Fast tensor decompositions for big data processing

Abstract

Tensors, as a natural extension of matrices, and their decompositions provide important tools in many disciplines such as psychometrics, signal processing, data communication, computer vision, and machine learning. The main objective of this paper is to briefly review several recent state-of-the-art approaches for large-scale tensor data which is a crucial part of big data. Moreover, we also introduce our own contributions on this topic.