Bocconi Postdoc Honored with 2024 Laplace Award for Advancements in Bayesian Statistics

Francesco Pozza, postdoctoral researcher at the Bocconi Institute for Data Science and Analytics, has been awarded the Laplace Award by the American Statistical Association Section on Bayesian Statistical Science (SBSS) in recognition of his outstanding contributions to the field of Bayesian statistics. His paper, “Skewed Bernstein-von Mises theorem and skew-modal approximations”, co-authored with Daniele Durante and Botond Szabo (both associate professors at the Bocconi’s Department of Decision Sciences), introduces novel methods which address the problem of characterizing forms of asymmetry that often arise in Bayesian inference. Pozza's research is celebrated for its balance between applied, computational, and theoretical advancements, which significantly improve the accuracy of approximations arising in Bayesian statistics.

The paper addresses important challenges in Bayesian statistics, where the posterior distribution is often asymmetric but intractable and is therefore typically approximated by symmetric solutions for computational convenience, potentially introducing severe bias into statistical inference. His innovative approaches mitigate this problem by introducing a general, theoretically motivated and computationally efficient class of asymmetric approximations. The application of these methods leads to more accurate results compared to current solutions and therefore has the potential to find wide applicability in the many scientific fields where Bayesian statistics is widely used.     

The Laplace Award, named after the renowned French mathematician Pierre-Simon Laplace, is presented annually to the best paper among the ten winners of the ASA-SBSS Student Paper Competition.