Merged $\pi ^{0} \to 2\gamma$ identification in LHCb for Run 4 using machine learning

The identification of high-energy neutral pions whose decay products (two photons) merge into a single signal cluster in the electromagnetic calorimeter is a problem in event reconstruction at the LHCb experiment, since such clusters can easily be misinterpreted as other types of clusters, e.g., those of single photons. This summer student project developed and investigated a possible classification of these two cluster types (merged pions and single photons) using machine learning. This, in turn, may help improve current studies of CP-violating decays as well as open up new possibilities in physics analysis at LHCb.