Estimators of Entropy and Information via Inference in Probabilistic Models

Abstract

Estimating information-theoretic quantities such as entropy and mutual information is central to many problems in statistics and machine learning, but challenging in high dimensions. This paper presents