Manuel Szewc: Machine Learning for Monte Carlo simulators in collider physics

Monte Carlo simulators are a fundamental part of the scientist's toolkit. In particle physics, simulators are needed to relate theory and experiment and have been the subject of continuous development. However, current and future collider experiments are pushing simulators to their limits, motivating a surge of Machine-Learning-based proposals. In this talk, we address one particular problem: the empirical modelling of hadronization, the transition from unobservable partons to measurable hadrons. We detail current efforts by the MLHad collaboration in designing, training and deploying a ML-based hadronization model that is physics-based, fits available data and can be made part of existing Monte Carlo simulators so as to be usable by the community.