The second meetup this month has been rescheduled to happen ONE WEEK EARLIER, i.e. this coming Sunday at noon. There are two reasons for this: Our speaker, Martin Magnusson, is available on that date, and both he and I are unavailable on the regular meetup day (fourth Sunday of the month).

Martin provided this abstract:

Logic Leaves the Lab: Building Smart Computer Game Characters

When I entered academia as a new PhD student in logic-based AI, I was haunted by a question. Why is everyone developing new logics while virtually no one actually uses them to solve practical problems?

Since then I've discovered just how hard a problem AI is. There's still plenty of work to do before we can formalize e.g. an encyclopedia entry. But I believe the advancements of the past 50 years of research enable promising (and fun!) applications in the relatively simple worlds of computer games.

In this talk I'll present some arguments for computer games as a practical application of logic-based AI, sketch some game scenarios that involve logical reasoning, and present my own attempts at solving them. The idea is to build an intelligent agent architecture based on logical reasoning. By unifying different tasks in logic we achieve a tight integration between them. This is necessary in games where the knowledge is relatively simple but changes quickly.

Read more about my work at http://www.martinmagn... .