Valentin Goranko: Logics for visual-epistemic reasoning in multi-agent systems

In this talk I will present a logical framework for multi-agent visual-epistemic reasoning, where each agent receives visual information from the environment via mobile camera with a given angle of vision in the plane. The agents can thus observe their surroundings and each other and can reason about each other's observation abilities and knowledge derived from these observations. I will introduce suitable logical languages for formalising such reasoning, involving atomic formulae stating what agents can see, multi-agent epistemic operators, as well as dynamic operators reflecting the ability of agents (or, their cameras) to move and turn around. I will then introduce several different types of models for these languages and will discuss their expressiveness and some essential validities. Lastly, I will discuss some basic model-theoretic problems arising in this framework that open up new directions of study, relating logic, geometry and graph theory.

This talk is partly based on a recent joint work with Olivier Gasquet (Univ. Paul Sabatier, IRIT, Toulouse) and Franois Schwarzentruber (ENS Rennes).