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

T: Fachverband Teilchenphysik

T 77: Deep Learning III

T 77.4: Vortrag

Donnerstag, 28. März 2019, 16:45–17:00, H06

Investigation of the top-quark mass precision using machine-learning techniques at the ATLAS experiment — •Steffen Ludwig, Andrea Knue, and Gregor Herten — University of Freiburg, Institute of Physics

The top quark is the heaviest known elementary particle in the Standard Model (SM) and its mass is a fundamental parameter. Its value is close to the scale of electroweak symmetry breaking and hence the top quark might serve as a window to physics beyond the SM.
Due to the high collision rate of the LHC, the ATLAS collaboration was able to measure the top-quark mass at subpercent level at √s = 8 TeV. Removing badly reconstructed events has shown to reduce the dominant signal modelling uncertainties using tt events in the lepton + jets channel.
Exploring this decay channel using pp collision data at √s = 13 TeV, the talk focuses on the influence of deep neural networks in comparison to boosted decision trees on the event reconstruction and selection purity, while studying the impact on the total systematic uncertainty of the top-quark mass.