Artificial Intelligence, 5 credits

About the course

"What is intelligence"? is a deep philosophical question. The topic of this course is more practical: how can we make computers perform tasks that - up to now - are difficult for computers? Tasks that require an "intelligent" approach, because computing power alone is not enough.

Even before the first computer was built, it was thought of as possibly intelligent. Obviously, computers would talk - a three-year-old child can talk. Playing chess, now that would be intelligent! A computer has since beaten the world Master in chess and fulfilled many other promises. This course is about the successes of artificial intelligence (AI). However, knowing how they work takes away part of the magic, or "intelligence".

Artificial intelligence requires knowledge representation and reasoning. There are many different approaches. In expert systems knowledge consists of facts and rules, clearly visible and separate from reasoning. In neural networks, knowledge is implicit and entangled in the reasoning process. The reasoning is not a straightforward computation. It is a search: making guesses, and trying different solutions. At the heart of AI is heuristic search: methods that try to make "intelligent" guesses.

Nowadays many applications contain AI, often in disguise. Google finds relevant search results. You can talk to a computer and book a railway ticket. The monsters in computer games come right at you. This diversity of applications is reflected in the topics of the course: expert systems, learning, planning, playing games like chess, and so on.