Generating samples from a given, possibly high-dimensional, probability distribution is a task that appears often in several sciences. The literature contains a high number of methods to perform that task and the talk will focus in one of them: the Hamiltonian Monte Carlo (HMC) algorithm. This widely used technique is remarkable in that it uses ideas from very many scientific fields, including classical mechanics, statistical physics and the theory of structure preserving discretizations of differential equations. The talk, which is aimed at a general audience and requires little background, will explore the connections between HMC and those different fields.

About the Speaker: J.M. Sanz-Serna is an emeritus professor at Universidad Carlos III de Madrid. He has contributed to numerical analysis, approximation theory, functional analysis, Monte Carlo methods and other areas. His main interest has been in the numerical analysis of stochastic and deterministic, ordinary or partial differential equations. He served as Universidad of Valladolid Vicechancellor 1998-2006 and as President of the Royal Academy of Sciences of Spain 2018-2024.