Erlangen 2026 – scientific programme

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

HK: Fachverband Physik der Hadronen und Kerne

HK 27: Poster Session

HK 27.24: Poster

Wednesday, March 18, 2026, 16:15–18:30, Redoutensaal

Fast detector simulation with machine learning for ALICE 3 — •Nils Meurer — Goethe-Universität Frankfurt am Main

High-energy physics experiments rely on simulations to correct measured collision data for detector effects. Performing these simulations causes high demands in computing resources. Machine learning (ML) models are considered a fast alternative for traditional simulation approaches.

In this poster, a conceptual study of ML-based detector response models using simulation data of the future ALICE 3 experiment is presented. The tracking performance and reconstruction efficiency for inclusive charged particles are compared to the full detector simulation and different ML approaches are explored. The current status of the project will be presented.

Supported by BMFTR and the Helmholtz Association.

Keywords: Machine Learning; fast detector simulation; ALICE 3