ResearchTrend.AI
  • Papers
  • Communities
  • Events
  • Blog
  • Pricing
Papers
Communities
Social Events
Terms and Conditions
Pricing
Parameter LabParameter LabTwitterGitHubLinkedInBlueskyYoutube

© 2025 ResearchTrend.AI, All rights reserved.

  1. Home
  2. Papers
  3. 1702.02267
29
22

Matrix Completion from O(n)O(n)O(n) Samples in Linear Time

8 February 2017
D. Gamarnik
Quan Li
Hongyi Zhang
ArXivPDFHTML
Abstract

We consider the problem of reconstructing a rank-kkk n×nn \times nn×n matrix MMM from a sampling of its entries. Under a certain incoherence assumption on MMM and for the case when both the rank and the condition number of MMM are bounded, it was shown in \cite{CandesRecht2009, CandesTao2010, keshavan2010, Recht2011, Jain2012, Hardt2014} that MMM can be recovered exactly or approximately (depending on some trade-off between accuracy and computational complexity) using O(n poly(log⁡n))O(n \, \text{poly}(\log n))O(npoly(logn)) samples in super-linear time O(na poly(log⁡n))O(n^{a} \, \text{poly}(\log n))O(napoly(logn)) for some constant a≥1a \geq 1a≥1. In this paper, we propose a new matrix completion algorithm using a novel sampling scheme based on a union of independent sparse random regular bipartite graphs. We show that under the same conditions w.h.p. our algorithm recovers an ϵ\epsilonϵ-approximation of MMM in terms of the Frobenius norm using O(nlog⁡2(1/ϵ))O(n \log^2(1/\epsilon))O(nlog2(1/ϵ)) samples and in linear time O(nlog⁡2(1/ϵ))O(n \log^2(1/\epsilon))O(nlog2(1/ϵ)). This provides the best known bounds both on the sample complexity and computational complexity for reconstructing (approximately) an unknown low-rank matrix. The novelty of our algorithm is two new steps of thresholding singular values and rescaling singular vectors in the application of the "vanilla" alternating minimization algorithm. The structure of sparse random regular graphs is used heavily for controlling the impact of these regularization steps.

View on arXiv
Comments on this paper