Gießen 2024 – wissenschaftliches Programm

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

HK: Fachverband Physik der Hadronen und Kerne

HK 56: Heavy-Ion Collisions and QCD Phases XIV

HK 56.4: Vortrag

Mittwoch, 13. März 2024, 18:15–18:30, HBR 62: EG 05

Development of a ML algorithm for neutral meson and photon reconstruction using PCM in ALICE — •Abhishek Nath for the ALICE Germany collaboration — Ruprecht Karl University of Heidelberg, Germany

Direct photons are unique probes to study and characterize the quark-gluon plasma (QGP) as they leave the medium unscattered. They are produced throughout all stages of the collision. Thus, they carry information about the space-time evolution and the temperature of the medium. However, they are present amidst a large background of mostly decay photons. So a precise estimate of decay photons is necessary. The Photon Conversion Method (PCM) is a great tool to identify photons, especially at low transverse momentum as they result in oppositely charged track pairs when they interact with detector materials.

Armed with the current machine learning algorithms, we try to reconstruct photons and their source mesons in heavy ion collision using PCM. The aim is to have an efficient estimate of the mesons along with photon samples with high purity and compare both with the current standardized cuts-based method implemented in the PCM analysis workflow. Our analysis is based on 2018 Pb-Pb data where we aim to explore various algorithms (XGBoost and others) to classify photons on-fly. Based on the analysis, a roadmap for analyzing high luminosity run 3 data is stated at the end.

Keywords: ALICE; Photon conversion; Machine learning