Göttingen 2025 – wissenschaftliches Programm
Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
AKPIK: Arbeitskreis Physik, moderne Informationstechnologie und Künstliche Intelligenz
AKPIK 2: AKPIK Poster Session
AKPIK 2.1: Poster
Mittwoch, 2. April 2025, 16:15–18:15, ZHG Foyer 1. OG
Exploring GNN-based trigger algorithms for underwater neutrino telescopes — •Avalon Rego1,2, Francesca Capel2, Christian Spannfellner3, and Li Ruohan3 — 1Ludwig-Maximilians-Universität, München, Deutchland — 2Max-Planck-Institut für Physik, Garching bei München, Deutchland — 3Technical University of Munich , Munich, Germany
Neutrinos are a window into a deeper understanding of both beyond standard model physics and various high-energy astrophysical phenomena. This is because they can easily escape dense environments due to their weakly interacting nature and can pinpoint their sources since they are not deflected by magnetic fields. We detect these weakly interacting particles by embedding detectors into massive volumes of naturally available water or ice and then detecting the Cherenkov radiation produced by their interactions. These detectors are sensitive to complex backgrounds such as bioluminescence signals which are a challenge for standard trigger algorithms. In this work we investigate the use of Graphnet, a GNN-based python framework, for signal classification and improving discrimination for bioluminescence signals in particular comparing it to a standard coincidence trigger. We also explore the possibility of using this trigger to lower the energy threshold for neutrino detection.
Keywords: Machine Learning; Graph Neural Networks; Neutrinos; IceCube; P-ONE