Active learning BSM parameter spaces

TL;DR

This paper demonstrates that active learning significantly improves the efficiency and accuracy of parameter space scans in beyond Standard Model (BSM) physics models, including MSSM, SMSQQ, and MDGSSM.

Contribution

It introduces active learning as a novel method for parameter scans, enhancing efficiency and precision in complex BSM models.

Findings

Active learning reduces computational effort in parameter scans.

More accurate bounds are obtained for MSSM.

Maximum dark matter singlet mass updated to 48.4 TeV.

Abstract

Active learning (AL) has interesting features for parameter scans of new models. We show on a variety of models that AL scans bring large efficiency gains to the traditionally tedious work of finding boundaries for BSM models. In the MSSM, this approach produces more accurate bounds. In light of our prior publication, we further refine the exploration of the parameter space of the SMSQQ model, and update the maximum mass of a dark matter singlet to 48.4 TeV. Finally we show that this technique is especially useful in more complex models like the MDGSSM.