The Cherenkov Telescope Array (CTA) will be the next generation ground-based very-high-energy gamma-ray observatory, sensitive from 20 GeV up to 300 TeV. In this energy range, the overall sensitivity of CTA and its angular resolution will significantly exceed the performance of all existing ground-based gamma observatories. Unprecedented precise data provided by CTA will shed light on the origin of high-energy cosmic rays, or the nature of dark matter. CTA will consist of several dozens of telescopes built on two sites selected to cover the entire sky. The Large-Sized Telescope prototype (LST-1) was inaugurated in October 2018 in La Palma (Spain) and it is currently in the commissioning phase. Cherenkov telescopes in general detect the cosmic gamma rays (which cannot penetrate the Earth's atmosphere) indirectly by observation of Cherenkov radiation of secondary shower particles, resulting from the interaction of the primary cosmic gamma-ray photon with atomic nuclei in the atmosphere. Reconstruction of properties of the primary gamma-ray photon requires precise Monte Carlo simulations including particle interaction models, models of the atmosphere, and models of the telescope itself.
The thesis will investigate advanced modern methods of shower reconstruction using convolutional neural networks. The thesis aims to characterize the performance of such reconstruction and its comparison with standard reconstruction methods, particularly for the classification of primary particles.