Speaker
Ethan Lewis Simpson
(The University of Manchester (GB))
Description
Precision measurements in top-quark physics require accurate kinematic reconstruction, definition of interesting observables, and performant signal-to-background separation. This talk will introduce a supervised learning method for generic event reconstruction, highlighting its application to top-quark physics specifically in channels with multiple neutrinos. The approach will be contrasted with physics-inspired generative regression methods for specific top measurements, including quantum tomography. Additionally, we will explore model applicability in Beyond Standard Model (BSM) scenarios, Monte Carlo generator dependencies, as well as the role of reconstruction as an auxiliary task in signal-background discrimination.
Author
Ethan Lewis Simpson
(The University of Manchester (GB))