The Bernoulli Society for Mathematical Statistics and Probability has recently announced the awardees of the 2024 Bernoulli Prize for an Outstanding Survey Article in Probability or Statistics. Prof. Christophe Ley from the University of Luxembourg together with Prof. Gesine Reinert from the University of Oxford and Prof. Yvik Swan from the Université Libre de Bruxelles have been selected for their article “Stein’s method for comparison of univariate distributions”.
New version of Stein’s method
Giovanni Peccati, Head of the Department in Mathematics at the University of Luxembourg and a specialist in probability theory explains: “Stein’s method is a collection of analytical and probabilistic techniques that allow one to answer the fundamental question: “how can we explicitly measure the discrepancy between two random phenomena?”. The crucial insight of the method, introduced by the statistician Charles Stein in 1972, is that one can perform such a task – to an astounding degree of precision! – by using some special class of ordinary differential equations. Stein’s method has become one of the most powerful and flexible tools for analysing probabilistic and statistical models, with applications to fields as diverse as mathematical physics, machine learning, mathematical finance, computer sciences, population genetics, and so on. The paper by Ley, Reinert and Swan proposes an original unifying approach to Stein’s method based on deep ideas of a functional analytical nature. Such a work is a landmark contribution to stochastic analysis; I am expecting that its impact on the field will continue to increase as time passes. Congratulations, Christophe!”
The aim of the Bernoulli Prize for an Outstanding Survey Article is to recognise authors of an influential survey publication in the areas of probability and statistics, respectively. The paper should be timely in addressing areas of active or emerging importance but have been in circulation long enough for there to be evidence of its impact. In this case, the paper “Stein’s method for comparison of univariate distributions” was published in 2017 in the journal Probability Surveys.
“I am very glad to receive this prestigious award with Gesine and Yvik, and in particular for this paper. This has several reasons. First, the paper has played a pivotal role for Stein’s Method to become influential in computational statistics and machine learning. Second, because it reminds me of our nice time spent at Oxford working on it. And third, because it took quite some time to get this work done and our viewpoint on Stein’s Method to be accepted in the community. The award means that all efforts have been worth it”, comments Prof. Ley.