Sommersemester 2024
- Probability Distances for Data Science:
- Important note: The lecture on 15th of May is cancelled. Instead, there will be a video uploaded on ILIAS.
- Lecture each Wednesday, 14:15-15:45
- Room N08 (Morgenstelle)
- Exercise classes on Tuesdays, 16:15-17:45 in C2A17 (Morgenstelle)
- All further information on ILIAS
- Description: The objective of this course are different concepts of distances between probability measures aimed at applications in data science. The classes of distances which are studied include optimal transport distances, f-divergences and integral probability metrics. The focus is on fundamental mathematical properties of these distances, like duality, famous inequalities, geometric aspects, and quantization. Several applications in the area of data science and machine learning are illustrated throughout, for instance related to clustering, autoencoders, GANs, image processing, and compression.
- Problem sets (password protected, in case you cannot access Ilias): PS1, PS2, PS2_sol, PS3
- Lectures notes (password protected, in case you cannot access Ilias): Script (12.04.24)
