Priyanka Nanayakkara
Harvard University
priyankan@g.harvard.edu
Bio
Priyanka Nanayakkara is a postdoctoral fellow at the Center for Research on Computation and Society (CRCS) at Harvard University, where she is hosted by Professor Salil Vadhan. Her research at the intersection of privacy and visualization develops human-centered tools to make differential privacy usable for people across the data ecosystem. In general, she is interested in problems where using a technology requires grappling with hard social questions. Priyanka’s work has appeared in venues such as IEEE S&P, USENIX Security, PoPETS, and AIES. She received a joint 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.
Areas of Research
- Human-Computer Interaction
Making Differential Privacy Usable Through Human-Centered Tools
It is often useful to learn patterns about a population while protecting individuals’ privacy. Differential privacy (DP) is a state-of-the-art framework for limiting how much is revealed about individuals during analysis. Under DP, statistical noise is injected into analyses to obscure individual contributions while maintaining overall patterns. The amount of noise is calibrated by a unit-less privacy loss parameter, epsilon, which controls a tradeoff between strength of privacy protections and accuracy of estimates. This tradeoff is difficult to reason about because it is probabilistic, non-linear, and inherently value-laden. However, for DP to be broadly usable, people across the data ecosystem must be able to effectively reason about it. In my work, I develop human-centered tools for data curators, data analysts, and data subjects to reason about DP. Specifically, I have developed (1) an interactive visualization interface for data curators setting epsilon, (2) an interactive paradigm instantiated in a visualization interface for analysts to spend epsilon efficiently during exploratory analysis, and (3) explanations of epsilon’s privacy guarantees for data subjects. Furthermore, I have conducted (4) an analysis of debates around the U.S. Census Bureau’s use of DP for the 2020 census to propose communication strategies that can facilitate more productive discussions and ensure smoother deployments. Together, these works aim to improve DP’s usability and further privacy-preserving data science.