CERN Accelerating science

Jet production is a fundamental probe of perturbative quantum chromodynamics (pQCD). They play a vital role also in other areas of high energy physics. Jet quenching is arguably one of the most spectacular proofs of the creation of quark-gluon plasma in ultrarelativistic collisions of heavy ions. Nowadays, the rise of novel experimental techniques, including jet substructure observables and the application of machine learning algorithms, are revolutionizing this field of study. New jet tagging capabilities allow for comparative studies between jet flavours. Substructure measurements open doors for direct observation of the effects, entangled into more generic observables. A perfect example is the dead-cone measurement by ALICE. Results shown in this thesis benefit from both of these advances. The first part describes the analysis of the beauty-jet production cross section, measured in pp collisions at $\sqrt{s}$ = 5.02 TeVby the ALICE experiment at the LHC. It is the first application of machine learning for heavy-flavour jet measurements in ALICE. The new method significantly improves tagging efficiency and purity, and shows a good stability over a wide range of these parameters. Results are consistent with the NLO pQCD predictions and the ALICE results obtained with other methods. The second part shows simulation studies for the dead-cone effect measurement for beauty jets in heavy-ion collisions. The study focuses on the removal of distortions introduced by uncorrelated heavy-ion background. The combination of jet reclustering and jet grooming allows for the restoration of the quantitative properties related to the dead-cone effect of jets. Additionally, this thesis highlights some potential issues that may arise during future measurements of this effect, which are not immediately apparent.