Speaker
Description
The origin of ultra-high-energy cosmic rays (UHECRs) is one of today's most pressing questions in astroparticle physics. Identifying their sources requires precise knowledge of the UHECR mass composition. Since the direct detection of UHECRs is not practical, composition analyses rely on mass-sensitive observables reconstructed from extensive air showers induced by UHECRs in the atmosphere. Ground-based detector arrays detect the footprint of extensive air showers with high uptime, enabling high-statistics measurements of these observables. The surface detector of the Pierre Auger Observatory comprises three sub-arrays of multi-detector stations arranged in differently spaced triangular grids, which detect the spatio-temporal signal of the shower footprint. In this contribution, we show how neural networks can be used to extract different mass-sensitive observables from local and global measurements taken by the different sub-arrays. Using the depth of the shower maximum as an example, we demonstrate how to transition from simulation to data. Moreover, we show that the reconstruction agrees with independent measurements by the fluorescence detector of the Pierre Auger Observatory.