Tensorlab+: A Case Study on Reproducibility in Tensor Research
Stijn Hendrikx, Raphaël Widdershoven, Nico Vervliet, Lieven De Lathauwer · Computing in Science & Engineering · 2023
Tensor methods emerge as an important class of basic techniques, generalizing matrix methods to multi-way data and models. We have recently released Tensorlab+, which is a downloadable archive of code and data that allows peers to reproduce the experiments reported in our publications on tensor decompositions and applications. We briefly discuss the basic tensor tools and give an introduction to the contents of Tensorlab+. We elaborate on the steps that were taken to ensure the reproducibility of the experiments and the quality of the code.