arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1802.07372math.OCcs.LGstat.ML
收藏

非凸Optimization的随机减方差三次正则化

Stochastic Variance-Reduced Cubic Regularization for Nonconvex Optimization

Ohio State University(俄亥俄州立大学);Georgia Institute of Technology(佐治亚理工学院)

Zhe Wang, Yi Zhou, Yingbin Liang, Guanghui Lan

摘要:

三次正则化(CR)是一种最优化方法,由于其能够逃逸鞍点并收敛于非凸优化的二阶稳定解而受到广泛关注。然而,CR遇到了一个高样本复杂度的问题,有限和问题,大的数据大小。已提出CR的各种不精确变体以改善样本复杂性。本文提出了一种随机抽样下的随机方差缩减的正则化方法,并研究了它的收敛性保证和样本复杂度。我们表明,SVRC实现$ε$精度内的二阶稳态解的迭代复杂度为$O(ε^{-3/2})$,这与CR类型方法的最新结果相匹配。此外,我们提出的方差缩减方案显着降低了每次迭代样本的复杂度。我们的SVRC得到的总Hessian样本复杂度是${\Oc}(N^{2/3} ε^{-3/2})$,它比最先进的结果好$O(N^{2/15})$倍。我们还研究了我们的SVRC下的随机不放回抽样方案,它产生了较低的每次迭代的样本复杂度,因此证明了其实用性。

英文摘要:

Cubic regularization (CR) is an optimization method with emerging popularity due to its capability to escape saddle points and converge to second-order stationary solutions for nonconvex optimization. However, CR encounters a high sample complexity issue for finite-sum problems with a large data size. %Various inexact variants of CR have been proposed to improve the sample complexity. In this paper, we propose a stochastic variance-reduced cubic-regularization (SVRC) method under random sampling, and study its convergence guarantee as well as sample complexity. We show that the iteration complexity of SVRC for achieving a second-order stationary solution within $ε$ accuracy is $O(ε^{-3/2})$, which matches the state-of-art result on CR types of methods. Moreover, our proposed variance reduction scheme significantly reduces the per-iteration sample complexity. The resulting total Hessian sample complexity of our SVRC is ${\Oc}(N^{2/3} ε^{-3/2})$, which outperforms the state-of-art result by a factor of $O(N^{2/15})$. We also study our SVRC under random sampling without replacement scheme, which yields a lower per-iteration sample complexity, and hence justifies its practical applicability.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑