Measuring the anomalous quartic gauge couplings in the $W^+W^-\to W^+W^-$ process at muon collider using artificial neural networks

TL;DR

This paper explores how artificial neural networks can identify anomalous quartic gauge couplings in vector boson scattering at a 30 TeV muon collider, providing new insights into effective field theories and unitarity bounds.

Contribution

It introduces a machine learning approach to extract specific scattering contributions and constrain dimension-8 operators in complex neutrino-rich final states at a muon collider.

Findings

ANN effectively isolates $WW\to WW$ contributions

Reconstruction of subprocess center of mass energy achieved

Constraints on dimension-8 operators demonstrated

Abstract

The muon collider provides a unique opportunity to study the vector boson scattering processes and dimension-8 operators contributing to anomalous quartic gauge couplings~(aQGCs). Because of the cleaner final state, it is easier to decode subprocess and certain operator couplings at a muon collider. We attempt to identify the anomalous $WWWW$ coupling in the exclusive $WW\to WW$ scattering in this paper. Since one aQGC can be induced by multiple dimension-8 operators, the study of one coupling can help to confine different operators. Meanwhile, singling out the $WW\to WW$ process can help to study the unitarity bounds. The vector boson scattering process corresponding to the anomalous $WWWW$ coupling is $\mu^+\mu^-\to \nu\nu\bar{\nu}\bar{\nu}\ell^+\ell^-$, with four (anti-)neutrinos in the final state, which brings troubles in phenomenological studies. In this paper, the machine learning method is used to tackle this problem. We find that, using the artificial neural network can extract the $W^+W^-\to W^+W^-$ contribution, and is helpful to reconstruct the center of mass energy of the subprocess which is important in the study of the Standard Model effective field theory. The sensitivities and the expected constraints on the dimension-8 operators at the muon collider with $\sqrt{s}=30$ TeV are presented. We demonstrate that the artificial neural networks exhibit great potential in the phenomenological study of processes with multiple neutrinos in the final state.