Using Machine Learning to Improve The Study of Jet Physics
Kyle Cranmer, CDS Professor of Data Science and Professor of Physics at the NYU College of Arts & Science, recently co-authored “Reframing Jet Physics with New Computational Methods” which is available at arXiv.org. The project was presented last month at the 25th International Conference on Computing in High-Energy and Nuclear Physics (vCHEP) by Kyle’s co-author and colleague Sebastian Macaluso, a postdoctoral associate in the NYU Physics Department.
What is a jet? When a single quark or gluon (both are types of elementary particles) flies off solo, it radiates new particles subsequently becoming a cloud of particles, flying in the same direction. This phenomenon in particle physics is referred to as a jet. “Reframing Jet Physics” presents a generative model, Ginkgo, which promotes machine learning research in the area of jet physics. The motivation in developing Ginkgo was to build a model that is simple and easy to describe but simultaneously one that captures the “essential ingredients of parton shower generators in full physics simulations”, the aim also being for it to have a python implementation but also few software dependencies.
In their paper, the team reframe common jet physics tasks in probabilistic terms, including “jet reconstruction, Monte Carlo tuning, matrix element — parton shower matching for large jet multiplicity, and efficient event generation of jets in complex, signal-like regions of phase space”.(1) Ginkgo facilitates research into these said tasks using techniques from machine learning, statistics, and combinatorial optimization. (Combinatorial optimization is a subfield of mathematical optimization that uses mathematical methods to improve an algorithm by limiting the size of possible solutions.)
To read/learn more about the project, please see the following resources listed below.
By Ashley C. McDonald