Towards Secured and Transparent Artificial Intelligence Technologies in Hierarchical Computing Networks

摘要

Researchers at NTT Network Innovation Laboratories have recently been focusing on the interdiscipline of transparent artificial intelligence (AI) technologies and hierarchical computing networks. A hierarchically distributed computing structure not only improves the quality of computation but also creates an extra degree of diversity for algorithm refinement. Sparse coding, an important transparent AI technique, is finding application in this new domain. We propose in this article a secure sparse coding scheme that enables computing directly on cipher-texts. We also demonstrate its application to image compression and face recognition in edge and cloud networks.