English

Stochastic Cubic Regularization for Fast Nonconvex Optimization

Machine Learning 2017-12-07 v2 Optimization and Control Machine Learning

Abstract

This paper proposes a stochastic variant of a classic algorithm---the cubic-regularized Newton method [Nesterov and Polyak 2006]. The proposed algorithm efficiently escapes saddle points and finds approximate local minima for general smooth, nonconvex functions in only O~(ϵ3.5)\mathcal{\tilde{O}}(\epsilon^{-3.5}) stochastic gradient and stochastic Hessian-vector product evaluations. The latter can be computed as efficiently as stochastic gradients. This improves upon the O~(ϵ4)\mathcal{\tilde{O}}(\epsilon^{-4}) rate of stochastic gradient descent. Our rate matches the best-known result for finding local minima without requiring any delicate acceleration or variance-reduction techniques.

Keywords

Cite

@article{arxiv.1711.02838,
  title  = {Stochastic Cubic Regularization for Fast Nonconvex Optimization},
  author = {Nilesh Tripuraneni and Mitchell Stern and Chi Jin and Jeffrey Regier and Michael I. Jordan},
  journal= {arXiv preprint arXiv:1711.02838},
  year   = {2017}
}

Comments

The first two authors contributed equally

R2 v1 2026-06-22T22:39:41.681Z