BriefGPT.xyz
Jun, 2016
近乎最优的鲁棒矩阵完成
Nearly-optimal Robust Matrix Completion
HTML
PDF
Yeshwanth Cherapanamjeri, Kartik Gupta, Prateek Jain
TL;DR
本文提出了一种简单的投影梯度下降方法来估计低秩矩阵,用于解决鲁棒矩阵完成问题,并且包括清除一些受损条目的步骤,并在低秩矩阵完成问题中获得了最优观测次数和最优破坏次数的解决方法。同时,本文的结果还意味着,对于低秩矩阵完成问题的时间复杂度界限,取得了重要的改进。最后,通过将结果应用于鲁棒PCA问题,得到了高效的解决方案
Abstract
In this paper, we consider the problem of
robust matrix completion
(RMC) where the goal is to recover a
low-rank matrix
by observing a small number of its entries out of which a few can be arbitrarily corrupted.
→