• Grossenbacher Bastian Alexander

• Grossenbacher Bastian Alexander

The goal of this course is to understand the foundation of Machine Learning as a field, providing the basis to master its many branches and applications. After an introduction about how machines learn, the focus will be on a short selection of key algorithms for supervised, unsupervised and reinforcement learning. The students will learn how parametrized function approximators can be used to take decisions, how to update their parametrization to modify their behavior, and how to leverage data and interactions in real-world applications.

The course is composed of theoretical lectures, explaining the inner working and intuitions behind the methods, interleaved with practical sessions to gain hands-on experience. The students will be introduced to both re-implementing (and customizing) some of the basic algorithms, and to applying the current standard libraries on practical applications.

The theoretical requirements include a basic understanding of linear algebra, calculus and statistic, as far as appropriate to understand the algorithms' inner workings. The practical sessions require a degree of familiarity with the Python programming language, and a working installation for the exercises. All lectures and material will be in English.

 

This course is set to provide a fundamental understanding of what is Machine Learning and how all of its methods operate, with practical experience on selected algorithms and libraries. The core objective is to support and facilitate further learning on the topic, both in the form of further self-study and as an introduction for advanced classes available in the Masters course.

Please register to the course on the students portal < https://my.unifr.ch >; in case of problems write an email with your name, Nr SIUS, Code and Course Name to Stephanie Fasel < stephanie.fasel@unifr.ch >.
All official communication will go through Moodle, please register at your earliest convenience https://moodle.unifr.ch/course/view.php?id=274767. All lectures will be online for the time being, using Microsoft Teams: access details will be made available on Moodle.

Les unités d’enseignement se composent généralement de deux heures de cours et une heure d’exercices par semaine. Nous vous prions de bien vouloir vous conformer aux délais d’inscriptions aux épreuves de la Faculté des sciences et de médecine.

- Y. Abu-Mostafa, M. Magdon-Ismail, L. Hsuan-Tien. Learning from Data. AMLBook.
  http://amlbook.com/
- M. Bishop. Pattern Recognition and Machine Learning. Springer.
  https://www.springer.com/gp/book/9780387310732
- I. Goodfellow, Y. Bengio, A. Courville. Deep Learning. MIT Press.
  https://www.deeplearningbook.org/
- R. Sutton, A. Barto. Reinforcement Learning. MIT Press.
https://mitpress.mit.edu/books/reinforcement-learning-second-edition

Selon modalité A de l'annexe du plan d'études en informatique

Access requires passing 09 out of 12 of the weekly assignments

Open book exam