Abstract:Randomized singular value decomposition (RSVD) is a class of computationally efficient algorithms for computing the truncated SVD of large data matrices. Given a $n \times n$ symmetric matrix $\mathbf{M}$, the prototypical RSVD algorithm outputs an approximation of the $k$ leading singular vectors of $\mathbf{M}$ by computing the SVD of $\mathbf{M}^{g} \mathbf{G}$; here $g \geq 1$ is an integer and $\mathbf{G} \in \mathbb{R}^{n \times k}$ is a random Gaussian sketching matrix. In this paper we study the statistical properties of RSVD under a general "signal-plus-noise" framework, i.e., the observed matrix $\hat{\mathbf{M}}$ is assumed to be an additive perturbation of some true but unknown signal matrix $\mathbf{M}$. We first derive upper bounds for the $\ell_2$ (spectral norm) and $\ell_{2\to\infty}$ (maximum row-wise $\ell_2$ norm) distances between the approximate singular vectors of $\hat{\mathbf{M}}$ and the true singular vectors of the signal matrix $\mathbf{M}$. These upper bounds depend on the signal-to-noise ratio (SNR) and the number of power iterations $g$. A phase transition phenomenon is observed in which a smaller SNR requires larger values of $g$ to guarantee convergence of the $\ell_2$ and $\ell_{2\to\infty}$ distances. We also show that the thresholds for $g$ where these phase transitions occur are sharp whenever the noise matrices satisfy a certain trace growth condition. Finally, we derive normal approximations for the row-wise fluctuations of the approximate singular vectors and the entrywise fluctuations of the approximate matrix. We illustrate our theoretical results by deriving nearly-optimal performance guarantees for RSVD when applied to three statistical inference problems, namely, community detection, matrix completion, and principal component analysis with missing data.

Matrix perturbation analysis of methods for extracting singular values from approximate singular subspaces

An exact $\sinΘ$ formula for matrix perturbation analysis and its applications

Perturbation expansions and error bounds for the truncated singular value decomposition

Perturbation Analysis of Randomized SVD and its Applications to High-dimensional Statistics

Perturbation Analysis for Subspace Decomposition with Applications in Subspace-Based Algorithms

High Order Numerical Methods To Approximate The Singular Value Decomposition

Singular Subspace Perturbation Bounds via Rectangular Random Matrix Diffusions

Analysis of singular subspaces under random perturbations

A Schatten-$q$ Low-rank Matrix Perturbation Analysis via Perturbation Projection Error Bound

Matrices with Gaussian Noise: Optimal Estimates for Singular Subspace Perturbation

Efficient Bounds and Estimates for Canonical Angles in Randomized Subspace Approximations

An Interval Perturbation Method for Singular Value Decomposition (SVD) with Unknown-but-bounded (UBB) Parameters

Probabilistic perturbation bounds of matrix decompositions

Randomized Nyström approximation of non-negative self-adjoint operators

Relative perturbation bounds for the eigenvalues of diagonalizable and singular matrices – Application of perturbation theory for simple invariant subspaces

Large-scale Nyström kernel matrix approximation using randomized SVD.

A universal matrix perturbation technique for structural dynamic modification using singular value decomposition

Matrix Perturbation: Davis-Kahan in the Infinity Norm

A note on perturbation analysis for T-product based tensor singular values

Small singular values can increase in lower precision