English

Stochastic Variance-Reduced Cubic Regularized Newton Method

Machine Learning 2018-02-14 v1 Optimization and Control

Abstract

We propose a stochastic variance-reduced cubic regularized Newton method for non-convex optimization. At the core of our algorithm is a novel semi-stochastic gradient along with a semi-stochastic Hessian, which are specifically designed for cubic regularization method. We show that our algorithm is guaranteed to converge to an (ϵ,ϵ)(\epsilon,\sqrt{\epsilon})-approximately local minimum within O~(n4/5/ϵ3/2)\tilde{O}(n^{4/5}/\epsilon^{3/2}) second-order oracle calls, which outperforms the state-of-the-art cubic regularization algorithms including subsampled cubic regularization. Our work also sheds light on the application of variance reduction technique to high-order non-convex optimization methods. Thorough experiments on various non-convex optimization problems support our theory.

Keywords

Cite

@article{arxiv.1802.04796,
  title  = {Stochastic Variance-Reduced Cubic Regularized Newton Method},
  author = {Dongruo Zhou and Pan Xu and Quanquan Gu},
  journal= {arXiv preprint arXiv:1802.04796},
  year   = {2018}
}

Comments

16 pages, 3 figures, 2 tables

R2 v1 2026-06-23T00:21:25.111Z