Ashwin Machanavajjhala, Duke computer science professor and privacy expert, spoke on “Public Use Data Products and Privacy Protection: Experiences from the Front Lines” at the AAAS Annual Meeting. Currently assisting the 2020 Census team to modernize their data publications using differential privacy (DP), Professor Machanavajjhala’s recent work includes “Utility-cost of Provable Privacy: A Case Study Using United States Census Data.”
Dr. John Abowd, the U.S. Census Bureau’s Chief Scientist and Associate Director for Research and Methodology also shared comments at this very popular AAAS session, along with other panelists like Duke statistical science professor, Jerome Reiter and previous Census Bureau director, Bob Groves.
During the panel briefing, Machanavajjhala discussed:
- How the 2020 Census will benefit from using DP: “The new system involves complex mathematical algorithms that inject ‘noise’ into the data, making it harder to get accurate information and providing ‘a very strong guarantee’ of privacy” as Seth Borenstein quoted Ashwin in his AP article on this session, Potential privacy lapse found in Americans’ 2010 census data.
- Why differential privacy is needed in other domains: for example, smart devices can now track nearby persons’ movements by using individual data collected.
- The immediate and overarching need to consider how data is analyzed and shared.
The annual meeting of the American Association for the Advancement of Science (AAAS) took place Feb. 14-17, 2019 in Washington, D.C.