Neural-Network Extraction of Unpolarized Transverse-Momentum-Dependent Distributions.

Area of Science:

• High Energy Physics
• Quantum Chromodynamics
• Particle Physics

Background:

• Transverse-momentum-dependent (TMD) distributions describe the motion of quarks inside hadrons.
• Extracting TMDs is crucial for understanding the partonic structure of matter.
• Drell-Yan processes provide valuable experimental data for TMD studies.

Purpose of the Study:

• To perform the first extraction of unpolarized quark TMDs using experimental Drell-Yan data.
• To investigate the efficacy of neural networks in parametrizing the nonperturbative component of TMDs.
• To establish the feasibility of machine learning techniques for probing hadron structure.

Main Methods:

• Utilized experimental Drell-Yan data.
• Employed neural networks for the nonperturbative parametrization of TMDs.
• Compared neural network performance against traditional parametrization methods.

Main Results:

• Successfully extracted unpolarized quark TMDs.
• Demonstrated that neural networks provide a more accurate description of the experimental data than traditional methods.
• Confirmed the superior performance of neural networks in modeling TMDs.

Conclusions:

• Neural networks are a feasible and effective tool for TMD extraction.
• Machine learning techniques enable more accurate determinations of the multidimensional partonic structure of hadrons.
• This work opens new avenues for precision studies in particle physics using AI.