Differential Privacy: Future Work & Open Challenges

View As Web Page Cybersecurity Insights a NIST Blog Differential Privacy: Future Work & Open Challenges Setting the Privacy Parameter The impact of the privacy parameter (or privacy budget) ε has been a consistent theme throughout this series. Conceptually, the privacy parameter is simple: smaller values of ε yield better privacy, and larger values yield worse privacy. But there's one very important question we haven't answered: what, exactly, does ε mean, and how should we set it? Unfortunately, we still don't have a consensus answer to this question.  Researchers have recommended ε values around 1. Can we set ε higher than 1 without opening ourselves up to privacy attacks? Existing deployments of differential privacy indeed have set ε larger than 1: Apple's differential privacy system uses ε values between 2 and 16 (per user/day) The US Census Bureau plans to use an ε of 19.61 for the 2020 Census redistricting data Google's Community Mobility Reports have used an ε value of 2.64 (per user/day) Despite concerns expressed by the academic community about the large ε values used in these systems, none of them has yet been the target of a successful privacy attack (in contrast to systems based on de-identification, which have been broken repeatedly). Read More

Connect with us Received this email from a friend? Subscribe here. You are subscribed to receive updates to NIST Cyber Blog from the National Institute of Standards and Technology (NIST). Subscriber services: Manage Preferences  |  Unsubscribe  |  Help

---

If you have questions or problems with the subscription service, please contact subscriberhelp.govdelivery.com. Technical questions? Contact inquiries@nist.gov. (301) 975-NIST (6478). This service is provided to you at no charge by National Institute of Standards and Technology (NIST). 100 Bureau Drive, Stop 1070 · Gaithersburg, MD 20899 · 301-975-6478