Module 24-M-PT-MML Mathematics of Machine Learning

Non-official translation of the module descriptions. Only the German version is legally binding.

Students expand and deepen their mathematical knowledge and skills in the field of Maschine Learning.

Course type Lecture with tutorials:
The students master the basic contents and methods of a special subject area of Mathematics of Maschine Learning, in particular they can carry out independently complex in this area requiring a high level of mathematical expertise.
Furthermore, students recognise further-reaching connections to previously acquired mathematical facts. They can transfer and apply the knowledge and methods they have learnt so far to other, deeper mathematical problem areas. Students also expand their mathematical intuition through further and more intensive study.
In the tutorials, students develop their ability to discuss mathematical topics and thus further prepare themselves for the requirements of the Master's module, in particular for the scientific discussion within the Master's seminar presentation and the defence of their Master's thesis.

Course type Seminar:

Students are able to give a specialised mathematical presentation independently. They can independently develop a mathematical problem from Mathematics of Maschine Learning, prepare it for a presentation and present it in an understandable way in the presentation and prepare a technically correct elaboration on the contents of the presentation. They will be able to independently fill any gaps, e.g. missing proofs/proof steps or missing illustrative examples.
With the seminar presentation and the preparation of the presentation, students develop both their ability to discuss and write mathematical texts. This prepares them further for the requirements of the Master's module, in particular the writing of the Master's thesis, the Master's seminar presentation including scientfic discussions and the defence of their Master's thesis.

In the lecture and tutorials, various aspects of a subject area of the mathematics of machine learning are taught.
In the seminar, students give a presentation on a mathematical problem of machine learning. The questions raised in the presentation are discussed with the participants of the seminar. Afterwards, the students prepare a paper on the presentation.

The courses in this module lead methodically and in terms of content to current research questions in the field of the mathematics of machine learning. Possible contents include

• Neural networks
• Generative networks
• Reinforcement learning,
• Unsupervised Learning
• Supervised and online learning
• Statistical learning theory
• Optimisation for machine learning
• Dimensionality reduction
• Kernel methods
• Model selection

Solid knowledge of probability theory and statistics. Depending on the chosen subject, the requirements will be specified in the course announcement.

—