Minimization Over the Nonconvex Sparsity Constraint Using A Hybrid First-order method

TL;DR

This paper introduces a hybrid first-order method combining Frank-Wolfe and gradient projection techniques to efficiently solve nonconvex optimization problems with sparsity constraints, achieving global convergence without smoothing.

Contribution

A novel hybrid algorithm that efficiently addresses nonconvex sparsity-constrained problems with proven convergence and practical advantages over existing methods.

Findings

Achieves $O(1/\sqrt{k})$ convergence rate.

Demonstrates efficiency through numerical experiments.

Operates without smoothing parameters.

Abstract

We investigate a class of nonconvex optimization problems characterized by a feasible set consisting of level-bounded nonconvex regularizers, with a continuously differentiable objective. We propose a novel hybrid approach to tackle such structured problems within a first-order algorithmic framework by combining the Frank-Wolfe method and the gradient projection method. The Frank-Wolfe step is amenable to a closed-form solution, while the gradient projection step can be efficiently performed in a reduced subspace. A notable characteristic of our approach lies in its independence from introducing smoothing parameters, enabling efficient solutions to the original nonsmooth problems. We establish the global convergence of the proposed algorithm and show the $O(1/\sqrt{k})$ convergence rate in terms of the optimality error for nonconvex objectives under reasonable assumptions. Numerical experiments underscore the practicality and efficiency of our proposed algorithm compared to existing cutting-edge methods. Furthermore, we highlight how the proposed algorithm contributes to the advancement of nonconvex regularizer-constrained optimization.