Organizer(s):

Cynthia Dwork, Harvard University

Ruobin Gong, Rutgers University

Weijie Su, University of Pennsylvania

Linjun Zhang, Rutgers University

Description

The workshop brings together research experts from many domains of social sciences, demography, public policy, statistics, and computer science to address key challenges in the use of the differentially private Census noisy measurement files to support social research and policy decisions.

With the implementation of differential privacy, Census data consumers need implementable tools to carry out analyses in the presence of noise introduced to protect privacy. Additionally, privacy researchers in statistics, computer science, and the quantitative social sciences need to understand the practical hurdles faced by data users. The following questions are at the center of the workshop’s inquiry:

• What are some important use cases in the many disciplines of social sciences in which the analysis of noisy measurement files can support reliable research conclusions and trustworthy policy decisions.
• What are the statistical and computational methodologies needed to support the analyses of the noisy measurement files? What are some challenges to their design and practical implementation? In particular, how to perform appropriate uncertainty quantification for analyses based on noisy measurement files?
• The Census Bureau may not be able to publicly release the full set of noisy measurement files at a sufficiently detailed geographic level. What are some inference methodologies that can support the analysis of minimally post-processed data files, or a combination of noisy measurements and post-processed data sources?
• Looking ahead, what are some good design principles for privacy-protected data files for public release, such that they are conducive to principled analysis?

The workshop aims to stimulate an in-depth exchange among a diverse group of research communities, and to supply a platform for researchers and data users to formulate privacy research questions, facilitate the development of useful tools for analyzing differentially private data products, and to provide guidance on the design of analysis-aware privacy protocols for official statistical agencies.

The workshop begins with a reception on April 27 and continues with its technical program on the following two days.

Workshop schedule: [PDF]

Workshop videos: [Youtube playlist]

Confirmed Participants

Session Directors:

• Gary King (Harvard University)
• Jae June Lee (Georgetown University)
• Amy O’Hara (Georgetown University)
• Adam Smith (Boston University)

Confirmed Speakers:

• Jordan Awan (Purdue University)
• Clark Bensen (Polidata)
• Claire Bowen (Urban Institute)
• Mark Bun (Boston University)
• Raj Chetty (Harvard University)
• Ryan Cumings-Menon (U.S. Census Bureau)
• Jeff Gill (American University)
• Eric Guthrie (Minnesota State Demographic Center)
• Michael Hawes (U.S. Census Bureau)
• Sam Hirsch (Jenner & Block)
• Frauke Kreuter (University of Maryland; Ludwig-Maximilian-University)
• Jeff Lewis (UCLA)
• Karin Mac Donald (Statewide Database, UC Berkeley)
• Ashwin Machanavajjhala (Duke University; Tumult Labs)
• Fred McBride (Southern Poverty Law Center)
• Jeanine Abrams McLean (Fair Count, Inc.)
• Audra McMillan (Apple)
• Terry Ao Minnis (Asian Americans Advancing Justice)
• Bill O’Hare (Partnership for America’s Children)
• Nancy Potok (George Washington University)
• Jayshree Sarathy (Harvard University)
• Ian Schmutte (University of Georgia)
• Jeremy Seeman (Penn State University)
• Dan Sheldon (UMASS Amherst)
• Sesa Slavkovic (Penn State University)
• Jan Vink (Federal-State Cooperative Program for Population Estimates)
• Rebecca Wright (Barnard College)

Code of Conduct

Program

Wednesday, April 27, 2022

Thursday, April 28, 2022

Friday, April 29, 2022

Call For Participation

Presentations at this workshop are by invitation, as is in-person attendance. The contents of the workshop will be available to all who are interested in two ways:

1. The workshop will be delivered via zoom for synchronous remote participation. To attend, please register using the link at the bottom of the page. A day or two before the start of the event, you will receive an email with information on how to join. Please note that you must register BEFORE the start of the event, as we will not be processing late registrations.
2. Presentations at the workshop will be videotaped and linked to this webpage when permitted by the presenters.

This event is held with the support of grant number G-2022-19314 from the Alfred P. Sloan Foundation.

Sponsored by Rutgers University Department of Statistics and DIMACS.

Registration for this event is closed.