English

Frank-Wolfe with Subsampling Oracle

Optimization and Control 2018-03-21 v1 Machine Learning Machine Learning

Abstract

We analyze two novel randomized variants of the Frank-Wolfe (FW) or conditional gradient algorithm. While classical FW algorithms require solving a linear minimization problem over the domain at each iteration, the proposed method only requires to solve a linear minimization problem over a small \emph{subset} of the original domain. The first algorithm that we propose is a randomized variant of the original FW algorithm and achieves a O(1/t)\mathcal{O}(1/t) sublinear convergence rate as in the deterministic counterpart. The second algorithm is a randomized variant of the Away-step FW algorithm, and again as its deterministic counterpart, reaches linear (i.e., exponential) convergence rate making it the first provably convergent randomized variant of Away-step FW. In both cases, while subsampling reduces the convergence rate by a constant factor, the linear minimization step can be a fraction of the cost of that of the deterministic versions, especially when the data is streamed. We illustrate computational gains of the algorithms on regression problems, involving both 1\ell_1 and latent group lasso penalties.

Keywords

Cite

@article{arxiv.1803.07348,
  title  = {Frank-Wolfe with Subsampling Oracle},
  author = {Thomas Kerdreux and Fabian Pedregosa and Alexandre d'Aspremont},
  journal= {arXiv preprint arXiv:1803.07348},
  year   = {2018}
}
R2 v1 2026-06-23T00:58:40.460Z