DSC291 (Fall 2024) Introduction to Differential Privacy: Theory, Algorithms and Applications |
|
Piazza: https://piazza.com/ucsd/fall2024/dsc291
Piazza is our main channel of communication. Questions should be posted here.
Gradescope: https://www.gradescope.com/courses/879296
This is where you submit your homeworks and project reports.
Office hours:
Instructor: After the lecture on Wednesday.
TA office hour to be announced soon.
Evaluation:
30% Homework, 40% Project, 20% Midterm Quiz, 10% for attendance / Participation.
Textbook:
Week | Dates | Lecture Materials | Readings | Assignments |
---|---|---|---|---|
1 | 27-Sep-24 | No lecture. Please read this notebook on Privacy challenges in the AI Era and DR Ch 1. | DR Ch 1 | |
2 | 30-Sep-24, 2-Oct-24, 4-Oct-24 | Overview. Privacy threats. Defining Privacy, Sensitivity and Laplace Mechanism [Lecture 1,Annotated] [Lecture 2,Annotated] | Notebook, DR Ch 2, Ch 3.1, Ch 3.2 | HW1 out on Oct 2 |
3 | 7-Oct-24, 9-Oct-24, 11-Oct-24 | Sparse Vector and Private Query Release [Lecture 3, Annotated]] | DR Ch 3.6, DR Ch 4.2 | Project proposal due Oct 11 |
4 | 14-Oct-24, 16-Oct-24, 18-Oct-24 | Private Multiplicative Weights, Exponential Mechanism, SmallDB [Lecture 4, Annotated] | Full proof of PMW.DR Ch 3.3, DR Ch 3.4 | HW1 due on Oct 16 HW2 out on Oct 18 |
5 | 21-Oct-24, 23-Oct-24, 25-Oct-24 | Privacy loss RV, Composition [Lecture 5, Annotated] | DR Ch 3.5 | |
6 | 28-Oct-24, 30-Oct-24, 1-Nov-24 | Gaussian Mechanism [Lecture 6, Annotated], midterm discussion | Balle & Wang (2018) | Midterm Quiz on 30-Oct-24 HW2 due on Nov 1 |
7 | 4-Nov-24, 6-Nov-24, 8-Nov-24 | Renyi DP [Lecture 7, Annotated], Privacy accounting and autodp [Lecture 8, Annotated], | Bun & Steinke (2018) | Project midterm report due Nov 8 |
8 | 13-Nov-24, 15-Nov-24 | Intro to DPML [Lecture 9, Annotated] | Chaudhuri et al., Kifer et al., Bassily et al. | HW3 out on Nov 12. No Lecture on Monday (Public Holiday) |
9 | 18-Nov-24, 20-Nov-24, 22-Nov-24 | DPML algorithms [Lecture 10, Annotated], DPML algorithms (Part 2) [Lecture 11, Annotated] | ||
10 | 25-Nov-24, 27-Nov-24 | DPML algorithms (Part 3)[Slides, Annotated], Private Deep Learning [Slides, Annotated] | Practicum [Start-up kit] HW3 due on Nov 27. No lecture on Friday (Thanksgiving) | |
11 | 2-Dec-24, 4-Dec-24, 6-Dec-24 | Mon: Data Adaptive DP [Slides, Annotated]. Wed: Project consultation in Lecture. Friday 1 - 5pm Project presentations | Vadhan Ch 3.1, Nissim et al. (2011) | Final project report due |
Code of Conduct: You can find UCSD's student code of conduct here.