Abstract:Randomized algorithms for low-rank matrix approximation are investigated, with the emphasis on the fixed-precision problem and computational efficiency for handling large matrices. The algorithms are based on the so-called QB factorization, where Q is an orthonormal matrix. First, a mechanism for calculating the approximation error in the Frobenius norm is proposed, which enables efficient adaptive rank determination for a large and/or sparse matrix. It can be combined with any QB-form factorization algorithm in which B's rows are incrementally generated. Based on the blocked randQB algorithm by Martinsson and Voronin, this results in an algorithm called randQB_EI. Then, we further revise the algorithm to obtain a pass-efficient algorithm, randQB_FP, which is mathematically equivalent to the existing randQB algorithms and also suitable for the fixed-precision problem. Especially, randQB_FP can serve as a single-pass algorithm for calculating leading singular values, under a certain condition. With large and/or sparse test matrices, we have empirically validated the merits of the proposed techniques, which exhibit remarkable speedup and memory saving over the blocked randQB algorithm. We have also demonstrated that the single-pass algorithm derived by randQB_FP is much more accurate than an existing single-pass algorithm. And with data from a scenic image and an information retrieval application, we have shown the advantages of the proposed algorithms over the adaptive range finder algorithm for solving the fixed-precision problem.

Low-Rank Approximations with Sparse Factors I: Basic Algorithms and Error Analysis

An algorithm for low-rank matrix factorization and its applications

Learning-Based Low-Rank Approximations

Sparse and Low-Rank Matrix Decompositions

Efficient Randomized Algorithms for the Fixed-Precision Low-Rank Matrix Approximation

Accurate Low Rank Approximation at a Low Computational Cost

Low Rank Approximation and Regression in Input Sparsity Time

A general error analysis for randomized low-rank approximation with application to data assimilation

Low-Rank Updates of Matrix Square Roots

Efficient Low Rank Matrix Approximation Via Orthogonality Pursuit and ℓ2 Regularization

Randomized algorithms for low-rank matrix factorizations: sharp performance bounds

Low-Rank Matrix Approximation with Stability.

Low rank matrix recovery using unconstrained smoothed lq minimization

Research Progress of Low-Rank Matrix Approximation and Optimization Problem

Robust Low-Rank Data Matrix Approximations

CUR Low Rank Approximation of a Matrix at Sublinear Cost

Leveraging subspace information for low-rank matrix reconstruction

Nonconvex Optimization Meets Low-Rank Matrix Factorization: An Overview

Low-Rank Modeling and Its Applications in Image Analysis

Single-pass randomized QLP decomposition for low-rank approximation

On Approximation Algorithm for Orthogonal Low-Rank Tensor Approximation