doc. Mgr. Viliam Lisý, MSc., Ph.D.
head_person_supervisor
Ing. Ondřej Kubíček
Department of Computer Science
Multi-agent Reinforcement Learning in Large Imperfect-information Games
Ing. David Milec
Department of Computer Science
Facing Sub-rational Opponent in Imperfect-Information Games
Ing. Michal Šustr
Department of Computer Science
Algorithms for Large Two-player Zero-sum Imperfect-information Games
Supervisor specialist
Ing. Petr Tomášek
Department of Computer Science
Solving Undiscounted One-Sided Partially Observable Stochastic Games
Ing. Jaromír Janisch
Department of Computer Science
Applications of Deep Reinforcement Learning in Practical Sequential Information Acquisition Problems
Dissertation topics
Multi-agent Reinforcement Learning in Large Imperfect-Information Games
- Branch of study: Computer Science – Department of Computer Science
- Department: Department of Computer Science
-
Description:
The student will explore the relation of multi-agent reinforcement learning with methods developed particularly for solving extensive form games. The goal is to understand which approach works better in which domain and ideally work towards a unified framework combining aspects of both approaches. The ultimate goal is to have a set of algorithms that are able to learn to play an arbitrarily complex game on a super-human level.