With organizations like Facebook currently proposing that employees and external third parties analyze sensitive user data, a need exists for the analysis of sensitive data by systems with strong privacy guarantees.
Machanavajjhala and Kotsogiannis will present compelling research on just this topic in their paper "Architecting a Differentially Private SQL Engine." At the biennial Conference on Innovative Data Systems Research on Jan. 14, 2019, they will unveil their revolutionary system to build databases that manage sensitive data about individuals, like Census Data, medical records, and etc. The proposed new system allows answering complex aggregate queries with strong privacy guarantees of differential privacy.
This paper and its groundbreaking solution are components of the larger DARPA Brandeis-funded project, System-P. Its goal is to develop easy to use systems for managing private data with provable privacy guarantees.
Duke computer science professor and privacy expert, Ashwin Machanavajjhala and Duke graduate students Ios Kotsogiannis and Yuchao Tao are leading this project, with collaborators Gerome Miklau from the University of Massachusetts at Amherst and Michael Hay of Colgate University.
