New algorithms for heavy hitters in data streams
David P. Woodruff
ICDT 2016
The CUR decomposition of an m × n matrix A finds an m × c matrix C with a subset of c < n columns of A, together with an r × n matrix R with a subset of r < m rows of A, as well as a c × r low-rank matrix U such that the matrix CUR approximates the matrix A, that is, ∥A-CUR∥2 F ≤ (1 + ϵ) ∥A-Ak∥2 F, where ∥. ∥F denotes the Frobenius norm and Ak is the best m × n matrix of rank k constructed via the SVD. We present input-sparsity-time and deterministic algorithms for constructing such a CUR decomposition where c = O(k/ϵ) and r = O(k/ϵ) and rank(U) = k. Up to constant factors, our algorithms are simultaneously optimal in the values c, r, and rank(U).
David P. Woodruff
ICDT 2016
Alexandr Andoni, Jiecao Chen, et al.
ITCS 2016
Kenneth L. Clarkson, David P. Woodruff
SODA 2015
Christos Boutsidis, David P. Woodruff
STOC 2014