Development of Online Learning Materials for Tensor Data Processing Exercises

Abstract

—Tensor decomposition is used in a wide range of research fields; however, its theory is difficult to understand. Therefore, basic education is essential when using it in programming. Currently, there are few Japanese universities that provide education on tensor decomposition; however, some overseas universities have already conducted it, and online learning materials are also substantial. Therefore, in this paper, we have developed online learning materials for basics and programming exercises of higher-order singular value decomposition (HOSVD), which is one of tensor decomposition, for the purpose of increasing the learning materials for tensor decomposition education. Our learning material is created on Microsoft Teams, and students can access this material channel and work on exercises on demand while watching explanatory videos including CG animation. As a result of the trial of this learning material, it was found that the students who used it can generally understand the processes related to tensor decomposition and can perform basic programming of them.