Erlangen 2026 – wissenschaftliches Programm

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

T: Fachverband Teilchenphysik

T 93: Data, AI, Computing, Electronics VIII

T 93.2: Vortrag

Freitag, 20. März 2026, 09:15–09:30, KH 00.024

Speeding up the MC Background Simulation at Belle II — •Oliver Schumann, Nikolai Krug, Thomas Kuhr, and Thomas Lück — Ludwig-Maximilians-Universität München (LMU), München, Germany

By striving for ever-higher luminosities, the Belle II detector is set to observe rare decay signals. However, these high luminosities correspond to an increased demand for MC-generated background. Therefore, an efficient algorithm to simulate background events for the detector in large quantities is vital for the successful interpretation of Belle II’s data. While the generation and skimming of events in the MC simulation chain are quick and easy to compute, the detector simulation and reconstruction of these particles are a slow task. A neural network (NN) is introduced into the chain to classify skim-passing events and discard those which fail to pass the NN before the detector response simulation, thereby saving valuable computing resources and time.

This work explores a new approach for parallelisation of the NN training process, in hopes of achieving convergence with the resources available in less (real) time. It involves an on-the-fly training pipeline on multiple machines, retrieving the updated NN, and redistributing it to the clients. This talk aims to provide an overview of the current work in progress and to give an outlook on future prospects.

Keywords: MC Simulation; Background; Transformer; Deep Learning; Belle II