Machine learning has been used for many years to identify the origin of a particle, e.g. does the jet result from the hadronic decay of a W boson (W-jet) or from the emission of a hard gluon? For this purpose, the substructure of the jets is analysed in order to distinguish the signal from the background. A major distinguishing feature is the mass of the jets. A W-jet has approximately the mass of the W boson, whereas a jet initiated by a gluon has a very low mass. Distinguishing W-jets from Z-jets (relevant for some searches for new physics) is experimentally very challenging, since W and Z bosons differ mainly by their mass and charge but barely by their substructure. In addition, the experimental resolution of the calorimeter is not good enough to distinguish the small mass difference. In recent years, the reconstruction of jets and the development of machine learning techniques have made great advances such that improvements in the differentiation between W-jet and Z-jet can be expected.

In this project, a new identification algorithm should be developed using machine learning to distinguish W-jets from Z-jets.

Back