Next to a common basis in AI, you choose courses from four areas:
An intelligent agent is an artificial (computer-based) entity that can act pro-actively, reactively, autonomously and rationally in a dynamic environment. Agents can reason about the situation they are in, plan their actions given their goals, revise their beliefs, learn from experience, adapt to the environment, communicate and cooperate. This area focuses on (1) the logical modelling, programming and application of intelligent agents and multi-agent systems; and (2) the use of probabilistic and sub-symbolic machine learning techniques for making agents learn, and adapt to their environment and to other agents.
The cognitive processing area focuses on how AI techniques can help to understand human behavior and cognition. Specifically, you learn how human behavior can be captured in computer simulations (cognitive models) and how the predictions of these simulations can be tested in experiments. You gain theoretical knowledge about and hands-on experience with modeling and experimentation.
The ability to reason is one of the primary forms of intelligence. This area addresses the question what correct reasoning is, how people can rationally reason with incomplete and uncertain information and resolve conflicts of opinion, how to formalise reasoning as a logic, what the best representations of a logic are, and how reasoning and language interact. The student will learn an array of skills ranging from formal methods to conceptual analysis.
Natural language plays a central role in many sub-domains of AI. Obviously, we would want artificially intelligent systems to be able to perform certain kinds of linguistic behaviour: natural language comprehension, generation, translation etc. Additionally, there are vast amounts of linguistically represented data - think of the proportion of the internet that is text-based - and we would like to be able to use that data for AI tasks. In this area you learn about models of linguistic human behaviour and linguistic output, using symbolic, logical and probabilistic methods. You learn to explore the suitability of these models for natural language processing. Finally, you will also learn a range of experimental techniques to develop and test models.