Münster 2017 – scientific programme

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

T: Fachverband Teilchenphysik

T 15: Kosmische Strahlung 1

T 15.6: Talk

Monday, March 27, 2017, 18:00–18:15, H 3

A Method of Reconstructing Ultra-High Energy Cosmic Rays at the Pierre Auger Observatory using Deep Learning — •Jonas Glombitza, David Walz, Marcus Wirtz, Gero Müller, and Martin Erdmann for the Pierre Auger collaboration — III. Physikalisches Institut A, Aachen, Deutschland

The surface detector of the Pierre Auger Observatory in Argentina measures the footprint of muons and electromagnetic particles of ultra-high energy cosmic ray induced air showers on ground level. Reconstructing the properties of the primary cosmic ray such as energy, direction and mass with optimal resolution remains a challenging task. Recently, great progress has been made in multiple fields of machine learning by using deep neural networks and associated techniques. In this talk we present a new method to reconstruct the properties of the ultra-high energy cosmic rays, by training deep neural networks to the detector response of the surface detector. By training the network to identify suitable features in all the available event information, this method has the potential to surpass currently employed methods which build on algorithms processing selected observables. In this context we discuss suitable data representations and compare different network architectures and training procedures. Finally, we assess the performance of the method on simulated air showers.