Progress in the Neural Network Determination of Polarized Parton Distributions

TL;DR

This paper reviews recent advances in determining polarized parton distributions using the NNPDF methodology, highlighting improved uncertainty estimates and computational techniques for polarized deep-inelastic scattering data analysis.

Contribution

It introduces the application of the NNPDF approach to polarized PDFs, including the FastKernel method for efficient DGLAP evolution and discusses physical constraints and uncertainty estimations.

Findings

Uncertainty on polarized PDFs, especially gluons, has been previously underestimated.

FastKernel method provides a fast, accurate solution for polarized DGLAP equations.

Preliminary results indicate larger uncertainties in polarized gluon distributions.

Abstract

We review recent progress towards a determination of a set of polarized parton distributions from a global set of deep-inelastic scattering data based on the NNPDF methodology, in analogy with the unpolarized case. This method is designed to provide a faithful and statistically sound representation of parton distributions and their uncertainties. We show how the FastKernel method provides a fast and accurate method for solving the polarized DGLAP equations. We discuss the polarized PDF parametrizations and the physical constraints which can be imposed. Preliminary results suggest that the uncertainty on polarized PDFs, most notably the gluon, has been underestimated in previous studies.