Hybrid Quantum-Classical Graph Neural Networks for Particle Track Reconstruction at the Large Hadron Collider

Download
2021-8-5
Tüysüz, Cenk
Particle collider experiments aim to understand Nature at small scales. Particle accelerators, such as the Large Hadron Collider (LHC) at the European Organization for Nuclear Research (CERN), collide particles at high rates (MHz) and high energies (TeV) in order to probe such small scales. High collision rates may bring many computational challenges. One of these challenges is particle track reconstruction, which is the task of distinguishing the trajectories of charged particles passing through the detector. The upcoming High Luminosity upgrade of the LHC is going to increase the collision rates and require more computational resources. Particle track reconstruction algorithms will also be under much more stress, as the current algorithms are scaling worse than quadratically. This work presents a hybrid Quantum-Classical model to solve the particle track reconstruction problem by combining novel Graph Neural Networks with Quantum Neural Networks that are compatible with Noisy Intermediate Scale Quantum (NISQ) computers. Results indicate that the hybrid model can match the performance of the classical model within the limits of 16 qubits and 16 hidden dimensions.
Citation Formats
C. Tüysüz, “Hybrid Quantum-Classical Graph Neural Networks for Particle Track Reconstruction at the Large Hadron Collider,” M.S. - Master of Science, Middle East Technical University, 2021.