Modeling Hadronization with Machine Learning

Hadronization is a fundamental step in high-energy particle physics simulations, encoding the transition between unobservable quarks and gluons to physically observable hadrons. Its non-perturbative nature, however, precludes first-principle calculations, and state-of-the-art event generators therefore rely on finely-tuned empirical models to reproduce experimental results. Motivated by the difficulties associated with such models, we present MLhad, a proposed alternative where the empirical model is replaced by a surrogate Machine Learning-based model to be ultimately data-trainable. In this poster, we detail the current stage of development, especially our recent work on generative models with quantifiable uncertainties to be fine-tuned in data.