Hacker News new | past | comments | ask | show | jobs | submit login
Is Differential Privacy practical? (2013) (mrtz.org)
29 points by luu on March 11, 2015 | hide | past | favorite | 4 comments

The Ironclad project[1] claims to have created a formally-verified differential privacy database. I recommend watching the demo video, the development efficiency they saw using formal verification is jaw-dropping.

[1] https://www.usenix.org/conference/osdi14/technical-sessions/...

Google recently put a tool called RAPPOR into production that's partially based on differentially private randomized response for collecting usage statistics about Chrome users.

Code is here: https://github.com/google/rappor Paper is here: http://static.googleusercontent.com/media/research.google.co...

I certainly hope Betteridge's law of headlines does not apply here. Differential privacy is an amazing criterion, and I'd love to hear about anyone who has put it into practice on data they're releasing.

It would be much appreciated if someone could explain, succinctly, how differential privacy works and its strengths and weaknesses. Thanks!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact