Practitioner’s Perspectives on a Differential Privacy Deployment Registry

• Date: Monday, November 10, 2025
• Time: 2-3pm CT
• Where: Morgridge Hall 5621 OR Zoom

Abstract:

Differential privacy (DP) is a principled approach to producing statistical data products, like summary statistics and machine learning models, with strong, mathematically provable privacy guarantees to people in the underlying dataset. In the past decade, DP has been increasingly adopted across government (e.g., the U.S. Census Bureau) and industry (e.g., Apple, Google, Microsoft). When applying DP, practitioners must make several implementation decisions with critical impacts to data privacy and/or utility. However, best practices for these choices largely do not exist. To promote shared learning and accountability, Dwork, Kohli, and Mulligan (2019) proposed a public-facing repository (“registry”) of DP deployments. In this talk, I will present our contributions to a community-wide effort to realize the vision of a differential privacy deployment registry. We (1) develop a rich, hierarchical schema to describe any given DP deployment and (2) design and implement a registry prototype as an interactive interface where practitioners can access information about past DP deployments. We (3) populate our interface with 21 real-world deployments and (4) conduct an exploratory user study with 16 DP practitioners to foresee challenges and opportunities around the registry’s adoption. In addition to describing these contributions, I will also discuss our efforts to develop a live version of the registry in collaboration with OpenDP, the open-source software project and Oblivious, a privacy start-up. Finally, I will share details about a recent proposal by the U.S. National Institute of Standards and Technology (NIST) to host the registry going forward.

Based on joint work with Elena Ghazi and Salil Vadhan

Preprint: https://arxiv.org/abs/2509.13509

Biography:

Priyanka Nanayakkara is a postdoctoral fellow in computer science at Harvard University, co-hosted by Professors Salil Vadhan and Martin Wattenberg. During the 2024-25 academic year, she was a fellow at the Center for Research on Computation and Society (CRCS) at Harvard. Her research develops interactive, visual tools to align technical approaches for data privacy with how people naturally reason about data. She holds a PhD in computer science and communication from Northwestern University, advised by Professor Jessica Hullman. During her PhD, she was also a visiting researcher at Columbia University, a visiting graduate student at UC Berkeley’s Simons Institute, and an intern at Microsoft Research. She is an MIT EECS Rising Star and UCSD Data Science Rising Star.