Development of systematic uncertainty-aware neural network trainings for binned-likelihood analyses at the LHC

TL;DR

This paper introduces a neural network training method that incorporates systematic uncertainties directly into the training process, improving the precision of Higgs boson production measurements at the LHC.

Contribution

It presents a novel systematic uncertainty-aware training approach for neural networks, extended to multiclass classification, applied to Higgs boson analysis at CMS.

Findings

12% reduction in uncertainty for gluon fusion signal strength

16% reduction in uncertainty for vector boson fusion signal strength

Enhanced neural network performance in systematic uncertainty handling

Abstract

We propose a neural network training method capable of accounting for the effects of systematic variations of the data model in the training process and describe its extension towards neural network multiclass classification. The procedure is evaluated on the realistic case of the measurement of Higgs boson production via gluon fusion and vector boson fusion in the $\tau\tau$ decay channel at the CMS experiment. The neural network output functions are used to infer the signal strengths for inclusive production of Higgs bosons as well as for their production via gluon fusion and vector boson fusion. We observe improvements of 12 and 16% in the uncertainty in the signal strengths for gluon and vector-boson fusion, respectively, compared with a conventional neural network training based on cross-entropy.