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

AKPIK: Arbeitskreis Physik, moderne Informationstechnologie und Künstliche Intelligenz

AKPIK 2: Applications in Particle and Astroparticle Physics

AKPIK 2.7: Talk

Tuesday, March 21, 2023, 18:30–18:45, ZEU/0118

Estimation of prediction uncertainties for data from Imaging Atmospheric Cherenkov Telescopes — •Cyrus Pan Walther and Maximilian Linhoff — Technische Universität Dortmund, Germany

One main step in the low-level analysis of astroparticle physics data is the reconstruction of the properties of primary particles that induced extensive air showers.

Various methods are applied in different experiments and software packages. In general, these are multi-output and combined regression and classification tasks. The estimation of prediction uncertainties is of crucial importance for the later scientific exploitation of these events. However, most methods do not in themselves provide reliable uncertainty estimates. In this contribution, we want to apply a method that has been used successfully in a Deep Learning reconstruction for the IceCube experiment to data from Imaging Atmospheric Cherenkov Telescopes used for gamma-ray astronomy.