Presented By: Department of Philosophy
Foundations of Belief & Decision Making Lecture: Entropy and Indifference
Anubav Vasudevan, University of Chicago
Abstract: The Judy Benjamin problem has often been cited as grounds for rejecting the use of entropy maximizing (or information minimizing) methods in Bayesian epistemology. In this talk, I will present an alternative explanation of the counterintuitive phenomenon that lies at the heart of the Judy Benjamin problem, in an effort to show that this critique is misguided. While the Judy Benjamin Problem does not invalidate the principle of maximum entropy, it does serve to bring into sharp relief the prerequisites for its proper application. Reflection on these prerequisites reveals that, contrary to the claims of E.T. Jaynes, who was the first to propose that the principle of maximum entropy be viewed as an a priori principle of rationality, maximum entropy methods should not be viewed as a supplement to the classical Bayesian theory of probabilistic reasoning based on the Laplacean principle of indifference. Rather, the principle of maximum entropy turns out to be a corollary of Laplacean principle of indifference, properly understood. I will conclude the talk by showing how the principle of maximum entropy can be used to resolve certain conceptual puzzles in the theory of probabilistic reasoning, including the apparent cotenability of symmetry and frequency-based assessments of probability.
Explore Similar Events
-
Loading Similar Events...