Date: October 21, 2011 Andreas Witzel, CIMS
Title:
Epistemic logic and dynamics of information change
Epistemic logic is the formal study of reasoning about knowledge and
beliefs. Starting from a philosophical motivation of defining exactly
what it means to "know" something, epistemic logic has classically
focused on describing static situations. Later, the focus was extended
to encompass changing information, raising a plethora of new
questions. Comparatively recently, a formalism was developed that
represents knowledge (or uncertainty) about events in much the same
way as knowledge (or uncertainty) about situations. This formalism,
called Dynamic Epistemic Logic, allows specifying situations and
events in a uniform and modular way, which makes it interesting for AI
applications.
In my talk, I will first give an introduction to modelling knowledge
and beliefs using epistemic modal logic, briefly reviewing some
underlying intuitions and philosophical foundations of classical
single and multiagent epistemic logic.
Then, I will present Dynamic Epistemic Logic, a conceptually simple,
but surprisingly rich framework for modeling the dynamics of
information change, which has become an active research area within
modal logic.
Due its modularity, the framework also seems to lend itself to
applications in AI. For example, one can envision a vitual world
keeping track of its inhabitants' knowledge about facts and about each
other, and enabling agents to respond adequately to events that change
their knowledge states; or automated planning using epistemic
conditions, effects, and goals. I will present some recent research
efforts in that direction.
I will not assume familiarity with modal logic, but some mathematical
inclination will be useful.
