Konstantin Genin
(Carnegie Mellon University)
Learning with Ockham: Simplicity in Inductive Inference
October 24th at 18:00, in ILLC Seminar Room (F1.15)
Is it possible to give a non-circular justification for the systematic scientific bias in favor of simpler theories? Combining ideas from belief revision, learning theory and topology, we give a justification for Ockham's razor grounded in truth-finding efficiency. Connections are demonstrated between monotonicity principles in belief revision, Popperian falsificationism and simplicity.