Speaker
Andrzej Novak
(RWTH Aachen (DE))
Description
A review of the heavy flavour identification methods and the performance results of the calibration of various taggers in Run 2 data at CMS is presented. The Machine Learning methods play an important role in the development of the identification algorithms and significantly improve the performance of heavy flavour jet tagging in the offline event reconstruction, as well as in the online event selection. An essential gain is also observed in the performance of the methods used in the identification of the decays of the lorentz-boosted objects resulting in reconstructed jets containing multiple hadrons. The outlook and projections of heavy flavour tagging performance at HL-LHC are also discussed.
Primary author
Arnd Meyer
(Rheinisch Westfaelische Tech. Hoch. (DE))
