Seminar session on Beliefs Dynamics and Probabilities on February 17th

This talk is a presentation of my paper of the same title. In the talk, I will present a logic combining Dynamic Epistemic Logic, a framework for reasoning about multi-agent communication, with a new multi-agent version of Justification Logic, a framework for reasoning about evidence and justification. This novel combination incorporates a new kind of multi-agent evidence elimination that cleanly meshes with the multi-agent communications from Dynamic Epistemic Logic, resulting in a system for reasoning about multi-agent communication and evidence elimination for groups of interacting rational agents.

Dynamic logic describes change, how processes, states, or situations change given certain actions or events. One way probability may play a role is if the results of the actions or events are not entirely certain, and this uncertainty can be quantified using probability. Another way probability and dynamics are combined is when it is the probabilities themselves that are subject to change as new information is provided. Bayesian updating is the most accepted mechanism for change in probability that results when such information received is fully understood and taken to be true. But less trusted or understood information may require different types of updating. One way to justify probabilistic updating mechanisms is to adapt to the new circumstances a betting game argument, called a Dutch book argument, that has been used to justify Bayesian updating through the possibility of a winning strategy in certain subgames. This talk provides an overview of ways dynamic logic and probability are combined.