Ultra High Energy Cosmic Ray Composition using Neural Network with Telescope Array Detectors in Hybrid Mode

Ultra-High Energy Cosmic Rays' (UHECRs) low flux makes direct detection infeasible, so indirect studies of the resulting particle cascade in the atmosphere are used to infer their mass composition. Telescope Array (TA), the largest cosmic ray detector in the Northern Hemisphere, combines three fluorescence telescope sites, each instrumented with 12-14 telescopes, surrounding an array of 507 surface detectors (SDs) to detect UHECRs with energies greater than 1018 eV. However, fluorescence telescopes can only run during dark, clear, moonless nights, while SDs have a ~100% duty cycle, meaning 90% of SD data does not have telescope coverage. We aim to increase TA's capabilities by utilizing neural networks to extract composition from TA SD data, improving the available statistics tenfold compared to the traditional hybrid analysis.