Accelerated Alternating Projections for Robust Principal Component Analysis

被引:0
作者
Cai, HanQin [1 ]
Cai, Jian-Feng [2 ]
Wei, Ke [3 ]
机构
[1] Univ Calif Los Angeles, Dept Math, Los Angeles, CA 90024 USA
[2] Hong Kong Univ Sci & Technol, Dept Math, Hong Kong, Peoples R China
[3] Fudan Univ, Sch Data Sci, Shanghai, Peoples R China
关键词
Robust PCA; Alternating Projections; Matrix Manifold; Tangent Space; Subspace Projection; RANK MATRIX COMPLETION; MINIMIZATION ALGORITHM;
D O I
暂无
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
We study robust PCA for the fully observed setting, which is about separating a low rank matrix L and a sparse matrix S from their sum D = L + S. In this paper, a new algorithm, dubbed accelerated alternating projections, is introduced for robust PCA which signi fi cantly improves the computational e ffi ciency of the existing alternating projections proposed in (Netrapalli et al., 2014) when updating the low rank factor. The acceleration is achieved by fi rst projecting a matrix onto some low dimensional subspace before obtaining a new estimate of the low rank matrix via truncated SVD. Exact recovery guarantee has been established which shows linear convergence of the proposed algorithm. Empirical performance evaluations establish the advantage of our algorithm over other state-of-theart algorithms for robust PCA.
引用
收藏
页数:33
相关论文
共 50 条
  • [31] Truncated Robust Principal Component Analysis and Noise Reduction for Single Cell RNA Sequencing Data
    Gogolewski, Krzysztof
    Sykulski, Maciej
    Chung, Neo Christopher
    Gambin, Anna
    JOURNAL OF COMPUTATIONAL BIOLOGY, 2019, 26 (08) : 782 - 793
  • [32] Alternating projections on manifolds
    Lewis, Adrian S.
    Malick, Jerome
    MATHEMATICS OF OPERATIONS RESEARCH, 2008, 33 (01) : 216 - 234
  • [33] Truncated Robust Principal Component Analysis and Noise Reduction for Single Cell RNA-seq Data
    Gogolewski, Krzysztof
    Sykulski, Maciej
    Chung, Neo Christopher
    Gambin, Anna
    BIOINFORMATICS RESEARCH AND APPLICATIONS, ISBRA 2018, 2018, 10847 : 335 - 346
  • [34] Automatic computation of regions of interest by robust principal component analysis. Application to automatic dementia diagnosis
    Lozano, Francisco
    Ortiz, Andres
    Munilla, Jorge
    Peinado, Alberto
    KNOWLEDGE-BASED SYSTEMS, 2017, 123 : 229 - 237
  • [35] Alternating Projections on Nontangential Manifolds
    Andersson, Fredrik
    Carlsson, Marcus
    CONSTRUCTIVE APPROXIMATION, 2013, 38 (03) : 489 - 525
  • [36] Strong Convergence of Alternating Projections
    Lira Melo, Italo Dowell
    da Cruz Neto, Joao Xavier
    Machado de Brito, Jose Marcio
    JOURNAL OF OPTIMIZATION THEORY AND APPLICATIONS, 2022, 194 (01) : 306 - 324
  • [37] Alternating Projections on Nontangential Manifolds
    Fredrik Andersson
    Marcus Carlsson
    Constructive Approximation, 2013, 38 : 489 - 525
  • [38] Strong Convergence of Alternating Projections
    Ítalo Dowell Lira Melo
    João Xavier da Cruz Neto
    José Márcio Machado de Brito
    Journal of Optimization Theory and Applications, 2022, 194 : 306 - 324
  • [39] Unlabeled Principal Component Analysis and Matrix Completion
    Yao, Yunzhen
    Peng, Liangzu
    Tsakiris, Manolis C.
    JOURNAL OF MACHINE LEARNING RESEARCH, 2024, 25 : 1 - 38
  • [40] Streaming Principal Component Analysis in Noisy Settings
    Marinov, Teodor, V
    Mianjy, Poorya
    Arora, Raman
    INTERNATIONAL CONFERENCE ON MACHINE LEARNING, VOL 80, 2018, 80