End-to-End Learning Between Compton Form Factors and Mellin Moments

In this study, we use machine learning algorithms to explore the relationship between Mellin moments and Compton Form Factors (CFFs). We develop an end-to-end deep learning approach aimed at extracting Mellin moments directly from CFFs or kinematics without the need for Generalized Parton Distributions (GPD) calculations. The CFF variables are related to Deeply Virtual Compton Scattering (DVCS) experimental observables. Our analysis employs a deep learning method to approximate Mellin moments from CFFs without the intermediate step of integrating GPDs, as well as an inverse model for predicting CFFs from Mellin moments. These machine learning algorithms are systematically evaluated using simulated data that approximates experimental ones.