Karlsruhe 2024 – scientific programme

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

T: Fachverband Teilchenphysik

T 55: Methods in particle physics 4 (HCAL, jets)

T 55.6: Talk

Wednesday, March 6, 2024, 17:15–17:30, Geb. 20.30: 2.066

Improving Hadron Reconstruction in the Belle II Electromagnetic Calorimeter using Graph Neural Networks — •Jonas Eppelt, Isabel Haide, and Torben Ferber — Institute of Experimental Particle Physics (ETP), Karlsruhe Institute of Technology (KIT)

Our aim is to refine hadron reconstruction in the Belle II Electromagnetic Calorimeter (ECL), specifically addressing overlapping clusters. We are using Graph Neural Network (GNN) architectures, such as GravNet, to enhance clustering accuracy. Improving clustering precision holds significant implications for physics analyses, especially in searches for final states that include missing energy like B → K ν ν. These searches will profit from refined selection criteria. This presentation outlines our ongoing efforts to optimize hadron clustering using GNNs, aiming for better precision within the Belle II ECL.

Keywords: Calorimeter; Machine Learning; Graph Neural Networks; Clustering; Reconstruction