Using Machine Learning techniques in phenomenological studies in flavour physics

TL;DR

This paper employs machine learning techniques combined with SHAP values to analyze New Physics effects violating Lepton Flavour Universality in flavour physics, using a global fit with Monte Carlo methods.

Contribution

It introduces the first application of machine learning and SHAP values in the context of phenomenological studies of flavour physics with effective field theory.

Findings

Machine learning effectively identifies relevant observables.

Monte Carlo analysis provides confidence intervals and correlations.

Results highlight the importance of mixing in the first generation.

Abstract

An updated analysis of New Physics violating Lepton Flavour Universality, by using the Standard Model Effective Field Lagrangian with semileptonic dimension six operators at $\Lambda = 1\,\mathrm{TeV}$ is presented. We perform a global fit, by discussing the relevance of the mixing in the first generation. We use for the first time in this context a Montecarlo analysis to extract the confidence intervals and correlations between observables. Our results show that machine learning, made jointly with the SHAP values, constitute a suitable strategy to use in this kind of analysis.