Introduction to Data Privacy
CS/SS 152
Fall 2014




Tentative class schedule

Note that a few lectures will be moved to Fridays; please don't forget to come! Homeworks are due on Thursdays at midnight; an additional office hour will be held by Juba from 3 to 4pm on Thursday afternoon when a homework is due.

September 29th Class overview. Readings and tasks for next class.
October 1st Properties of the DP definition. Sensitivity. The Laplace Mechanism. Readings and tasks for next class. HW1 out.
October 6th Applications of the Laplace Mechanism. Readings and tasks for next class.
October 8th Randomized Response. Exponential Mechanism. Assignment to reading groups.
October 9th HW1 due.
October 13th Exponential mechanism, continued. Sparse vector technique. Readings and tasks for next class.
October 15th Sparse vector, continued. HW2 out.
October 20th Sparse vector, continued. Advanced composition. The Small Database Mechanism. Readings and tasks for next class.
October 22nd SmallDB, continued.
October 23rd HW2 due.
October 27th Blatant non-privacy. Reconstruction attacks.
October 29th Finish reconstruction attacks. Multiplicative Weights. The Multiplicative Weights Mechanisms.
November 3rd Private Peer Prediction.
November 5th Student lecture: Privacy and Learning.Project Proposals due. HW 3 out.
November 7th Friday student lecture: Lower Bounds via Cryptography.
November 10th Value of Private Information. Mechanism design.
November 12th Guest lecture: Moritz Hardt. The noisy power method.
November 14th Friday student lecture: Privacy of the Analyst.
November 17th The local model, SQ learning, PAC learning.
November 19th No class.
November 20th HW3 due.
November 24th Streaming.
November 26th No class. Happy Thanksgiving!
December 1st Private matchings and allocations.
December 3rd Project presentations. Final lecture: sample-and-aggregate, propose-test-release. Final discussion.
December 5th Project reports due.