On Fast Computation of Gradients for CANDECOMP/PARAFAC Algorithms

arXiv1204.1586MaRDI QIDQ6232153FDOQ6232153

Anh-Huy Phan, Petr Tichavský, Andrzej Cichocki

Publication date: 6 April 2012

Abstract: Product between mode-$n$ unfolding $$ of an $N$-D tensor $Y$ and Khatri-Rao products of $(N-1)$ factor matrices $$, $m=1,\mathrm{...},n-1,n+1,\mathrm{...},N$ exists in algorithms for CANDECOMP/PARAFAC (CP). If $Y$ is an error tensor of a tensor approximation, this product is the gradient of a cost function with respect to factors, and has the largest workload in most CP algorithms. In this paper, a fast method to compute this product is proposed. Experimental verification shows that the fast CP gradient can accelerate the CP_ALS algorithm 2 times and 8 times faster for factorizations of 3-D and 4-D tensors, and the speed-up ratios can be 20-30 times for higher dimensional tensors.

This page was built for publication: On Fast Computation of Gradients for CANDECOMP/PARAFAC Algorithms

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6232153)