Introduction在线视频

Abstract: 

I present two simple logics for reasoning about the process of inductive learning from successive observations. This formalism combines ideas from Epistemology, Learning Theory, Dynamic Epistemic Logics and Subset Space Logics. Semantically, we take intersection spaces (a type of subset spaces that are closed under finite non-empty intersections), with points interpreted as possible worlds and neighborhoods interpreted as observations or"information states", and enhance these structures with an AGM learner L, i.e. a function mapping every information state to a conjecture (representing the learner’s strongest belief in this state), that satisfies the AGM postulates for belief revision. At the syntactic level, we extend Subset Space Logic with dynamic observation modalities, as well as with a learning or "belief" operator. I give a complete axiomatization of this logic, study its expressivity and use it to characterize various notions of knowledge, belief, knowability and learnability. Time-permitting, I provide a topological characterization of inductive solvability of empirical problems, and use it to prove that AGM-style belief revision is "universal": every inductively solvable problem can be solved by AGM learners. This talk is based on joint papers with Nina Gierasimczuk, Aybuke Ozgun, Ana Lucia Vargas and Sonja Smets.

Speaker：

Alexandru Baltag is now an Associate Professor at ILLC (Institute for Logic, Language and Computation), University of Amsterdam, NL. His research interest includes but is not restricted to the following topics: modal logic, dynamic logic, epistemic logic, temporal logic; models for multi-agent information flow and information merge (learning, belief revision, communication, persuasion, belief aggregation); quantum logic and quantum information flow; coalgebras, non-well-founded sets, Universal Set Theory, models for self-reference, circularity and fixed-points; rationality and action in Game Theory; formal epistemology, philosophy of information and philosophy of science.