Master Thesis defense by Magnus Guldbæk Hansen

Title: Search for the Higgs to Z Gamma decay in the ATLAS experiment using new hybrid training approach.

Abstract: The asymmetric Higgs boson decay to a Z boson and a photon has yet to be discovered. This decay is interesting, as it is sensitive to new particles and physics. The main limitation in its discovery is the low signal statistics. This thesis aims to improve the ATLAS Run3 analysis reach by optimising the H → Z (→ ee)γ event selection using an array of Machine Learning models.

Identification and isolation models are made for electrons, which are then combined with additional variables to produce an efficient Z → ee selection model. Various training and testing schemes are used and compared. The performance is also studied in real data to ensure that the models are capturing structures present in the real data. Photon identification and isolation models are produced to be combined with the output of the Z → ee model, thus in the end yielding a model for the selection of these rare Higgs decays.

No improvement in Z → ee selection efficiency (at same back- ground level as the current ATLAS selection) of 2% (YY%) in MC. Finally, it failed to deliver a increased efficiency for H → Zγ. Cou- pled with the increased data size of Run3, this is expected to push the analysis beyond the 3 sigma discovery limit.