Erlangen 2026 – wissenschaftliches Programm

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

T: Fachverband Teilchenphysik

T 5: Methods in Particle Physics I

T 5.2: Vortrag

Montag, 16. März 2026, 16:30–16:45, KH 00.020

Validation of the Graph Neural Network tracking at Belle II — •Jonas Lotz and Slavomira Stefkova — Physikalisches Institut der Rheinischen Friedrich-Wilhelms-Universität Bonn, Bonn, Germany

A novel end-to-end multi-track reconstruction algorithm based on a graph neural network (GNN) architecture was developed for the Belle II collaboration (Comput Softw Big Sci 9, 6 (2025)). These studies demonstrate substantial improvements in key performance metrics such as tracking efficiency, with particularly strong gains for tracks originating from displaced vertices. In addition, different hit-to-track association provides an opportunity for improved event clean-up, which may be particularly useful for missing energy decays. In this presentation, building on this development, the applicability of this GNN-based tracking algorithm in the context of searches for B⁰ → K_s⁰ ν ν decays will be studied. We will focus on validating the algorithm against established tracking methods and assessing potential benefits for rare decay searches. The talk will present the status of integrating the new GNN tracking framework into the B⁰ → K_s⁰ ν ν analysis.

Keywords: Graph Neural Network; Tracking; Belle II; Rare Decay Search