A generative adversarial network as the basis for a semi-inclusive deep inelastic lepton scattering generator on a polarized proton

Abstract

A neural network, that allows someone to obtain results for semi-inclusive deep inelastic scattering of charged leptons on polarized protons, with the production of pions or strange K mesons, has been developed in this study. The research covered both transverse and longitudinal polarizations of the proton. A range of initial energies of colliding particles was chosen from 20 to 100 GeV in a central mass system. The range is typical for electron-ion colliders currently being designed. It has been shown that it is possible to predict the physical characteristics of the final lepton and hadron with high accuracy as well as different variants of proton polarization using the proposed neural network.

Keywords

• semi-inclusive deep inelastic scattering
• asymmetries
• machine learning
• neural network
• generative-adversarial network