SMuK 2023 – wissenschaftliches Programm

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

T: Fachverband Teilchenphysik

T 34: ML Methods II

T 34.6: Vortrag

Dienstag, 21. März 2023, 18:15–18:30, HSZ/0405

ANN for Pulse Shape Analysis in GERDA — •Vikas Bothe for the GERDA collaboration — Max-Planck-Institute for Nuclear physics, Heidelberg

The GERDA experiment searches for the neutrinoless double-beta decay of 76Ge using enriched high-purity Germanium diodes as a source as well as a detector. For such a rare event search, the sensitivity of the experiment can be improved by employing active background suppression techniques. The time-profile analysis of the signals, called pulse shape analysis (PSA), generated by energy deposits within the detectors is employed to discriminate signal and background events. An effective PSA with artificial neural networks can reject the background events like alpha particles and Compton scattered photons while preserving a high signal efficiency for double beta decay-like events.

Coaxial detectors due to their geometry have significantly homogenous weighting potential adding a spatial dependence to pulse shapes. This makes the signal-background differentiation difficult with the use of simple mono-parametric cuts and to overcome this, we employ a multi-variate analysis with artificial neural networks which are capable of modeling complex relationships.

I will give a review of the methodology in building these ANN and their performance for PSA in GERDA.