Privacy-Preserving Sparse Data Modeling in Encrypted Domain

Abstract

With the arrival of the big data era, the amount of digital content on the network has rapidly increased. The use of edge/cloud computing has become widespread in various fields such as big data analysis. Examples include data compression to reduce the amount of data, data mining to extract significant information from data, and machine learning to automatically learn models for the classification and prediction of data. However, the use of edge/cloud computing is premised on the reliability of service providers, and there is a concern that privacy invasion problems may occur as a result of the lack of reliability and the unauthorized use or loss of data owing to accidents. To solve the problems, privacy-preserving sparse modeling based on the random unitary transform has been proposed. The features of privacy-preserving sparse modeling are that 1) it can process encrypted data without changing the algorithm of sparse modeling and 2) it guarantees no degradation of sparse modeling performance, without a significant increase in the amount of calculation. In this article, we overview the basic mechanism of the privacy-preserving sparse modeling and introduce application examples to edge AI.