Using Taylor-Approximated Gradients to Improve the Frank-Wolfe Method for Empirical Risk Minimization

TL;DR

This paper introduces Taylor-approximated gradient methods to enhance the Frank-Wolfe algorithm, significantly reducing computational dependence on data size while maintaining optimal convergence in empirical risk minimization tasks.

Contribution

It proposes a novel Taylor series-based gradient approximation for Frank-Wolfe, applicable in both deterministic and stochastic settings, with adaptive step-size and proven computational guarantees.

Findings

Significant speed-ups over existing methods on real datasets

Reduced dependence on data size in large-scale problems

Achieved optimal convergence rates in convex and non-convex settings

Abstract

The Frank-Wolfe method has become increasingly useful in statistical and machine learning applications, due to the structure-inducing properties of the iterates, and especially in settings where linear minimization over the feasible set is more computationally efficient than projection. In the setting of Empirical Risk Minimization -- one of the fundamental optimization problems in statistical and machine learning -- the computational effectiveness of Frank-Wolfe methods typically grows linearly in the number of data observations $n$. This is in stark contrast to the case for typical stochastic projection methods. In order to reduce this dependence on $n$, we look to second-order smoothness of typical smooth loss functions (least squares loss and logistic loss, for example) and we propose amending the Frank-Wolfe method with Taylor series-approximated gradients, including variants for both deterministic and stochastic settings. Compared with current state-of-the-art methods in the regime where the optimality tolerance $\varepsilon$ is sufficiently small, our methods are able to simultaneously reduce the dependence on large $n$ while obtaining optimal convergence rates of Frank-Wolfe methods, in both the convex and non-convex settings. We also propose a novel adaptive step-size approach for which we have computational guarantees. Last of all, we present computational experiments which show that our methods exhibit very significant speed-ups over existing methods on real-world datasets for both convex and non-convex binary classification problems.