Department of Computer Science
D329, Levine Research Science Building
Duke University, Durham NC 27708
- [Jan 2017] I was invited to moderate a breakout session on building differentially private systems: opportunities and challenges at the NSF Security and Trustworthy Computing (SaTC) PI Meeting.
- [Dec 2016] Yan Chen's paper on differentially private regression diagnostics at ICDM 2016 selected as one of the best papers at the conference (and invited for publication in KAIS journal).
- [Dec 2016] I was invited to speak at the 2016 Federal Committee on Statistical Methodology (FCSM) Policy Conference.
- [Nov 2016] Samuel Haney's paper (in collaboration with US Census Bureau) titled Utility Cost of Formal Privacy for Releasing National Employer-Employee Statistics accepted at SIGMOD 2017
- [Oct 2016] Ios Kotsogiannis's paper on directed edge recommendations accepted at WSDM 2017.
- [Sep 2016] Ios Kotsogiannis's paper on differentially private algorithm selection accepted at SIGMOD 2017.
- [Sep 2016] Ben Stoddard's paper on focussed task lifelogging accepted at MOBICASE 2016
- [Sep 2016] Xi He and Nisarg Raval's demo on privacy utility tradeoffs in releasing GPS trajectory at VLDB 2016 is awarded "Best Demo"!
- [Sep 2016] We are presenting a tutorial, a talk and a demo on differential privacy at VLDB 2017 in New Delhi, India
- [Aug 2016] DPComp.org is now public: explore differentially private algorithms and their performance.
Selected Recent Papers:
- "Utility Cost of Formal Privacy for Releasing National Employer-Employee Statistics", with Samuel Haney, John Abowd, Matthew Graham, Mark Kutzbach, Lars Vihuber accepted at SIGMOD 2017
- "Directed Edge Recommendation Systems", with Ios Kotsogiannis and Elena Zheleva accepted at WSDM 2017
- "Pythia: Data Dependent Differentially Private Algorithm Selection", with Ios Kotsogiannis, Michael Hay and Gerome Miklau accepted at SIGMOD 2017
- "Differentially private regression diagnostics", with Yan Chen, Jerome Reiter and Andres Barrientos accepted at ICDM 2016
- "Ayumu: Efficient lifelogging with focused tasks", with Ben Stoddard, Kate O'Hanlon, Brian Lin and Landon Cox accepted at MOBICASE 2016
- "What You Mark is What Apps See", with Nisarg Raval, Animesh Srivastava, Ali Razeen, Kiron Lebeck, Landon Cox, MOBISYS 2016
- "Principled evaluation of differentially private algorithms using DPBench", with Michael Hay, Gerome Miklau, Dan Zhang, Yan Chen, SIGMOD 2016 [arxiv]
- "Design of Policy-Aware Differentially Private Algorithms" with Sam Haney and Bolin Ding, VLDB 2016 vol 9.
Short Bio: My
research interests lie in privacy in statistical databases, privacy in
augmented reality systems, and systems for massive data analytics. Previously, I was a Senior Research
Scientist at Yahoo! Research. My thesis was awarded the 2009 SIGMOD Jim Gray Dissertation Award, Honorable Mention.
Curriculum Vitae and full bio.
- Kiron Lebeck (BS) currently PhD candidate at U Washington
- Yoon Choi (BS)
- Annie Xie (Senior, NCSSM)
- Bharat Chelapalli (MS) current @ Amazon
Prospective Students: I am primarily interested statistical privacy,
or the problem of disclosing aggregate statistics about data collected
from individuals, while ensuring that individuals level sensitive
properties are not disclosed. This problem appears in a number of
diverse application settings including in the US Census, social networks,
search logs, location
tracking, etc. Read one of the following articles to get an idea of my research: (ACM Crossroads article, no free lunch in privacy, CACM article, Blowfish privacy).
I am also interested in privacy of visual secrets in augmented reality
systems -- an evergrowing threat due to the popularity of camera driven
devices like smart phones and Google Glasses. Read this paper for an overview of our research.