This paper presents an algorithm based on Graph Neural Network for charged-particle track reconstruction in the ATLAS Inner Tracker. Using realistic simulation data, we demonstrate the performance of the algorithm in comparison with the state-of-the-art technique.
We explore the use of conditional normalizing flow in the simulation of interaction between hadronic particles and atomic nuclei in ordinary matter. We trained generative models to reproduce data simulated by the state-of-the-art simulator, conditioned on the kinematics of the incoming hadron.
We studied a support vector machine with a quantum kernel estimator (SQVM-Kernel) for classification of proton-proton final states, targeting the Higgs boson production associated with a pair of top quarks. Using a dataset of 50000 events, we demonstrated the evquivalent performance of the quantum algorithm compared to the classical counterparts.