Game AI workshop
AIM organizes a workshop together with the University of Maastricht.
- The theory of learning in games is one of the cornerstones of machine learning theory. Key concepts, which are today considered as the gold standard of the theory, like regret and approachability, have their roots in statistical decision theory and nowadays are standard tools in the theory of online learning. The mathematical umbrella of all these notions is the theory of “min-max” to robustify the learning outcomes. Besides robustness, more and more optimization problems in AI involve hierarchical decision making and bilevel structures. This becomes relevant when optimizing over hyperparameters, in sequential experimental design, or in adversarial training of deep neural networks. Such techniques are new instances of leader-follower games, a classical model problem in mathematical programming and Stackelberg game theory. Finally, driven by the huge success of AI in general game playing, a rigorous analysis of search methods is much in demand to understanding the sample complexity of such methodologies better. With all these innovations in mind, the aim of this workshop is to work out cutting-edge research questions arising in these advance the mathematical foundations of AI within 4 key domains of research, each led by renowned experts in the field.
In this workshop, we are focusing on four research domains
- Online learning in complex environments (leader: Tim van Erven (UvA))
- Game AI & Search (leader: Wouter Koolen (CWI))
- Robust AI via optimal transport & mean field games (leader: Christoph Brune (UT))
- Fast methods for large-scale bi-level programming (leader: Tristan van Leeuwen (CWI))
Venue and Date:
Date: March 30 and 31,2023. Lunches and coffee will be provided.
Location: Department of Advanced Computing Sciences (DACS), Faculty of Science and Engineering, Maastricht University, Paul-Henri Spaaklaan 1, 6229 EM, Maastricht.
- Maastricht Centre Mathematics (MCM)
If you have any questions, please contact the organisers
- Mathias Staudigl (UM)
- Barbara Franci (UM)
or send an email to firstname.lastname@example.org
Please note: We only have a limited number of spaces available.
Register now by filling in the contact form below and join us at our workshop!