HadML: A Novel Deep Generative Approach to Hadronisation

by Andrzej Siodmok (CERN)

Tuesday 27 May 2025, 16:00 → 17:00 Europe/Berlin

ENC-D308

The general-purpose Monte Carlo generators Herwig, Pythia and Sherpa play a key role in high-energy physics. These tools provide fully exclusive simulations of particle collisions, enabling direct comparisons with experimental data, even when complicated experimental cuts have been applied. Significant progress has been made in improving the perturbative aspects of the generators over the past few years. Increased control of perturbative corrections means that the precision of LHC measurements is now more often limited by the non-perturbative components of the generators, such as hadronisation. In my presentation, I will briefly discuss the two hadronisation models currently available — the cluster model and the string model — before focusing on HadML, a novel approach to hadronisation based on machine learning.

Siegen_Siodmok_2025.pdf