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

T: Fachverband Teilchenphysik

T 8: Higgs Boson: Decay in Fermions 1

T 8.2: Talk

Monday, March 21, 2022, 16:30–16:45, T-H21

Optimization of di-tau mass reconstruction in the ATLAS experiment using a deep neural network — Klaus Desch, Philip Bechtle, Christian Grefe, Lena Herrmann, and •Ramy Hmaid — Rheinische-Friedrich-Wilhelms-Universität Bonn

A major challenge of identifying Higgs decays to tau leptons is the similarity of Higgs- and Z-Boson decays, due to their similar mass and at least least two unobservable neutrinos in the final state.

We will present a regression neural network that determines the invariant mass of the reconstructed di-tau system and compare its performance with existing solutions for this problem. In addition, the stability of the neural network response under different training conditions will be discussed.