Gijs Hofstee - Different approaches to belief revision in multi-agent systems

Logic has been studied since ancient Greece (and probably before). In more recent ages, it has been used and applied in different fields like mathematics, philosophy and robotics, each suited to their own purpose. Not many look at all three fields at the same time, to determine what differences there are, or to which extend they can serve the same purpose. In this colloquium we shall investigate three different techniques to model beliefs and belief change. Our first technique is Jeffrey conditioning, which is based on the Bayes Theorem and a typical case of probability theory. The second technique is Interpreted Systems, which aims to model belief and belief change over a series of observations, and as it turns out is largely compatible with the AGM theorem. Finally, we shall look at Dynamic Epistemic Logic (DEL), and how it can use event models to handle belief change. For all three techniques a short explanation shall be given about what these techniques are about, and how they can be applied to model and agents beliefs when given public and private announcements.