Model-based reinforcement learning combines planning and learning in a single algorithmic approach. Some example directions in which we work:

• Algorithmic connections of planning and learning

• Combinations of planning and learning in games

• Interpretation of value equivalent models

People:

• Aske Plaat

• Thomas Moerland

• Joost Broekens

• Hui Wang

• Andrius Bernatavicius

• Joery de Vries

• Ken Voskuil

Example papers: