Elements: Machine Learning Quark Hadronization

Particle physics explores the fundamental building blocks of nature, and their interactions. This is accomplished by colliding particles like protons together at high energies, producing many particles which are then measured by experimental detectors. To interpret the data, detailed simulations of particle collisions are needed.

Monte Carlo event generators provide the simulations and are at the core of particle physics discoveries, such as the discovery of the Higgs boson at the Large Hadron Collider in 2012, forming a keystone of the cyberinfrastructure for particle physics. This project addresses a bottle-neck in Monte Carlo event generators — how quarks are bound together inside protons and other composite particles through the process of hadronization.

Hadronization models are improved by this project, including the speed of simulating the process of hadronization, fulfilling a critical step for understanding the scientific impact of upcoming large-scale particle, neutrino, and nuclear physics projects.

Event generators simulate particle collisions in three steps: a high energy collision at the small distances, evolution of the event into a large distance (low energy) configuration, and finally a hadronization into observable particles. The hadronization step is particularly challenging, because it cannot be directly calculated from first principles.

Since hundreds of particles must be produced for each particle collision, algorithmic efficiency is necessary and requires innovative approaches. This project is applying Machine Learning (ML) techniques to hadronization in order to: connect to existing models of hadronization, understand the underlying description of data by current event generators augmented with ML, and provide a fast and accurate ML based simulation of hadronization.

As part of the project, a platform is being built where user submitted modules, such as this ML hadronization module, can be incorporated in the most widely used event generator in the particle physics community, Pythia.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.