Equivariant neural networks for robust $\textit{CP}$ observables

TL;DR

This paper presents a novel application of equivariant neural networks to improve the detection of CP symmetry violations in particle physics experiments, enhancing the accuracy and efficiency of observable construction.

Contribution

It introduces equivariant neural networks tailored for CP violation analysis, demonstrating improved convergence and optimal observable construction over existing methods.

Findings

Enhanced numerical convergence with equivariant models

Construction of observables that better reflect CP symmetry properties

Significant improvement over previous methodologies in CP violation searches

Abstract

We introduce the usage of equivariant neural networks in the search for violations of the charge-parity ($\textit{CP}$) symmetry in particle interactions at the CERN Large Hadron Collider. We design neural networks that take as inputs kinematic information of recorded events and that transform equivariantly under the a symmetry group related to the $\textit{CP}$ transformation. We show that this algorithm allows to define observables reflecting the properties of the $\textit{CP}$ symmetry, showcasing its performance in several reference processes in top quark and electroweak physics. Imposing equivariance as an inductive bias in the algorithm improves the numerical convergence properties with respect to other methods that do not rely on equivariance and allows to construct optimal observables that significantly improve the state-of-the-art methodology in the searches considered.