Sequential Decision Making

Sequential Decision Making under uncertainty is an important skill of an intelligent agent, as planning ahead will allow it to achieve its goals better. This course will deepen the students' knowledge of AI decision making by focusing on the following core aspects. First, we discuss how the goals of an agent can be represented properly. Next, we cover search and/or learning algorithms that use the Markov Decision Process (MDP) framework to compute (sub)optimal decisions in sequential settings. Extensions of the basic MDP framework that are relevant for real-world problems such as partial information, multiple objectives and acting in multiagent systems are also covered. Next, we discuss techniques to improve the safety and robustness of the decisions of agents. Finally, the gained theoretical knowledge is applied in several lab assignments, where students need to design and program an algorithmic solution for decision-making problems.