Matteo Maffei
Fachrichtung Informatik
Universität des Saarlandes
Saarbrücken, Germany
Differential privacy, theory and applications
Abstract
Personal information (e.g., patient records, browsing histories, social graphs, and behavioral data used for advertising) is today disseminated in a wealth of databases spread across different institutions and services. On the one hand, disclosing information about these data is often desirable for improving services, analyzing trends, performing marketing studies, conducting research, and so on. On the other hand, this information leakage may irremediably compromise the privacy of users.
Differential privacy has emerged as the de-facto standard notion of privacy for queries on statistical databases. Intuitively, a query is differentially private if it behaves statistically similarly on any pair of databases differing in one entry. In other words, the contribution of each single entry to the query result is bounded by a small constant factor, even if all remaining entries are known. A deterministic query can be made differentially private by perturbing the result with a certain amount of noise, thus reducing the accuracy of the answer.
This course will cover the theoretical foundations of differential privacy, as well as its applications in modern data processing services. In particular, it will overview the fundamental mechanisms to sanitize queries, the cryptographic protocols to achieve differential privacy in a distributed setting, and programming languages techniques to mechanize differential privacy proofs.
Course materials
- M. Maffei. Differential privacy, theory and applications. Slides from the EWSCS 2016 course.
- Videos from the lectures.
- C. Dwork, A. Roth. The algorithmic foundations of differential privacy. Foundations and Trends in Theoretical Computer Science, v. 9, n. 3-4, pp. 211-407, 2014. [doi link]
- F. Eigner, M. Maffei. Differential privacy by typing in security protocols. In Proc. of 26th IEEE Computer Security Foundations Symp., CSF 2013, pp. 272-286. IEEE CS Press, 2013. [doi link]
- F. Eigner, A. Kate, M. Maffei, F. Pampaloni, I. Pryvalov. Differentially private data aggregation with optimal utility. In Proc. of 30th Annual Computer Security Applications Conf., ACSAC 2014, pp. 316-326. ACM Press, 2014. [doi link]
Last changed
May 7, 2016 12:53 Europe/Helsinki (GMT +03:00)
by
local organizers, ewscs16(at)cs.ioc.ee
EWSCS'16 page:
http://cs.ioc.ee/ewscs/2016/