Karlsruhe 2024 – scientific programme

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

T: Fachverband Teilchenphysik

T 74: Top physics 3 (single top)

T 74.7: Talk

Wednesday, March 6, 2024, 17:30–17:45, Geb. 30.95: Audimax

Event classification of t-channel single top-quark production in proton-proton collisions at a centre-of-mass energy of 13 TeV with the ATLAS detector using Graph Neural Networks. — •Lukas Kretschmann, Joshua Reidelstürz, Dominic Hirschbühl, and Wolfgang Wagner — Bergische Universität Wuppertal, Wuppertal, Germany

For the differential cross section of single top-quark t-channel production, a high-purity signal region with high statistics in single top t-channel production events and low statistics in background processes is necessary. The definition of the signal region for the total t-channel cross section analysis is used as a starting point for defining the high-purity signal region. An additional cut on the NN distribution produced by the neural network trained for the total cross section analysis is applied to define the high-purity signal region.

To improve the separation between signal and background events, we investigate the use of Graph Neural Networks (GNNs) as an alternative to traditional feed forward networks for constructing a final discriminant. Studies on the separation power and signal over background ratios for various cuts on the output values will be presented using simulated data.

Keywords: Single Top; t-channel; Graph Neural Networks; LHC; ATLAS