Enrolment options

This course aims to give an overview of the methods of differential privacy with specific applications in machine learning and artificial intelligence.

The following topics will be covered:

  • What is privacy and how to define it from a mathematical perspective?
  • What is differential privacy, and how to build algorithms that give guarantees of differential privacy?
  • How can an adversary/attacker re-identify or reconstruct data from aggregate queries?
  • What are the variants of differential privacy and how to adapt it to different learning paradigms like distributed learning?
We will then discuss the modern challenges of differential privacy, which is still a very active research field.
 
The courses will be taught in english, with a 2h lecture in the morning and a 2h tutorial/lab in the afternoon (generally 1h of mathematical tutorial and 1h of lab).
For the afternoon labs, bring your computer or find a friend with whom to work during the computer lab.
 
Labs will be in Python on jupyter notebook: be sure that you can open and edit such files before the course! 
 
Some references that might be useful for this class:
Guests cannot access this course. Please log in.