An efficient randomized QLP algorithm for approximating the singular value decomposition

DOI10.1016/J.INS.2023.119464MaRDI QIDQ6052614zbMATH OpenOpenAlexFDO

Authors

Publication date 22 September 2023

Published in Information Sciences (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/2110.01011

zbMATH Keywords

communication-avoiding algorithmhigh-performance computingpivoted QLPrandomized samplingsingular value decomposition

Mathematics Subject Classification ID

Numerical computation of eigenvalues and eigenvectors of matrices (65F15) Orthogonalization in numerical linear algebra (65F25)

Abstract: In this paper, we introduce a randomized QLP decomposition called Rand-QLP. Operating on a matrix $$, Rand-QLP gives $$, where $$ and $$ are orthonormal, and $$ is lower-triangular. Under the assumption that the rank of the input matrix is $k$, we derive several error bounds for Rand-QLP: bounds for the first $k$ approximate singular values and for the trailing block of the middle factor $$, which show that the decomposition is rank-revealing; bounds for the distance between approximate subspaces and the exact ones for all four fundamental subspaces of a given matrix; and bounds for the errors of low-rank approximations constructed by the columns of $$ and $$. Rand-QLP is able to effectively leverage modern computational architectures, due to the utilization of random sampling and the unpivoted QR decomposition, thus addressing a serious bottleneck associated with classical algorithms such as the singular value decomposition (SVD), column-pivoted QR (CPQR) and most recent matrix decomposition algorithms. To assess the performance behavior of different algorithms, we use an Intel Xeon Gold 6240 CPU running at 2.6 GHz with a NVIDIA GeForce RTX 2080Ti GPU. In comparison to CPQR and the SVD, Rand-QLP respectively achieves a speedup of up to 5 times and 6.6 times on the CPU and up to 3.8 times and 4.4 times with the hybrid GPU architecture. In terms of quality of approximation, our results on synthetic and real data show that the approximations by Rand-QLP are comparable to those of pivoted QLP and the optimal SVD, and in most cases are considerably better than those of CPQR.

Recommendations

• Randomized QLP decomposition
• Single-pass randomized QLP decomposition for low-rank approximation
• Stewart's pivoted QLP decomposition for low‐rank matrices
• The QLP Approximation to the Singular Value Decomposition
• Randomized quaternion QLP decomposition for low-rank approximation

Cites work

• A BLAS-3 Version of the QR Factorization with Column Pivoting
• A block QR algorithm and the singular value decomposition
• A set of level 3 basic linear algebra subprograms
• Algorithm 656: an extended set of basic linear algebra subprograms: model implementation and test programs
• An efficient algorithm for computing the approximate t-URV and its applications
• Basic Linear Algebra Subprograms for Fortran Usage
• Communication avoiding rank revealing QR factorization with column pivoting
• Determining the number of communities in degree-corrected stochastic block models
• Efficient Low-Rank Approximation of Matrices Based on Randomized Pivoted Decomposition
• Efficient randomized algorithms for the fixed-precision low-rank matrix approximation
• Functions of Matrices
• scientific article; zbMATH DE number 47363 (Why is no real title available?)
• scientific article; zbMATH DE number 1012640 (Why is no real title available?)
• scientific article; zbMATH DE number 961607 (Why is no real title available?)
• Low-rank revealing \(UTV\) decompositions
• Matrix mathematics. Theory, facts, and formulas
• Numerical methods in matrix computations
• On the convergence of Stewart's QLP algorithm for approximating the SVD
• Parallel QR Factorization of Block Low-rank Matrices
• Practical sketching algorithms for low-rank matrix approximation
• Principal regression for high dimensional covariance matrices
• Projection-Based QLP Algorithm for Efficiently Computing Low-Rank Approximation of Matrices
• randUTV: a blocked randomized algorithm for computing a rank-revealing UTV factorization
• Rang revealing QR factorizations
• Singular value decomposition approximation via Kronecker summations for imaging applications
• Space-time reduced order model for large-scale linear dynamical systems with application to Boltzmann transport problems
• Stable and efficient spectral divide and conquer algorithms for the symmetric eigenvalue decomposition and the SVD
• Subspace-Orbit Randomized Decomposition for Low-Rank Matrix Approximations
• The QLP Approximation to the Singular Value Decomposition
• The singular value decomposition: anatomy of optimizing an algorithm for extreme scale
• The University of Florida sparse matrix collection
• Updating a Rank-Revealing ULV Decomposition

Cited in

(2)

• Rank-deficient prewhitening with quotient SVD and ULV decompositions
• On the convergence of Stewart's QLP algorithm for approximating the SVD

This page was built for publication: An efficient randomized QLP algorithm for approximating the singular value decomposition

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6052614)