Searching for gluon quartic gauge couplings at muon colliders using the auto-encoder

TL;DR

This paper explores using an auto-encoder, a machine learning technique, to efficiently identify gluon quartic gauge couplings at muon colliders, demonstrating improved event selection over traditional methods.

Contribution

It introduces an auto-encoder based anomaly detection method for studying dimension-8 operators in SMEFT related to gQGCs at muon colliders, enhancing analysis efficiency.

Findings

AE-based anomaly detection outperforms traditional event selection methods

The approach effectively identifies gQGC signals in collider data

Potential for acceleration via quantum computing

Abstract

One of the difficulties one has to face in the future phenomenological studies of the new physics~(NP), is the need to deal with increasing amounts of data. It is therefore increasingly important to improve the efficiency in the phenomenological study of the NP. Whether it is the use of the Standard Model effective field theory~(SMEFT), the use of machine learning~(ML) algorithms, or the use of quantum computing, all are means of improving the efficiency. In this paper, we use a ML algorithm, the auto-encoder~(AE), to study the dimension-8 operators in the SMEFT which contribute to the gluon quartic gauge couplings~(gQGCs) at muon colliders. The AE is one of the ML algorithms that has the potential to be accelerated by the quantum computing. It is found that the AE-based anomaly detection algorithm can be used as event selection strategy to study the gQGCs at the muon colliders, and is effective compared with traditional event selection strategies.