Deep Neural Networks (DNNs) are a powerful and flexible tool for information extraction and modeling. In this study, we use DNNs to extract the Sivers functions by globally fitting Semi- Inclusive Deep Inelastic Scattering (SIDIS) and Drell-Yan (DY) data. To make predictions of this Transverse Momentum-dependent Distribution (TMD), we construct a minimally biased model using data from COMPASS and HERMES. The resulting Sivers function model, constructed using SIDIS data, is also used to make predictions for DY kinematics specific to the valence and sea quarks, with careful consideration given to experimental errors, data sparsity, and complexity of phase space.
@article{arxiv.2304.14328,
title = {Extraction of the Sivers function with deep neural networks},
author = {I. P. Fernando and D. Keller},
journal= {arXiv preprint arXiv:2304.14328},
year = {2023}
}