Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
HK: Fachverband Physik der Hadronen und Kerne
HK 40: Heavy-Ion Collisions and QCD Phases VI
HK 40.4: Vortrag
Donnerstag, 19. März 2026, 17:00–17:15, PHIL C 601
ML-based direct photon and neutral meson measurement in Pb–Pb collisions at √sNN = 5.02 TeV in the ALICE experiment at LHC — •Abhishek Nath for the ALICE Germany collaboration — Ruprecht Karl University of Heidelberg, Heidelberg, Germany
The ALICE experiment at LHC-CERN aims to analyze the properties of the quark-gluon plasma (QGP) formed during heavy-ion collisions. Neutral meson yields determine RAA and constrain parton energy loss, whereas direct photons from thermal and hard-scattering sources probe the QGP temperature. However, large photon backgrounds from neutral meson decays hinder direct-photon extraction, causing significant loss of precision at low pT in Run 2 heavy-ion Pb–Pb data. To overcome this limitation, we introduce a machine learning-based approach for photon candidate selection within the Photon Conversion Method (PCM). An XGBoost classifier trained on Monte Carlo simulations anchored to the Run 2 Pb–Pb √sNN = 5.02 TeV dataset replaces traditional cut-based selections to provide data samples with simultaneously optimized photon efficiency and purity.
In this talk, I will present the application of this ML-enhanced PCM analysis, showing updated transverse momentum spectra for π0, η, and direct photons. The resulting RAA, η/π0 ratio, and direct photon excess ratio (Rγ) are then compared with the traditional cut-based measurements as well as with state-of-the-art theoretical model predictions.
Keywords: ALICE; heavy-ion collision; XGBoost; neutral meson; direct photon