ToTs & TiPs

Privacy Preservation Methods for Genetic and Location Data Massive data sets and sophisticated learning algorithms are routinely used by companies and governments to predict people’s behavior and to deduce key properties of individuals. While the ability to infer detailed knowledge about people can be tied to many privacy threats, existing models of statistical disclosure are limited in modeling the threat and in suggesting effective solutions. In many cases, the data about each individual is processed separately, ruling out solutions such as anonymization and attribute obfuscation. Therefore, we generalize the concept of inferential privacy, and suggest Decision Theory as a solid foundation that can model inferential attacks. In addition, we use Principle Component Analysis (PCA) to quantify the potential privacy loss through the ability of an attacker to infer an individual’s properties. We show initial results from two studies, one in the field of location databases and one in the field of genomic databases, in which we empirically describe an upper bound to privacy loss. Finally, the talk will discuss several privacy- enhancing technologies that rely on inferential privacy, and demonstrate how they can be applied to real-world problems. Speakers: Prof. Eran Toch is a member of the faculty of Engineering at Tel-Aviv University, at the department of Industrial Engineering. Eran’s research group is working on usable privacy and security, currently running several projects funded by agencies such as the Israeli Science Foundation (ISF), Horizon 2020, DARPA, and Israel Ministry of Science. Prior to joining Tel Aviv University, Eran was a post-doc fellow at the Carnegie Mellon University, School of Computer Science, and he has a Ph.D. from the Technion – Israel Institute of Technology.