Speaker
Juan Montoya
(Universidad de Antioquia's estudent)
Description
We explore quantum machine learning (QML) for b-tagging at pT < 20 GeV, where taggeing b-jets from light/c-jets is challenging. We propose classifiers based on variational quantum circuits (VQCs) implemented in PennyLane, compared against classical ML techniques with equivalent parameter counts. The pipeline shares preprocessing with classical methods for fair comparison. We evaluate generalization, sample complexity, AUC. This study aims to identify conditions where QML offers advantages for challenging HEP tasks in the low-pT regime.
Authors
Jose Ruiz
(Universidad de Antioquia (CO))
Juan Montoya
(Universidad de Antioquia's estudent)
Sebastian Duque Mesa
Tomas Sosa Giraldo
(Universidad de Antioquia)
