Gradient Descent Methods for Regularized Optimization

TL;DR

This paper reviews gradient descent and proximal gradient descent methods for regularized optimization, introduces a variable step size algorithm based on local Lipschitz estimates, and demonstrates improved efficiency through experiments.

Contribution

It proposes a novel proximal gradient descent algorithm with adaptive step size estimation based on local Lipschitz constants, enhancing convergence speed.

Findings

The proposed method reduces the number of iterations needed for convergence.

It improves computational efficiency in synthetic and real-data experiments.

The adaptive step size outperforms fixed step size in proximal gradient descent.

Abstract

Regularization is a widely recognized technique in mathematical optimization. It can be used to smooth out objective functions, refine the feasible solution set, or prevent overfitting in machine learning models. Due to its simplicity and robustness, the gradient descent (GD) method is one of the primary methods used for numerical optimization of differentiable objective functions. However, GD is not well-suited for solving $\ell^1$ regularized optimization problems since these problems are non-differentiable at zero, causing iteration updates to oscillate or fail to converge. Instead, a more effective version of GD, called the proximal gradient descent employs a technique known as soft-thresholding to shrink the iteration updates toward zero, thus enabling sparsity in the solution. Motivated by the widespread applications of proximal GD in sparse and low-rank recovery across various engineering disciplines, we provide an overview of the GD and proximal GD methods for solving regularized optimization problems. Furthermore, this paper proposes a novel algorithm for the proximal GD method that incorporates a variable step size. Unlike conventional proximal GD, which uses a fixed step size based on the global Lipschitz constant, our method estimates the Lipschitz constant locally at each iteration and uses its reciprocal as the step size. This eliminates the need for a global Lipschitz constant, which can be impractical to compute. Numerical experiments we performed on synthetic and real-data sets show notable performance improvement of the proposed method compared to the conventional proximal GD with constant step size, both in terms of number of iterations and in time requirements.