Fast and stable randomized low-rank matrix approximation

Abstract: Randomized SVD has become an extremely successful approach for efficiently computing a low-rank approximation of matrices. In particular the paper by Halko, Martinsson, and Tropp (SIREV 2011) contains extensive analysis, and has made it a very popular method. The typical complexity for a rank-$r$ approximation of $m\times n$ matrices is $O(mn\log n+(m+n)r<sup>2)$ for dense matrices. The classical Nystr{\"o}m method is much faster, but applicable only to positive semidefinite matrices. This work studies a generalization of Nystr{\"o}m method applicable to general matrices, and shows that (i) it has near-optimal approximation quality comparable to competing methods, (ii) the computational cost is the near-optimal $O(mn\log n+r<sup>3)$ for dense matrices, with small hidden constants, and (iii) crucially, it can be implemented in a numerically stable fashion despite the presence of an ill-conditioned pseudoinverse. Numerical experiments illustrate that generalized Nystr{\"o}m can significantly outperform state-of-the-art methods, especially when $r\gg 1$, achieving up to a 10-fold speedup. The method is also well suited to updating and downdating the matrix.