Göttingen 2025 – wissenschaftliches Programm

Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

T: Fachverband Teilchenphysik

T 54: Data, AI, Computing, Electronics V (Anomaly Detection, Event Selection)

T 54.6: Vortrag

Mittwoch, 2. April 2025, 17:30–17:45, VG 2.101

Sterile Neutrino Search with Neural Networks at KATRIN — •Luca Fallböhmer for the KATRIN collaboration — Technical University Munich

The KATRIN experiment aims to search for keV sterile neutrinos in the full beta decay spectrum of tritium using the TRISTAN detector and DAQ system after the end of the neutrino mass measurement. Thanks to the high source activity of KATRIN, a sterile neutrino signature can be probed down to the parts per million level. Because the modelling of the deep differential tritium spectrum is very complex and the involved Monte Carlo simulations require long computing times to reach the necessary statistics, a fit of the sterile parameters is very challenging with the current model. Thus, neural networks are used to search directly for the sterile neutrino signature. In this talk, we demonstrate the sensitivity of the neural network method to the sterile neutrino signature. Additionally, we discuss the robustness of the neural network approach in the presence of experimental effects, their uncertainties, and modelling inaccuracies.

Keywords: Neural Networks; KATRIN; Sterile Neutrinos; TRISTAN