Heidelberg 2022 – wissenschaftliches Programm
Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
T: Fachverband Teilchenphysik
T 76: Search for Dark Matter 3
T 76.8: Vortrag
Mittwoch, 23. März 2022, 18:00–18:15, T-H35
Performance of different MET reconstruction methods in a monotop DM analysis — •Jost von den Driesch1, Sebastian Wieland1, Michael Waßmer1, Nikita Shadskiy1, Ulrich Husemann1, Matteo Cremonesi2, Lindsey Gray3, and Yihui Lai4 — 1Institute of Experimental Particle Physics (ETP), Karlsruhe Institute of Technology (KIT) — 2University of Notre Dame (ND) — 3Fermilab — 4University of Maryland (UMD)
Missing transverse momentum (MET) is an important quantity in many analyses at hadron colliders. Especially Dark Matter (DM) analyses often make use of this quantity as DM particles leave the detector without interactions and therefore create large amounts of MET. However, due to its origin from non-detectable particles, MET cannot be measured directly, but must be estimated from the transverse momentum of all reconstructable particles.
Over the years, various MET reconstruction methods have been developed and applied at CMS. The latest approaches use machine learning methods, e.g. Convolutional Neural Networks (DeepMET) or Graph Neural Networks (GraphMET). Monte Carlo studies show an improvement of MET reconstruction performance by these novel reconstruction methods compared to the older ones. Yet, it remains unclear how large this effect will be in a full analysis.
This talk will introduce the aforementioned MET reconstruction methods and compare their expected impact on a monotop analysis, aimed at the search for Dark Matter in events with a single top quark and large MET.