Static and Dynamic Robust PCA and Matrix Completion: A Review

Abstract

Principal Components Analysis (PCA) is one of the most widely used dimension reduction techniques. Robust PCA (RPCA) refers to the problem of PCA when the data may be corrupted by outliers. Recent work by Candès, Wright, Li, and Ma defined RPCA as a problem of decomposing a given data matrix into the sum of a low-rank matrix (true data) and a sparse matrix (outliers). The column space of the low-rank matrix then gives the PCA solution. This simple definition has lead to a large amount of interesting new work on provably correct, fast, and practical solutions to RPCA. More recently, the dynamic (time-varying) version of the RPCA problem has been studied and a series of provably correct, fast, and memory efficient tracking solutions have been proposed. Dynamic RPCA (or robust subspace tracking) is the problem of tracking data lying in a (slowly) changing subspace, while being robust to sparse outliers. This article provides an exhaustive review of the last decade of literature on RPCA and its dynamic counterpart (robust subspace tracking), along with describing their theoretical guarantees, discussing the pros and cons of various approaches, and providing empirical comparisons of performance and speed.A brief overview of the (low-rank) matrix completion literature is also provided (the focus is on works not discussed in other recent reviews). This refers to the problem of completing a lowrank matrix when only a subset of its entries are observed. It can be interpreted as a simpler special case of RPCA in which the indices of the outlier corrupted entries are known.

The medical AI insurgency: what physicians must know about data to practice with intelligent machines

“…Computing run times of simple AI diagnostic modules (i.e., auto-encoders and support vector machines), advanced convolutional or recurrent neural networks, and other DL algorithms depend greatly on the dimensionality of the inputted data. To achieve computing efficiencies, data scientists use a wide variety of dimensionality reduction and feature selection techniques: principal component analysis (PCA), 2 generalized spike models, 22 robust PCA, 23,24 PCA whitening, 25 robust subspace tracking, 24 low rank plus sparse [L + S] data decomposition 26 and algorithms (i.e., t -distributed stochastic neighbor embedding). 27 With proper preprocessing of dynamic datasets, AI technologies are becoming more efficient at signal processing (i.e., satellite communications and seismology), computer vision (i.e., video surveillance and traffic patterns), and network traffic analysis.…”

A Robust Principal Component Analysis for Outlier Identification in Messy Microcalorimeter Data

“…The search for robust methods of PCA is a very active area of research in statistics and machine learning. 19 This search is sometimes divided into robust PCA, where individual matrix elements of M are outliers, 20 and robust subspace recovery, where entire columns may be outliers. 21 We have considered L1-norm PCA, 22,23,24,25 in which principal components minimize not the variance of the data residuals (the L2 norm) but the absolute deviation of the residuals (the L1 norm).…”

Point Cloud Denoising with Principal Component Analysis and a Novel Bilateral Filter

“…Ali et al [16] review the major tensor decomposition methods with a focus on problems targeted by classical PCA. Namrata et al [17] studied the dynamic (time-varying) version of the RPCA problem and proposed a series of provably correct, fast, and memoryefficient tracking solutions. The above studies have improved the normal vector estimation of the point cloud to some extent.…”