Karlsruhe 2024 – wissenschaftliches Programm

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

T: Fachverband Teilchenphysik

T 43: Data, AI, Computing 3 (pointclouds & graphs)

T 43.7: Vortrag

Dienstag, 5. März 2024, 17:30–17:45, Geb. 30.33: MTI

Improvement of GNN energy regression for KM3Net/ORCA with weighted training samples — •Bastian Setter for the ANTARES-KM3NET-ERLANGEN collaboration — Erlangen Centre for Astroparticle Physics (ECAP), Friedrich-Alexander-Universität Erlangen-Nürnberg

KM3NeT/ORCA is a water Cherenkov detector currently under construction in the Mediterranean Sea, near the coast of France. It specializes in the detection of atmospheric neutrinos in the GeV range. It will be used for many different types of analysis such as the determination of the neutrino mass ordering, constraining the elements of the PMNS matrix or Lorentz invariance violation. For each of these analyses a good resolution in energy regression is important. This talk will present the impact of so-called weighted data-sets in the training of Graph Neural Networks for an early stage of KM3NeT/ORCA with 6 detection units. In addition, it will discuss the increase in performance that could be achieved compared to training strategies without data-set optimisation and to traditional reconstructions using maximum-likelihood estimator techniques.

Keywords: Machine Learing; Graph Neural Network; KM3NeT; Weighted training samples; Neutrino telescope