Karlsruhe 2024 – scientific programme

Parts | Days | Selection | Search | Updates | Downloads | Help

T: Fachverband Teilchenphysik

T 16: Data, AI, Computing 1 (anomaly detection)

T 16.8: Talk

Monday, March 4, 2024, 17:45–18:00, Geb. 30.33: MTI

ANNs for enhanced Pulse Shape Discrimination in GERDA — •Vikas Bothe — Max-Planck-Institute for Nuclear Physics, Heidelberg

The GERDA experiment searches for the rare neutrinoless double-beta decay of 76Ge using enriched high-purity Germanium diodes as a source as well as a detector. The experimental sensitivity can be improved significantly by employing active background suppression techniques such as Pulse Shape Discrimination(PSD) based on the analysis of time-profile of signals.

The unique challenge arises from coaxial detectors showcasing spatial dependence of pulse shapes which makes traditional mono-parametric PSD techniques ineffective. To address this, we implement artificial neural networks (ANNs) in a multivariate analysis, leveraging their capacity to model complex patterns. This work presents advancements in ANN based PSD within the GERDA experiment to effectively reject background events, such as alpha particles and Compton scattered photons, while maintaining high signal efficiency for double beta decay-like events.

I will give a brief review of the development of ANNs for PSD in GERDA, highlighting the exploration of various machine learning models and diverse approaches to input feature manipulation to achieve improved PSD performance.

Keywords: Neutrinoless double-beta decay; Pulse Shape Analysis; Artificial Neural Networks; Germanium detectors; Background mitigation