Göttingen 2025 – wissenschaftliches Programm

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

T: Fachverband Teilchenphysik

T 83: Methods in Particle Physics IV (Lepton Reconstruction)

T 83.3: Vortrag

Donnerstag, 3. April 2025, 16:45–17:00, VG 4.101

Photon identification at the CMS experiment using particle flow candidates and individual calorimeter energy deposits — •Caio Daumann and Johannes Erdmann — III. Physikalisches Institut A, RWTH Aachen University

Many physics processes under study at the Large Hadron Collider are characterized by the presence of photons in the final state. Consequently, the performance of photon identification algorithms is crucial for the physics reach of the CMS experiment. Currently, the photon identification algorithm is based on a Boosted Decision Tree that utilizes high-level variables as input, such as shower shapes and isolation variables. Instead of relying on high-level variables, we investigate the performance of a photon classifier trained on low-level quantities, such as individual energy deposits in the calorimeter and particle-flow candidates surrounding the photon, from which high-level information is typically derived. Modern machine learning architectures are well-known for their ability to extract informative features directly from raw training data, often outperforming classifiers based on high-level variables. In this study, we report the performance of a classifier trained using such low-level information.

Keywords: High energy physics; Photons; Classification