Göttingen 2025 – wissenschaftliches Programm
Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
T: Fachverband Teilchenphysik
T 101: Neutrino Physics VIII
T 101.3: Vortrag
Freitag, 4. April 2025, 09:30–09:45, VG 3.103
Bayesian analysis of KATRIN neutrino mass data using a neural network — Philipp Krönert1, Susanne Mertens2, Oliver Schulz3, and •Alessandro Schwemmer2 for the KATRIN collaboration — 1Helmholtz-Institut für Strahlen- und Kernphysik, Bonn — 2Physik Department, Technische Universität München, Garching — 3Max-Planck-Institut für Physik, München
The Karlsruhe Tritium Neutrino (KATRIN) experiment probes the effective electron anti-neutrino mass by precisely measuring the tritium beta-decay spectrum near its endpoint. A world-leading upper limit of 0.45 eV c−2 (90 % CL) has been set with the first five measurement campaigns following a frequentist analysis procedure. A neural network has been developed in this context, enabling fast and precise model calculations. Utilizing this neural network, a new Bayesian framework has been built in the Julia programming language. It allows for efficient sampling of the posterior density using Hamiltonian Monte Carlo methods implemented by BAT.jl. In this talk, we will present the current development status of the Bayesian framework and its application to the analysis of the first five KATRIN measurement campaigns.
This work is supported by the Helmholtz Association and by the Ministry for Education and Research BMBF (grant numbers 05A23PMA, 05A23PX2, 05A23VK2, and 05A23WO6).
Keywords: neutrino mass; bayesian analysis; neural network; julia; katrin