Reconstruction of event kinematics in semi-inclusive deep-inelastic scattering using the hadronic final state and Machine Learning

TL;DR

This paper presents machine learning methods to improve the reconstruction of event kinematics in semi-inclusive deep-inelastic scattering at the Electron-Ion Collider, enabling more precise mapping of nucleon structure.

Contribution

It introduces novel machine learning techniques that utilize the hadronic final state and scattered electron for more accurate SIDIS kinematic reconstruction.

Findings

Enhanced accuracy in reconstructing virtual photon four momentum.

Improved kinematic resolution across the inclusive DIS coverage.

Potential for more detailed 3D nucleon structure mapping.

Abstract

Semi-inclusive deep-inelastic scattering (SIDIS) at the Electron-Ion Collider will allow for precise mapping of the 3D momentum and spin structure of nucleons and nuclei over a large kinematic region. In this contribution, we demonstrate methods utilizing the hadronic final state and scattered electron, as well as machine learning, to more reliably reconstruct the virtual photon four momentum and SIDIS kinematics across the inclusive DIS coverage at the EIC.