Course content

This is an advanced course in machine learning, focusing on modern probabilistic/Bayesian methods: Bayesian linear regression, Bayesian networks, latent variable models and Gaussian processes, as well as methods for exact and approximative inference in such models. The course also contains necessary probability theory and methods for data dimensionality reduction.

The course includes theory (e.g., derivations and proofs) as well as practice. The practical part will be implemented using Python.

Two versions

The course exists in two versions - 1RT705 (5 course credits) and 1RT003 (7.5 course credits).

If you are registered as a PhD student, the course codes are FTN0092 (5 course credits) and FTN0204 (7.5 course credits).

Course Structure

Lectures: 14 (1RT003), 10 (1RT705)
Exercise sessions: 11 (1RT003), 8 (1RT705)
Mini project: 1 (mandatory for all)
Peer-review: 1 (mandatory for all)
Computer lab: 1 (mandatory only for 1RT003)
Exam: 1 (mandatory for all, shorter exam for 1RT705)
Literature : A few different sources, all available online
Language of instruction: English

Click on the link above to get further info about each course element.

Schedule

Campus vs online teaching

The course will a hybrid campus and online teaching. The lectures will be in general be given on campus. Note that lecture 2 is only given on zoom. There might be more lectures later in the course rescheduled to be on zoom depending on how many students end up following the lectures on campus. Our aim is also to stream the lectures on zoom and record that streaming. If this online-solution does not work out, we will provide links to the recorded lectures from last year in case you are unable to attend the lectures on campus. However, be aware that the sequence of the lectures is slightly different from last year.

The pen and paper exercise sessions will be given both on zoom and on campus and computer exercise session only on zoom. You find all information what is given on campus and zoom in the schedule.

We will use the same zoom meeting room for all teaching activities that are on zoom, both for the streamed lectures and the exercise sessions, which you find here or via "Zoom" in the left menu. The zoom meeting room is named "Advanced probabilistic machine learning, HT2021". It is required that you are logged in to Studium and registered to the course to see the meeting link.

Covid-plan

In case the covid-situation gets worse, all teaching will be given via zoom at the same time points as they should have been given on campus. If this happens, this decision will be announcement via an Announcement.

Exercise sessions

During the exercise sessions, the teaching assistant will give a short introduction of the material related to that session and during the session provide insight and comments on the solutions. Except for this, you are supposed to be active and work with the suggested exercises and the teaching assistants will be available for questions and discussion. Take this opportunity to interact! The exercises will be available ahead of time before the session. There are two time slots for each exercise session in the schedule. Choose the one that fits best with your schedule.

Formalities

The course is 5 credits for 1RT705 and 7.5 credits for 1RT003. Entry requirements are: 120 credits, including Statistical Machine Learning, Probability and Statistics, Linear Algebra II, Single Variable Calculus, a course in multivariable calculus and one basic programming course.

Course evaluation and course report

Course evaluation and course report for the 2021 course version can be found here (login required).

Teachers


Niklas Wahlström (course responsible, lecturer)	Jens Sjölund (lecturer)	Antônio H. Ribeiro (lecturer)	Thomas Schön (lecturer)

David Widmann (teaching assistant)	Carl Andersson (teaching assistant)	Paul Häusner (teaching assistant)

Advanced Probabilistic Machine Learning, (samläsning) HT2021