Vision Transformer for Gamma-Hadron Classification with the HAWC Observatory

The High Altitude Water Cherenkov (HAWC) gamma-ray observatory consists of 300 water Cherenkov detectors, each of which contains four photomultiplier tubes (PMTs). In the atmosphere, both cosmic rays and gamma rays produce air showers containing cascades of ionized particles and electromagnetic radiation. We train a neural network to distinguish gamma-ray events from cosmic ray events using a novel semi-supervised method, which uses real data to train the network. We show that the network trained using our method outperforms a network trained with simulated data.