Stochastic trust-region algorithm in random subspaces with convergence and expected complexity analyses

TL;DR

This paper introduces STARS, a scalable stochastic trust-region method that optimizes large-scale problems efficiently by working in random low-dimensional subspaces, with proven convergence and complexity guarantees.

Contribution

It extends stochastic derivative-free optimization by incorporating random subspace minimization, reducing computational costs while maintaining convergence guarantees.

Findings

STARS achieves convergence to a stationary point.

Expected iteration complexity is comparable to existing methods.

Subspace dimension is independent of problem size.

Abstract

This work proposes a framework for large-scale stochastic derivative-free optimization (DFO) by introducing STARS, a trust-region method based on iterative minimization in random subspaces. This framework is both an algorithmic and theoretical extension of an algorithm for stochastic optimization with random models (STORM). Moreover, STARS achieves scalability by minimizing interpolation models that approximate the objective in low-dimensional affine subspaces, thus significantly reducing per-iteration costs in terms of function evaluations and yielding strong performance on large-scale stochastic DFO problems. The user-determined dimension of these subspaces, when the latter are defined, for example, by the columns of so-called Johnson--Lindenstrauss transforms, turns out to be independent of the dimension of the problem. For convergence purposes, both a particular quality of the subspace and the accuracies of random function estimates and models are required to hold with sufficiently high, but fixed, probabilities. Using martingale theory under the latter assumptions, an almost sure global convergence of STARS to a first-order stationary point is shown, and the expected number of iterations required to reach a desired first-order accuracy is proved to be similar to that of STORM and other stochastic DFO algorithms, up to constants.