Heidelberg 2022 – scientific programme

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

T: Fachverband Teilchenphysik

T 70: Experimental Methods (general) 3

T 70.7: Talk

Wednesday, March 23, 2022, 17:45–18:00, T-H29

A machine-learning based method to improve isolation variables for photon identification with the ATLAS detector — Johannes Erdmann, Olaf Nackenhorst, and •Michael Windau — TU Dortmund University, Department of Physics

The study of photons is crucial for finding and measuring many processes at colliders. Predominantly, prompt photons, which are created during the collisions, play an important role and have to be distinguished from hadrons decaying into photons. Different methods are used to distinguish this signal from the background. One of these is the use of isolation variables. These are based on track measurements and information from the calorimeters, where they are defined by the activity in a cone around the candidate object. They are currently built in ATLAS by discrete cuts.

In this talk, studies on improving isolation variables using deep neural networks will be presented.