Gamma Ray Source Localization for a Time Projection Chamber Telescope using Convolutional Neural Networks

Diverse phenomena such as antimatter throughout the Milky Way, merging binary neutron stars, and dark matter can be better understood by studying their gamma ray emission. Despite their importance, low MeV gamma rays have been insufficiently explored at sensitivities that would allow for deeper insight into the nature of the gamma emitting objects. In response, a liquid argon time projection chamber (TPC) gamma ray observatory called GammaTPC has been proposed and promises exploration of the entire sky with a large field of view, large effective area, and high polarization sensitivity. Optimizing the pointing capability of this instrument is crucial and can be accomplished by leveraging convolutional neural networks to reconstruct electron recoil paths from Compton scattering events within the detector. In this investigation, we modify a previously developed machine learning model architecture to accommodate a larger data set of high fidelity simulated electron tracks. We create two modified model architectures: one to predict the electron recoil track origin and one for the initial scattering direction. We find that these models predict the true origin and direction with high accuracy, thereby optimizing the observatory's estimates of the sky location of gamma ray sources.