On Thursday, November 21, we will have a LIRa session with Dominik Klein.
Everyone is cordially invited!
Speaker: Dominik Klein (University of Tilburg)
Title: Reasoning about Knowledge in Games
Date and Time: Thursday, November 21, 2013, 15:30-17:30
Venue: Science Park 107, Room F1.15
Abstract:
In this talk, we are interested in epistemic logic as a language to reason about interactive situations. The talk is divided into two, interrelated parts.
The first part of the talk deals with the relationship between epistemic logic and epistemic game theory. Generally, the play of a game does not only depend upon the game structure, but also upon the epistemic status of the players involved. Therefore epistemic game theory has developed models of game situations that incorporate the epistemic status of agents and uncertainty thereof by an explicit parameter, the agents epistemic type.
Especially for multi round games, such models need to be rich enough to account for all kinds of dynamics that could occur: The moves players make over time as well as in-play communication or external events influencing the game. Thus, realistic epistemic game models tend to become very complex and unhandy, especially for sequential games.
On the other hand, dynamic epistemic logic has developed methods such as product updates to incorporate informational change from outside a given model. These methods allow for simple and intuitive representations of an initial situations while still maintaining the possibility to allow for all kinds of informational events.
In this talk we show how to connect epistemic logic to epistemic game theory, thereby allowing for smaller and more intuitive models of game situations.
In the second part of the talk, we focus on limited reasoning about an interactive situation. Agents might not be interested in the full wealth of facts expressible in the langauge of epistemic logic about a certain situation. Due to interest or cognitive limitations, they might only be intersted in a certain kind of statements, for instance positive knowledge about some specific fact. In this part of the talk we study how the fragment of the language an agent picks for reasoning influences his expressive power.