Speaker
Description
We present a study of the muon content of extensive air showers registered by the Telescope Array Surface Detector (TA SD). We first produce dedicated simulations of the TA SD detector response for two modern high-energy hadronic interaction models — EPOS LHC-R and SIBYLL 2.3e in a muon-enhanced tuning (denoted SIBYLL*) — in which the muon component of each event is recorded separately, giving the true muon content of the simulated events in each model. We then train a neural network on these simulated events to estimate the muon signal directly from the recorded SD waveforms. Applying the trained network to TA SD data, we obtain an estimate of the integral muon signal and compare it with the predictions of both models, providing a data-driven assessment of the muon content of air showers and of a possible muon excess.