arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1809.10477cs.LGmath.OCstat.ML
收藏

低秩非光滑矩阵问题的快速随机算法

Fast Stochastic Algorithms for Low-rank and Nonsmooth Matrix Problems

Technion - Israel Institute of Technology(以色列理工学院)

Dan Garber, Atara Kaplan

摘要:

同时包含非光滑项和低秩提升项的复合凸优化问题在机器学习和信号处理中具有重要的应用,例如当人们希望恢复同时是低秩和稀疏的未知矩阵时。然而,这样的问题是非常具有挑战性的大规模解决:低秩提升项禁止有效实现复合优化的近似方法,甚至简单的次梯度方法。另一方面,为低秩优化量身定制的方法,例如条件梯度型方法,其通常应用于非光滑目标的光滑近似,是缓慢的,因为它们的运行时间尺度与平滑梯度向量的大Lipshitz参数和$1/ε$。在本文中,我们开发了有效的算法\textit{stochastic}优化的强凸目标,其中包括一个非光滑项和一个低秩促进长期。特别是,据我们所知,我们提出了第一个算法,享有所有以下关键属性的大规模问题:i)(接近)最佳的样本复杂性,ii)每次迭代只需要一个单一的\textit{low-rank} SVD计算,iii)总数量的瘦SVD计算规模仅与$\log{1/ε}$(而不是$\textrm{poly}(1/ε)$在以前的方法)。我们还给出了一个算法的密切相关的有限和设置。在我们的研究结果的核心在于一个新的组合的方差减少技术和使用的\textit{weak-proximal oracle},这是关键,同时获得所有上述三个属性。

英文摘要:

Composite convex optimization problems which include both a nonsmooth term and a low-rank promoting term have important applications in machine learning and signal processing, such as when one wishes to recover an unknown matrix that is simultaneously low-rank and sparse. However, such problems are highly challenging to solve in large-scale: the low-rank promoting term prohibits efficient implementations of proximal methods for composite optimization and even simple subgradient methods. On the other hand, methods which are tailored for low-rank optimization, such as conditional gradient-type methods, which are often applied to a smooth approximation of the nonsmooth objective, are slow since their runtime scales with both the large Lipshitz parameter of the smoothed gradient vector and with $1/ε$. In this paper we develop efficient algorithms for \textit{stochastic} optimization of a strongly-convex objective which includes both a nonsmooth term and a low-rank promoting term. In particular, to the best of our knowledge, we present the first algorithm that enjoys all following critical properties for large-scale problems: i) (nearly) optimal sample complexity, ii) each iteration requires only a single \textit{low-rank} SVD computation, and iii) overall number of thin-SVD computations scales only with $\log{1/ε}$ (as opposed to $\textrm{poly}(1/ε)$ in previous methods). We also give an algorithm for the closely-related finite-sum setting. At the heart of our results lie a novel combination of a variance-reduction technique and the use of a \textit{weak-proximal oracle} which is key to obtaining all above three properties simultaneously.

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。

↑