Trusted AI methods for sequential decision making 2025
The Technical Doctoral School of IT and Design at Aalborg University
Description: The theme of this course is automatic reasoning and learning, with a focus on decision making and in particular sequential decision making. These techniques are relevant whenever repeated interaction with an environment is required. As an example, consider the task of moving a robot from A to B. While a complete path can be statically computed, if the environment is dynamic and uncertain, the planned path can become inaccessible. Thus, on-the-fly adaptation is required.
This course will introduce a number of algorithmic solutions such as planning and safe reinforcement learning, explaining the algorithmic principles behind several offline and online sequential decision-making approaches.
The course will consist of several lectures as well as some hands-on exercises.
Prerequisites: You will need background in a mathematical discipline; in particular, you should have attended a course covering propositional logic. You also need basic programming experience (read and write simple computer programs) and familiarity with standard algorithms like search algorithms. Basic familiarity with machine learning is useful but not required.
Learning objectives: The students are familiar with algorithmic solutions to sequential decision making.
Key literature: TBA
Organizer: Christian Schilling
Lecturers: Alvaro Torralba and Christian Schilling
ECTS: 2.0
Time: 10, 12, 17, 19 and 26 March 2025
Place: Aalborg University (Room: TBA)
Zip code: 9220
City: Aalborg
Maximal number of participants: 20
Deadline: 17 February 2025