MCNNTUNES: tuning Shower Monte Carlo generators with machine learning

TL;DR

This paper introduces a machine learning approach using neural networks to improve the tuning of Shower Monte Carlo generators, addressing the complexity and computational challenges of traditional methods.

Contribution

It presents a novel neural network-based tuning procedure for event generators, offering an alternative to polynomial parametrisation methods.

Findings

Successful implementation tested with ATLAS data

Closure testing confirms the method's validity

Potential for more efficient generator tuning

Abstract

The parameters tuning of event generators is a research topic characterized by complex choices: the generator response to parameter variations is difficult to obtain on a theoretical basis, and numerical methods are hardly tractable due to the long computational times required by generators. Event generator tuning has been tackled by parametrisation-based techniques, with the most successful one being a polynomial parametrisation. In this work, an implementation of tuning procedures based on artificial neural networks is proposed. The implementation was tested with closure testing and experimental measurements from the ATLAS experiment at the Large Hadron Collider.