Nested Sampling for Rare Event Estimation

Abstract

Nested Sampling is a relatively recent development and could be framed as “Lebesgue Monte Carlo integration” (as opposed to usual Monte Carlo methods which could be said to perform “Riemann Monte Carlo integration”). I will give an introduction to Nested Sampling and how it can be used to compute the log-probability of (very) rare events. This is ongoing work with Jonas Latz and Doris Schneider.