Content

The course covers central topics of supervised machine learning. As a theoretical foundation it introduces the theory of generalization and the VC bound, as well as practical methods like regularization and validation. The lecture complements courses on neural networks by focusing on classic learning algorithms for so-called tabular data: nearest neighbor prediction, various algorithms using linear models (Perceptron, linear regression, LASSO regression, logistic regression), random forests, and support vector machines. Extensive exercises based on the Python library sklearn are a central component of the course.

Learning Outcomes

The participants can explain statistical learning theory and its implications for what machine learning can and cannot do. They know the most important methods for learning from tabular data, and they can apply these methods to practical problems using Python.

Learning Methods

The course is taught as a one week intense block course, combining lecture and exercise sessions, followed by self-study.

Assessment

Written exam of 90 minutes, likely an electronic exam.

Prof. Dr. Tobias Glasmachers Lecturer

(+49) 234-32-25558 tobias.glasmachers@ini.rub.de NB 3/27

Course type

Lectures

Credits

3 CP

Term

Summer Term 2026

Lecture

Takes place every day from 10:00 to 16:00 in room IC 03/610.
First appointment is on 07.09.2026
Last appointment is on 11.09.2026

Participants need a solid understanding of basic linear algebra (linear functions, vectors, matrices, inner product) and probability theory (distribution, expectation, variance, quantile, combinatorics). The exercise sessions as well as the exam require basic Python programming.

Literature

Book “Learning From Data” by Abu-Mostafa, Malik, and Lin