DeepMET: Improving missing transverse momentum estimation with a deep neural network

Jul 28, 2025, 3:50 PM
20m
Barus and Holley 168 (Brown University)

Barus and Holley 168

Brown University

184 Hope St, Providence, RI 02912
Oral Presentation Plenary Talks

Speaker

Yongbin Feng

Description

Missing transverse momentum is a crucial experimental observable for many analyses of data from detectors at hadron colliders. In the standard model, missing transverse momentum originates from neutrinos. Moreover, beyond the standard model particles like dark matter candidates are also expected to leave the detector undetected. This talk presents a novel missing transverse momentum estimator based on deep neural networks, called “DeepMET”. DeepMET was developed for the CMS experiment at the LHC. It produces a weight for each reconstructed particle in an event. The DeepMET estimator is the negative of the vector sum over all particles of the weighted transverse momenta. The talk presents performance improvements compared to estimators previously employed by CMS and resilience against the effect of additional proton-proton interactions accompanying the collision of interest.

Presentation materials