Schedule
Lecture Schedule (Tentative)
Week | Lecture | Topic | Readings | Slides Notes |
---|---|---|---|---|
Week 1 | Lecture 1 | Course Overview | Slides | |
Definitions and Basic Techniques | ||||
Week 1 | Lecture 2 | Reconstruction Attacks (Part 1) | Reading | Slides Notes |
Week 2 | Lecture 3 | Reconstruction Attacks (Part 2) | Reading | Slides Notes |
Week 2 | Lecture 4 | Definitoin of Differential Privacy Randomized Response; Laplace Mechanism |
Reading/Video | Slides Notes |
Week 3 | Lecture 5 | Properties of Differential Privacy Composition; Post-Processing; Group Privacy |
Slides Notes |
|
Week 3 | Lecture 6 | Selection problem (part 1) Exponential Mech. |
Slides Notes |
|
Week 4 | Lecture 7 | Selection problem (part 2) Report noisy max |
Slides | |
Week 4 | Lecture 8 | DP and Mechanism Design | Slides | |
Private Synthetic Data | ||||
Week 5 | Lecture 9 | Private GAN (Online) Private Multiplicative Weights |
||
Private (Non)-Convex Optimization | ||||
Week 5 | Lecture 10 | (Strong) Convexity, smoothness Output/Objective Perturbation |
||
Week 6 | Lecture 11 | Private Gradient Descent (Part 1) | ||
Week 6 | Lecture 12 | Private Gradient Descent (Part 2) | ||
Week 7 | Lecture 13 | Private Deep Learning (Part 1) | ||
Week 7 | Lecture 14 | Private Deep Learning (Part 2) | ||
Practical Deployments of DP | ||||
Week 8 | Lecture 15 | Local Differential Privacy | ||
Week 8 | Lecture 16 | Shuffling | ||
Week 9 | Lecture 17 | US Census Deployment 2020 | ||
Week 9 | Lecture 18 | Buffer | ||
Connections and Applications | ||||
Week 10 | Lecture 19 | Online Learning Follow-the-perturbed-leader via DP |
||
Week 10 | Lecture 20 | Adaptive Data Analysis Algorithmic Stability |
Acknowledgement: Some of course materials are based on those developed by Gautam Kamath, Jonathan Ullman, and Adam Smith.
Deadlines
TBA