Rafael D’Oliveira — Clemson University

Abstract: Since its inception in 2006, differential privacy has emerged as one of the main privacy-preserving tools for sharing information from datasets that contain sensitive information about individuals. Notable applications of differential privacy include large-scale private learning of users’ preferences by companies like Google, Microsoft, and Apple, and the 2020 United States Census’ privatization method to provide data privacy protection, each impacting hundreds of millions of individuals.

In this talk, we present a general framework for designing optimal differentially private mechanisms for binary functions via a graph representation of datasets. The topic will be introduced in a self-contained manner, requiring no prior knowledge on differential privacy.