SMuK 2023 – scientific programme

Parts | Days | Selection | Search | Updates | Downloads | Help

T: Fachverband Teilchenphysik

T 88: Gamma Astronomy IV

T 88.5: Talk

Wednesday, March 22, 2023, 18:30–18:45, POT/0151

MAGIC Event Reconstruction with Deep Learning — •Jarred Gershon Green for the MAGIC collaboration — Max Planck Institute for Physics, Munich, Germany

The Major Atmospheric Gamma Imaging Cherenkov (MAGIC) telescope is a stereoscopic system used for detecting gamma rays in the GeV to TeV range. When gamma rays and cosmic rays interact with the atmosphere, an air shower is initiated which itself emits Cherenkov photons detectable by MAGIC. After parametrizing the images of each shower, machine learning algorithms like random forests are used to reconstruct the properties of each primary particle, including their type, energy, and arrival direction. Convolutional Neural Networks offer a promising way to perform this reconstruction directly on pixelated camera images. In this contribution, we explore how deep learning algorithms like convolutional and graph neural networks can be used to reconstruct events, first by introducing architectures and then showing their performance as applied to real MAGIC data.