Speaker
Description
The precise identification and momentum measurement of muons are crucial for the ATLAS physics programme, enabling a wide range of precision measurements and searches for new phenomena. This contribution presents the muon reconstruction performance in Run 3 using proton–proton collision data recorded by the ATLAS detector at the LHC.
Measurements of muon reconstruction, identification and isolation efficiencies are reported, together with studies of the muon momentum scale and resolution. Resonant dimuon decays of J/Psi and Z boson are exploited to calibrate the muon momentum and to validate the modelling of reconstruction and identification efficiencies over a wide kinematic range. The performance of established muon isolation working points is reviewed and compared with PLIT (Prompt Lepton Isolation Tagger), a novel machine-learning-based isolation algorithm developed for Run-3 data. Using a transformer encoder architecture and track-level information within a cone around the lepton, PLIT enhances the separation between prompt and non-prompt leptons while improving robustness against pile-up effects.
The impact of the upgraded detector and reconstruction on key physics observables is also discussed.