Inhoud: The course aims to show how (reinforcement) learning techniques can be used to create adaptable agents. I.e. agents that can dynamically adapt their behaviour to the environment based on the results of their actions on that environment. The largest part of the course will be about reinforcement learning techniques using Q-learning, SARSA, etc. The last part of the course will show some other machine learning techniques such as neural networks. Finally we show how all these techniques can be used in the decision making mechanism of the agents and how multi-agent learning can be done.

http://www.cs.uu.nl/education/vak.php?vak=INFOMAA&jaar=2008