Energy-weighted message passing: an infra-red and collinear safe graph neural network algorithm

• Published in: The journal of high energy physics Vol. 2022; no. 2; pp. 60 - 30
• Main Authors: Konar, Partha, Ngairangbam, Vishal S., Spannowsky, Michael
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 08.02.2022; Springer Nature B.V; SpringerOpen
• Subjects: Algorithms; Bosons; Classical and Quantum Gravitation; Criteria; Deep learning; Elementary Particles; Energy flow; Gluons; Graph neural networks; High energy physics; Jets; Large Hadron Collider; Machine learning; Message passing; Neural networks; Physics; Physics and Astronomy; QCD Phenomenology; Quantum chromodynamics; Quantum Field Theories; Quantum Field Theory; Quantum Physics; Quarks; Regular Article - Theoretical Physics; Relativity Theory; Safety; String Theory; Jets; QCD Phenomenology
• ISSN: 1029-8479; 1126-6708; 1127-2236; 1029-8479
• DOI: 10.1007/JHEP02(2022)060

A bstract Hadronic signals of new-physics origin at the Large Hadron Collider can remain hidden within the copiously produced hadronic jets. Unveiling such signatures require highly performant deep-learning algorithms. We construct a class of Graph Neural Networks (GNN) in the message-passing formalism that makes the network output infra-red and collinear (IRC) safe, an important criterion satisfied within perturbative QCD calculations. Including IRC safety of the network output as a requirement in the construction of the GNN improves its explainability and robustness against theoretical uncertainties in the data. We generalise Energy Flow Networks (EFN), an IRC safe deep-learning algorithm on a point cloud, defining energy weighted local and global readouts on GNNs. Applying the simplest of such networks to identify top quarks, W bosons and quark/gluon jets, we find that it outperforms state-of-the-art EFNs. Additionally, we obtain a general class of graph construction algorithms that give structurally invariant graphs in the IRC limit, a necessary criterion for the IRC safety of the GNN output.