Topics in Sequential Monte Carlo Samplers

Abstract

This represents the original developments of Sequential Monte Carlo Samplers in the class of solutions that generalise SMC filtering methods to the case of a fixed state-space. This makes such methods exact and applicable for Bayesian inference in context otherwise typically treated by Markov Chain Monte Carlo methods. It is the original work developed on SMC Samplers as a research thesis in Cambridge University that led to subsequent works on this topic.

Keywords: Sequential Monte Carlo Samplers