In fall 2024, Advanced Topics in Machine Learning (ATML) will be taught by Amartya Sanyal, Nirupam Gupta, and Rasmus Pagh. As machine learning models are often trained on data containing sensitive private information, it is important to provide reasonable privacy protection to the individuals whose data is included. This course focuses on advanced topics in privacy-preserving machine learning, with a focus on techniques that maintain data privacy without compromising the performance of the learning algorithm. Key concepts covered include differential privacy and federated learning. By the end, students will be prepared to pursue a master's thesis on privacy in machine learning.
The course is relevant for computer science students as well as students from other studies with a good mathematical background, including Statistics, Actuarial Mathematics, Mathematics-Economics, Physics, etc. We assume that the students have previously passed Machine Learning A+B courses offered by DIKU.
The course is theoretically oriented and assumes that you have a strong mathematical background.Â
Important: We assume that the students have previously passed Machine Learning A+B courses offered by DIKU.
Recommended Academic Background: If you have a strong mathematical background and basic knowledge of machine learning it is possible to take ATML without taking ML-B. In this case, please, check taking PriMAL before/without ML and self-preparation instructions carefully. You should be able to solve the self-preparation assignment before joining the course, otherwise we strongly advise taking ML first. Be aware that our ML course is more theoretical than average machine learning courses and if you have taken a machine learning course elsewhere it does not necessarily prepare you for ATML. In particular, Introduction to Data Science (IDS) course given by DIKU is much lighter than ML and does not prepare you for ATML.
The course will be held in Block 1 Schedule C. Additional details regarding timing of lectures and tutorials are yet to be finalised.