Bonn 2020 – scientific programme

The DPG Spring Meeting in Bonn had to be cancelled! Read more ...

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

T: Fachverband Teilchenphysik

T 47: Neural networks and systematic uncertainties

T 47.9: Talk

Wednesday, April 1, 2020, 18:30–18:45, H-HS IV

Multivariate analysis methods in the analysis of single top-quark production in association with a heavy boson at ATLAS — •Christian Kirfel, Ian Brock, Richard Baumann, and Piet Nogga — Physikalisches Institut Bonn

Single top-quark production in association with a heavy boson gives rise to a multitude of interesting analyses including the production of a top quark and a Higgs boson. To separate the signal from background events, multivariate analysis methods are a common choice.

Previous analyses of a top quark associated with a Z boson featured a shallow neural network provided by the NeuroBayes package. Since the long-term support for this program is unclear, efforts are being made to create an algorithm using the open source software Keras to replace NeuroBayes.

A summary of approaches is presented, ranging from a performance comparison between the formerly used NeuroBayes algorithm and a deep neural network built in Keras, to the use of Lorentz-invariant variables. The different approaches are introduced and the results are discussed not only in the context of replacing the NeuroBayes network but also focusing on the general usability of the methods in particle physics analyses.