Heidelberg 2022 – wissenschaftliches Programm

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

T: Fachverband Teilchenphysik

T 88: Top Quarks: Production (Exp.) 3

T 88.8: Vortrag

Donnerstag, 24. März 2022, 18:00–18:15, T-H19

Graph neural network applications in tt+heavy flavor studies at the CMS experiment — •Emanuel Pfeffer, Max Erhart, Ulrich Husemann, Rufa Rafeek, Jan van der Linden, and Michael Waßmer — Institute of Experimental Particle Physics (ETP), Karlsruhe Institute of Technology (KIT)

Graph neural networks are a promising and novel method of artificial intelligence. In contrast to other classes of machine learning, graphs may be better at mapping relationships and dependencies among objects such as jets and at learning interconnections. The application of these networks is particularly interesting in processes consisting a pair of top quarks produced in association with bottom quarks (tt+bb) and other indistinguishable processes such as tt+H with H→bb or tt+Z with Z→bb. Initial studies show promising results in identifying the b jets not originating from the decay of the top quarks in such processes. The graph neural network studies show better results than previous ones based on conventional deep neural networks.
  This talk gives an overview of the applications and results of jet assignments in tt+heavy flavor processes at the CMS experiment using graph neural networks.