Inference by Multiple Identical Observers

Abstract

We consider models for inference which involve observers which may have multiple copies, such as in the Sleeping Beauty problem. We establish a framework for describing these problems on a probability space satisfying Kolmogorov's axioms, and this enables the main competing solutions to be compared precisely.

This version (v3) has a new title, and is shorter and more mathematical than v2. Most of sections 4,5,6,7 of v2 have been omitted