Machine Learning Techniques for Calorimetry

Abstract

The Compact Muon Solenoid (CMS) is one of the general purpose detectors at the CERN Large Hadron Collider (LHC), where the products of proton–proton collisions at the center of mass energy up to 13.6 TeV are reconstructed. The electromagnetic calorimeter (ECAL) is one of the crucial components of the CMS since it reconstructs the energies and positions of electrons and photons. Even though several Machine Learning (ML) algorithms have been already used for calorimetry, with the constant advancement of the field, more and more sophisticated techniques have become available, which can be beneficial for object reconstruction with calorimeters. In this paper, we present two novel ML algorithms for object reconstruction with the ECAL that are based on graph neural networks (GNNs). The new approaches show significant improvements compared to the current algorithms used in CMS.