Probabilistic Machine Learning (Summer 2020)

Open Educational Resources

In the "Corona Summer" of 2020, Prof. Dr. Philipp Hennig remotely taught the course on Probabilistic Machine Learning within the Tübingen International Master Programme on Machine Learning. The course consists of two ~90min lectures per week (26 lectures in total) plus a weekly practical / tutorial. Videos of all lectures are available on the youtube channel of the Tübingen Machine Learning Groups. The tutorials were taught by members of the Chair: Alexandra Gessner, Julia Grosse, Filip de Roos, Jonathan Wenger, Marius Hobbhahn, Nicholas Krämer, and Agustinus Kristiadi. The exercises and other material from these tutorials are available only to Tübingen students, via Ilias

The course is aimed at Master students of computer science and machine learning in particular. It provides an introduction to core concepts of machine learning from the probabilistic perspective (the lecture titles below give a rough overview of the contents). The course is designed to run alongside an analogous course on Statistical Machine Learning (taught, in the Summer of 2020, by Prof. Dr. Ulrike von Luxburg). The students who takes this course in Tübingen have also often taken an introductory math refresher, a course on deep learning, and a basic introduction to statistics.

This page provides links to all videos, along with the slides for each lecture. For your convenience, each slide contains a link (the small "play"-icon at the bottom right) directly to a time-stamp in the video of the lecture. Many thanks to Tim Rebig for administrating this functionality. Some of the slides have embedded animations. These are only functional when viewed in Adobe Reader(tm).

The slides and videos are provided under the Creative Commons 3.0 BY-NC-SA license. In particular, this allows non-commercial use for teaching purposes at public institutions of learning, as long as due credit is given and any public distribution of the material happens under the same or a functionally equivalent license. The material comes with no guarantee, expressed or implicit, for correctness, completeness, or anything else. 

If you want to cite the lectures in your academic work, feel invited to do so as "Hennig, P., 2020. Probabilistic Machine Learning. lecture course, University of Tübingen". [here is a bibtex]

Lecture 9 — Gaussian Processes

click here for slides (51MB, hosted externally). Animations require Adobe Reader. 

click here for notebook