You are warmly invited to attend our upcoming statistics seminar hosted by the Department of Mathematics.

On 28 May 2026, Professor Nicolas Chopin from ENSAE, Institut Polytechnique de Paris will give a talk on 'On the complexity of standard and waste-free SMC samplers.'

This talk will take place in person at King's College London in Room G39 in the King's Building from 14:15.

This event is open to all, but to ensure smooth entry on the day, we ask that attendees who are not current members of staff at King's to please click 'Register for this event' to sign up for an In-Person Ticket. 

 

Abstract

On the complexity of standard and waste-free SMC samplers

SMC (Sequential Monte Carlo) samplers are a class of iterative algorithms that generate Monte Carlo approximations on a sequence of target distributions. They may be used either in genuine sequential scenarios (i.e. for on-line learning, where data are processed sequentially), or when there is only target distribution of practical interest (and then one designs an artificial sequence to interpolate between an easy-to-sample distribution and the target distribution). This talk will be based on two recent papers; one (with Hai-Dang Dau, NTU) who introduced waste-free SMC samplers as a better alternative over standard SMC samplers, and another (with Yvann Le Fay, ENSAE), where we study how the complexity (number of likelihood evaluations) of these samplers scale with respect to various aspects of the target distributions, such as the length of the sequence, the mixing times of the Markov kernels, the dimension of the ambient space, and so on. These complexity results leads to practical guidelines on how to obtain optimal performance for these algorithms, as I will argue at the end of my talk.

About the Speaker: 

Professor Nicolas Chopin is a Professor of Data Science, Statistics, and Machine Learning at ENSAE, Institut Polytechnique de Paris. His research focuses on Bayesian computation, particularly the development of algorithms for Bayesian inference. His interests include Monte Carlo methods, especially Sequential Monte Carlo, as well as Markov chain Monte Carlo, quasi-Monte Carlo, and fast approximation techniques such as Expectation Propagation and variational Bayes. He currently serves as a joint editor of JRSSB.