Somewhat Homomorphic Encryption and its Attractiveness in Privacy-Enhancing Technologies
Andreas Peter, University of Twente
November 8, 2017 10:30am, in DC 1304
In 2009, Gentry presented the first Fully Homomorphic Encryption (FHE) scheme, which roughly speaking supports arbitrary computation over encrypted data without the need to decrypt. While, theoretically, FHE schemes can be used in many settings to provide data privacy, in practice they suffer from severe efficiency drawbacks. Somewhat Homomorphic Encryption (SWHE), on the other hand, constitutes a relaxed form of FHE which only supports the computation of a limited set of functions on encrypted data (i.e., functions that can be represented as an arithmetic circuit with fixed and restricted amounts of additions and multiplications).
In this talk, I would like to give a short introduction to SWHE and show that the limited functionality of SWHE schemes can be sufficient to provide data privacy in specific application scenarios and, most importantly, can lead to constructions efficient enough for practical use. Concretely, I will look at two example applications from different domains: (1) enhancing privacy in recommender systems based on social networks and (2) privately outsourcing forensic image recognition. Using the different characteristics of these settings, I will identify some features of SWHE which make the further study of this type of encryption within the area of Privacy-Enhancing Technologies particularly worthwhile.
Andreas Peter graduated with a M.Sc. in mathematics at both the University of Cambridge (UK) and the University of Oldenburg (Germany) in 2008 and 2009, respectively. Subsequently, he received the Ph.D. in computer science from the Technical University of Darmstadt (Germany) in 2013. His Ph.D. thesis deals with the topic of secure outsourcing of computation with a special focus on homomorphic encryption. Since 2014, he is employed as an Assistant Professor at the chair of Services, Cybersecurity and Safety at the University of Twente (NL). His research interests include both fundamental and applied security and privacy aspects in IT systems with a focus on privacy-enhancing technologies, cryptographic protocol design and analysis, as well as the application and theory of functional and homomorphic encryption. He served on the program committees of several workshops and conferences devoted to information security and privacy. Since 2015, he serves on the Editorial Board of the MDPI Open Access Journal on Cryptography and the SpringerOpen EURASIP Journal on Information Security.