Bayesian techniques and applications to QCD

Abstract

Realizing the full potential of interconnecting the large amounts of data created in physics experiments, phenomenological models and theory simulations requires robust tools for statistical inference. Here I review a particularly promising branch, Bayesian statistics, which over the past decade has found manifold use in high-energy physics. After a brief introduction to Bayesian statistics I will present two concrete examples, where Bayesian thinking has led to progress in understanding strongly interacting matter: unfolding problems in the form of lattice QCD spectral functions (in spirit similar to detector corrections), as well as the efficient estimation of quark-gluon-plasma parameters from a systematic comparison of experimental heavy-ion collision data and phenomenological models.

Comments: 12 pages, 4 figures, talk given at the XIIIth Quark Confinement and the Hadron Spectrum Conference 2018, Maynooth, Ireland