The eigenvalues and eigenvectors of finite, low rank perturbations of large random matrices

DOI10.1016/J.AIM.2011.02.007MaRDI QIDQ531808zbMATH OpenOpenAlexFDO

Authors Florent Benaych-Georges, Raj Rao Nadakuditi

Publication date 20 April 2011

Published in Advances in Mathematics (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/0910.2120

zbMATH Keywords

phase transitionprincipal components analysiseigenvalue distributionrandom matricesfree probabilityrandom perturbationsymmetric matrixHaar measurerandom eigenvaluesrandom eigenvectorssample covariance matricesinformational limit

Mathematics Subject Classification ID

Factor analysis and principal components; correspondence analysis (62H25) Eigenvalues, singular values, and eigenvectors (15A18) Random matrices (algebraic aspects) (15B52) Random matrices (probabilistic aspects) (60B20) Phase transitions (general) in equilibrium statistical mechanics (82B26) Free probability and free operator algebras (46L54)

Abstract: We consider the eigenvalues and eigenvectors of finite, low rank perturbations of random matrices. Specifically, we prove almost sure convergence of the extreme eigenvalues and appropriate projections of the corresponding eigenvectors of the perturbed matrix for additive and multiplicative perturbation models. The limiting non-random value is shown to depend explicitly on the limiting eigenvalue distribution of the unperturbed random matrix and the assumed perturbation model via integral transforms that correspond to very well known objects in free probability theory that linearize non-commutative free additive and multiplicative convolution. Furthermore, we uncover a phase transition phenomenon whereby the large matrix limit of the extreme eigenvalues of the perturbed matrix differs from that of the original matrix if and only if the eigenvalues of the perturbing matrix are above a certain critical threshold. Square root decay of the eigenvalue density at the edge is sufficient to ensure that this threshold is finite. This critical threshold is intimately related to the same aforementioned integral transforms and our proof techniques bring this connection and the origin of the phase transition into focus. Consequently, our results extend the class of `spiked' random matrix models about which such predictions (called the BBP phase transition) can be made well beyond the Wigner, Wishart and Jacobi random ensembles found in the literature. We examine the impact of this eigenvalue phase transition on the associated eigenvectors and observe an analogous phase transition in the eigenvectors. Various extensions of our results to the problem of non-extreme eigenvalues are discussed.

Recommendations

• Mesoscopic perturbations of large random matrices
• The singular values and vectors of low rank perturbations of large rectangular random matrices
• Limits of spiked random matrices. II
• Large deviations for the largest eigenvalues and eigenvectors of spiked Gaussian random matrices
• Limits of spiked random matrices. I

Cites work

• scientific article; zbMATH DE number 47363 (Why is no real title available?)
• scientific article; zbMATH DE number 1465044 (Why is no real title available?)
• scientific article; zbMATH DE number 3244317 (Why is no real title available?)
• A Fourier view on the \(R\)-transform and related asymptotics of spherical integrals
• An introduction to random matrices
• Analysis of the limiting spectral distribution of large dimensional random matrices
• Asymptotics of sample eigenstructure for a large dimensional spiked covariance model
• Character expansion method for the first order asymptotics of a matrix integral
• Eigenvalue separation in some random matrix models
• Eigenvalues of large sample covariance matrices of spiked population models
• Finite sample approximation results for principal component analysis: A matrix perturbation approach
• Free Random Variables
• Generic behavior of the density of states in random matrix theory and equilibrium problems in the presence of real analytic external fields
• Infinitely divisible distributions for rectangular free convolution: classification and matricial interpretation
• Integration with respect to the Haar measure on unitary, orthogonal and symplectic group
• Large deviations of the extreme eigenvalues of random deformations of matrices
• Matrix Analysis
• On a surprising relation between the Marchenko-Pastur law, rectangular and square free convolutions
• On the distribution of the roots of certain symmetric matrices
• On the eigenvectors of large dimensional sample covariance matrices
• Phase transition of the largest eigenvalue for nonnull complex sample covariance matrices
• Product of random projections, Jacobi ensembles and universality problems arising from free probability
• Rank-one modification of the symmetric eigenproblem
• Rectangular random matrices, related convolution
• Refined perturbation bounds for eigenvalues of Hermitian and non-Hermitian matrices
• Restricted rank modification of the symmetric eigenvalue problem: Theoretical considerations
• Some limit theorems on the eigenvectors of large dimensional sample covariance matrices
• The concentration of measure phenomenon
• The largest eigenvalue of rank one deformation of large Wigner matrices
• The largest eigenvalue of small rank perturbations of Hermitian random matrices
• The largest eigenvalues of finite rank deformation of large Wigner matrices: Convergence and nonuniversality of the fluctuations
• The polynomial method for random matrices
• The singular values and vectors of low rank perturbations of large rectangular random matrices
• Tracy-Widom limit for the largest eigenvalue of a large class of complex sample covariance matrices
• Uniform asymptotics for polynomials orthogonal with respect to varying exponential weights and applications to universality questions in random matrix theory

Cited in

(only showing first 100 items - show all)

• Covariance estimation for distributions with \({2+\varepsilon}\) moments
• Outlier eigenvalues for deformed i.i.d. random matrices
• Large deviations for the largest eigenvalues and eigenvectors of spiked Gaussian random matrices
• Spectra of nearly Hermitian random matrices
• On sample eigenvalues in a generalized spiked population model
• Limiting distribution of the sample canonical correlation coefficients of high-dimensional random vectors
• Eigenvectors and controllability of non-Hermitian random matrices and directed graphs
• Spectral statistics of Erdős-Rényi graphs. I: Local semicircle law
• Circular law for noncentral random matrices
• Large deviations of the extreme eigenvalues of random deformations of matrices
• Singular vectors under random perturbation
• Eigenvector delocalization for non‐Hermitian random matrices and applications
• A note on the CLT of the LSS for sample covariance matrix from a spiked population model
• Local stability of the free additive convolution
• Asymptotic mutual information for the balanced binary stochastic block model
• Spectral statistics of sparse Erdős-Rényi graph Laplacians
• Outliers in the single ring theorem
• Outliers in the spectrum for products of independent random matrices
• The isotropic semicircle law and deformation of Wigner matrices
• Phase transition in the spiked random tensor with Rademacher prior
• On finite rank deformations of Wigner matrices
• Strong convergence of the empirical distribution of eigenvalues of sample covariance matrices with a perturbation matrix
• On the eigenvalues of specially low-rank perturbed matrices
• On singular values of large dimensional lag-\(\tau\) sample auto-correlation matrices
• On finite rank deformations of Wigner matrices. II: Delocalized perturbations
• Random perturbation of low rank matrices: improving classical bounds
• Principal components in linear mixed models with general bulk
• Complex outliers of Hermitian random matrices
• Estimation of low-rank matrices via approximate message passing
• \(e\)PCA: high dimensional exponential family PCA
• Detection, reconstruction, and characterization algorithms from noisy data in multistatic wave imaging
• Eigenvectors of random matrices: A survey
• The outliers of a deformed Wigner matrix
• A note on the prediction error of principal component regression in high dimensions
• On bilinear forms based on the resolvent of large random matrices
• Batch latency analysis and phase transitions for a tandem of queues with exponentially distributed service times
• Detection thresholds in very sparse matrix completion
• Empirical spectral distribution of a matrix under perturbation
• A survey on the eigenvalues local behavior of large complex correlated Wishart matrices
• Target detection and characterization from electromagnetic induction data
• Kernel spectral clustering of large dimensional data
• On the principal components of sample covariance matrices
• Canonical correlation coefficients of high-dimensional Gaussian vectors: finite rank case
• Spectral measures of spiked random matrices
• Fundamental limits of symmetric low-rank matrix estimation
• Robust spiked random matrices and a robust G-MUSIC estimator
• Limits of spiked random matrices. I
• An Eigenvalue Ratio Approach to Inferring Population Structure from Whole Genome Sequencing Data
• Limits of spiked random matrices. II
• Eigenvectors and eigenvalues in a random subspace of a tensor product
• Large bidiagonal matrices and random perturbations
• Subordination for the sum of two random matrices
• Cleaning large correlation matrices: tools from random matrix theory
• The largest eigenvalue of real symmetric, Hermitian and Hermitian self-dual random matrix models with rank one external source. I
• Large complex correlated Wishart matrices: fluctuations and asymptotic independence at the edges
• Low rank perturbations of large elliptic random matrices
• Additive/multiplicative free subordination property and limiting eigenvectors of spiked additive deformations of Wigner matrices and spiked sample covariance matrices
• The largest eigenvalue of small rank perturbations of Hermitian random matrices
• On the spectrum of random anti-symmetric and tournament matrices
• Separation of the largest eigenvalues in eigenanalysis of genotype data from discrete subpopulations
• Dynamics of a rank-one perturbation of a Hermitian matrix
• Statistical limits of spiked tensor models
• Optimality and sub-optimality of PCA. I: Spiked random matrix models
• Optimal shrinkage of eigenvalues in the spiked covariance model
• The singular values and vectors of low rank perturbations of large rectangular random matrices
• Singular vector distribution of sample covariance matrices
• Random matrix theory in statistics: a review
• High-dimensional dynamics of generalization error in neural networks
• Eigenvalue distributions of variance components estimators in high-dimensional random effects models
• The two-to-infinity norm and singular subspace geometry with applications to high-dimensional statistics
• Entrywise eigenvector analysis of random matrices with low expected rank
• Sequential testing for structural stability in approximate factor models
• Universality and sharp matrix concentration inequalities
• Eigenvalues outside the bulk of inhomogeneous Erdős-Rényi random graphs
• scientific article; zbMATH DE number 7415122 (Why is no real title available?)
• Fundamental limits of low-rank matrix estimation with diverging aspect ratios
• Eigenvalues of stochastic blockmodel graphs and random graphs with low-rank edge probability matrices
• Rank 1 perturbations in random matrix theory — A review of exact results
• Solving orthogonal group synchronization via convex and low-rank optimization: tightness and landscape analysis
• A Unifying Tutorial on Approximate Message Passing
• Near-optimal bounds for phase synchronization
• High‐dimensional limit theorems for SGD: Effective dynamics and critical scaling
• Universality of approximate message passing algorithms
• Free probability of type B and asymptotics of finite-rank perturbations of random matrices
• Spiked sample covariance matrices with possibly multiple bulk components
• Free energy subadditivity for symmetric random Hamiltonians
• Freeness of type B and conditional freeness for random matrices
• High dimensional deformed rectangular matrices with applications in matrix denoising
• Approximate message passing algorithms for rotationally invariant matrices
• Outliers in spectrum of sparse Wigner matrices
• Robust high-dimensional factor models with applications to statistical machine learning
• Distribution of rare saddles in the p -spin energy landscape
• Spiked multiplicative random matrices and principal components
• The statistics of spectral shifts due to finite rank perturbations
• Singular vector and singular subspace distribution for the matrix denoising model
• Efficient estimation of linear functionals of principal components
• Principal eigenportfolios for U.S. equities
• First eigenvalue/eigenvector in sparse random symmetric matrices: influences of degree fluctuation
• Order Determination for Spiked Type Models
• Relative perturbation bounds with applications to empirical covariance operators

Describes a project that uses

Uses Software

• RMTool

This page was built for publication: The eigenvalues and eigenvectors of finite, low rank perturbations of large random matrices

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q531808)