Inexact Successive Quadratic Approximation for Regularized Optimization

TL;DR

This paper analyzes the iteration complexity of inexact successive quadratic approximation methods for regularized optimization, demonstrating that approximate solutions to subproblems can achieve convergence rates comparable to exact methods across various problem types.

Contribution

It provides a global complexity analysis for inexact second-order proximal methods, allowing flexible second-order choices without increasing subproblem precision over iterations.

Findings

Inexact solutions within a fixed multiplicative precision suffice for convergence.

The methods achieve linear convergence for strongly convex problems.

For general convex problems, convergence is linear initially and $O(1/k)$ overall.

Abstract

Successive quadratic approximations, or second-order proximal methods, are useful for minimizing functions that are a sum of a smooth part and a convex, possibly nonsmooth part that promotes regularization. Most analyses of iteration complexity focus on the special case of proximal gradient method, or accelerated variants thereof. There have been only a few studies of methods that use a second-order approximation to the smooth part, due in part to the difficulty of obtaining closed-form solutions to the subproblems at each iteration. In fact, iterative algorithms may need to be used to find inexact solutions to these subproblems. In this work, we present global analysis of the iteration complexity of inexact successive quadratic approximation methods, showing that an inexact solution of the subproblem that is within a fixed multiplicative precision of optimality suffices to guarantee the same order of convergence rate as the exact version, with complexity related in an intuitive way to the measure of inexactness. Our result allows flexible choices of the second-order term, including Newton and quasi-Newton choices, and does not necessarily require increasing precision of the subproblem solution on later iterations. For problems exhibiting a property related to strong convexity, the algorithms converge at global linear rates. For general convex problems, the convergence rate is linear in early stages, while the overall rate is $O(1/k)$. For nonconvex problems, a first-order optimality criterion converges to zero at a rate of $O(1/\sqrt{k})$.