Are you an EPFL student looking for a semester project?
Work with us on data science and visualisation projects, and deploy your project as an app on top of Graph Search.
This lecture covers the concept of differential privacy, focusing on partial secrecy and hypothesis testing. It explains the conditions for a mechanism to be (ε,δ)-differentially private and how to test hypotheses for differential privacy. The lecture also delves into the composition of differentially private mechanisms, discussing homogeneous and heterogeneous compositions. The speaker explores the privacy region for differential privacy and the optimal privacy region under composition. The lecture concludes by addressing the question of how privacy degrades after repeated accesses and the tightest bound on privacy under composition.