Machine learning-based identification of boosted objects and search for pair production of heavy vector-like quarks in fully-hadronic final states with the ATLAS detector

Abstract

In this thesis, the data recorded by the ATLAS experiment at the LHC in 2015 and 2016, corresponding to an integrated luminosity of 36.1 fb-1, is analyzed to study the performance of the boosted object identification algorithms and to search for pair-produced heavy vector-like quarks in fully hadronic final states. The first part of the thesis focuses on the use of jet moments as inputs to deep neural networks to build binary jet classifiers which discriminate W-boson or top-quark jets from the gluon and light-quark jet background. The second part of the thesis presents a search for pair-produced vector-like quarks in fully-hadronic final states with small missing transverse momentum. The analysis strategy is optimized assuming that the pair produced vector-like quarks decay into a Standard Model boson and a third-generation quark. No significant deviation from the Standard Model expectation is observed and upper limits are set on the production cross-sections.