In this talk, I will begin by introducing a hypothesis-testing-based formulation of differential privacy in classical computation. The Gaussian Differential Privacy paper (Dong–Roth–Su '22) established a central limit theorem for the composition of multiple private mechanisms. Building on this, I will present a Poisson extension of their result and show how both the Gaussian and Poisson limits are unified under the framework of infinitely divisible privacy, revealing connections to statistics, probability, and discrete mathematics. I will conclude with discussions on quantizing privacy in quantum computation and deriving central limit theorems in a framework of quantum differential privacy.

 

Short Bio:  Aaradhya is a fifth-year PhD Student at Princeton ORFE. He completed his Bachelor's in mathematics at IISc in 2021. His research interests are at the interface of probability theory, statistics, and information theory with applications in (quantum) differential privacy, machine unlearning, and spin glasses. When most approaches in a field are analytic, he tries to develop algebraic ones — and vice versa.