ESSLLI 2012 Course on Interactive learning, formal social epistemology and group belief dynamics: logical, probabilistic and game-theoretic models
Alexandru Baltag, Sonja Smets
Content of the course:
This course is addressed to students and researchers interested in formal approaches to multi-agent belief change, interactive learning and doxastic group attitudes induced by various forms of inter-agent communication or interaction. In this sense, we present, compare and relate various models for group knowledge and group belief dynamics, developed by researchers in areas as different as Dynamic Epistemic Logic, Belief Revision theory, the Bayesian theory of rational belief change, Computational Learning Theory, Game theory, as well as Epistemology (especially the new areas of Formal Epistemology and Social Epistemology). We take the Logical perspective as fundamental, and look at ways to integrate within it the other mentioned approaches. We look at applications to: rationality and equilibria in games; epistemic paradoxes; the formalization of classical epistemological conceptions; the issue of tracking the truth by individual belief revision strategies or by pooling up information via merging strategies exploiting the “wisdom of the crowds”; as well as the apparently “irrational” aspects of epistemic group dynamics (informational cascades, epistemic bandwagon effect, pluralistic ignorance).
Timeliness and relevance to ESSLLI
The field of Dynamic Epistemic Logic has rapidly developed in the last 12 years, successfully dealing with various forms of group learning, communication and interaction. In parallel, the new philosophical fields of Formal and Social Epistemology are currently thriving. On the other hand, Belief Revision Theory, Computational Learning theory, the Bayesian theory of probabilistic belief change and the Game- Theoretic study of interactive learning in repeated games are well-established fields, with clear relevance to ESSLLI and with a well-known set of classical results and open problems.
Courses on Dynamic Epistemic Logic, as well as courses on Interactive Belief Revision, on Rational Agency, on Computational Learning and on Bayesian theory of belief update have already been given at several ESSLLI’s. But no ESSLLI course has touched all the approaches mentioned above, despite the fact that great progress in connecting and comparing each two of these different approaches has been registered in recent years (though a number of exciting conceptual challenges and technical open problems remain to be tackled). So we think it is timely to have such a course at ESSLLI.
Course Slides
Lectures 1 to 5 are available to students who registered for the school.
Prerequisites
This course is essentially self-contained. We presuppose some (but very little) background knowledge in Logic: basically, some previous familiarity with the syntax and semantics of modal logic, and maybe with the usual axioms of Epistemic Logic. This should be enough for understanding the course, though obviously having more prior knowledge, in particular of elementary notions of Probability theory, and/or the basic concepts of Game Theory, would be very welcome. More importantly, we assume that participants in the course possess both some degree of mathematical maturity (as can be expected from graduate students in Logic) and a live interest in interdisciplinary connections and applications of Logic.