Differential Privacy: From Theory to Practice
February 12-16, 2017
The concept of differential privacy is central to the rigorous foundational approach to private data analysis that has emerged in cryptography in the last decade. The development of this approach was motivated by the vast amounts of personal information that are collected in today’s information environment, and by a rapidly growing body of work demonstrating how traditional approaches to privacy, such as de-identification, fail to provide adequate privacy preservation. The framework of differential privacy provides a rigorous mathematical treatment of privacy, with concrete provable guarantees that are robust against adversaries with arbitrary computational power and with arbitrary auxiliary knowledge. There is now a large body of theoretical work in this vein, and many established relationships to scientific fields including statistics, machine learning, databases, algorithms, information theory, program verification, and game theory. The products of this research are also making their first strides into use in real world applications where sensitive personal information is analyzed, with algorithms currently deployed by the US Census Bureau, Google, Yahoo, and Apple.
The target audience for the school is graduate students and postdocs in cryptography (we will assume that participants have taken at least one university-level course in cryptography). However, all faculty, undergrads and professionals with the necessary background are welcome. The winter school is open to participants from all over the world; all talks will be in English.
- Katrina Ligett, Hebrew University and Caltech
- Kobbi Nissim, Georgetown University and Harvard
- Vitaly Shmatikov, Cornell Tech
- Adam Smith, Pennsylvania State University
- Jon Ullman, Northeastern University
Where: The winter school will take place at the Rayman hall at Kfar Hamaccabiah events & conference center in Ramat Gan
When: Sunday, February 12, 2017 to Thursday February 16, 2017
Registration: Due to rising costs, registration is 750 shekels for the entire school. Registration is free for overseas participants (due to costs already incurred due to travel). Israeli participants who have difficulty paying the registration fee can request a waiver; please include this in the special request box in the registration form. After registration has been confirmed, Israeli participants will receive a link to carry out the payment. Registration will be considered complete only after payment.
Registration includes school participation, lunch, refreshments and the excursion (accommodation is not included). Please register by December 31, 2016.
Contact: For any questions or queries, please send an e-mail to: email@example.com
Hotel: We have arranged a special rate at the where the conference center is located. The rate is $160 a night for a single room, $180 a night for a double room (with two occupants) and $210 for a triple room (with three occupants). The rate includes breakfast. Hotel reservation form will be available soon.
Support: A limited number of stipends of $800 each (for flight and accommodation) will be awarded for overseas students needing support. The deadline for stipend application is December 15, 2016. Please have your advisor send a letter justifying the need for financial support.
Sponsorship: This winter school is graciously sponsored by the European Research Council under the European Union’s Seventh Framework Programme (FP/2007-2013) / ERC Grant Agreement n. 615172 (HIPS), and the BIU Center for Research in Applied Cryptography and Cyber Security in conjunction with the Israel National Cyber Bureau in the Prime Minister’s Office, Bar-Ilan University.
The detailed schedule for the winter school can be downloaded here.
Sunday, February 12, 2017
- Vitaly Shmatikov: The Anonymization/De-identification Paradigm (slides)
- Vitaly Shmatikov: Re-identification Attacks
- Vitaly Shmatikov: Inference Attacks (slides)
- Katrina Ligett: Introduction to Differential Privacy, Randomized Response, Basic Properties (slides)
- Katrina Ligett: Basic Tools – Laplace Mechanism, Exponential Mechanism (slides)
- Katrina Ligett, Kobbi Nissim and Adam Smith: Further Examples/Problem Solving (PDF)
Monday, February 13, 2017
- Katrina Ligett: Private Data Release (slides)
- Jon Ullman: Lowerbound Techniques: Packing Argument (slides)
- Jon Ullman: Private Data Release – Intractability
- Kobbi Nissim: Advanced Tools: Tree Algorithm, Advanced Composition (slides)
- Katrina Ligett: Advanced Tools – Sparse Vector (slides)
- Jon Ullman: Private Multiplicative Weights (slides)
Tuesday, February 14, 2017
- Adam Smith: Advanced Tools – Stability Based Techniques, Choosing/Gap Mechanism (slides)
- Jon Ullman: Lowerbound Techniques: Fingerprinting Codes (slides)
- Adam Smith: Local Algorithms (slides)
- Katrina Ligett, Kobbi Nissim and Adam Smith: Further Examples/Problem Solving
Wednesday, February 15, 2017
- Vitaly Shmatikov: Privacy – What It Is and What It Isn’t? (slides)
- Vitaly Shmatikov: Privacy Meets Machine Learning
- Kobbi Nissim: Generalization and Privacy (slides)
- Kobbi Nissim: Private Learning 1 (slides)
- Kobbi Nissim: Private Learning 2 (slides)
- Adam Smith: Private ERM (slides)
Thursday, February 16, 2017
- Adam Smith: Privacy and Statistics
- Jon Ullman: Lowerbound Techniques: Reconstruction Attacks (slides)
- Jon Ullman: Tracing Attacks
- Adam Smith: Analyzing Graphs (slides)
- Katrina Ligett: Privacy and Game Theory 1 (slides)
- Katrina Ligett: Privacy and Game Theory 2