Extracting GPDs from experimental data, moments of QCD models and lattice calculations using ML

GPDs are an important tool in understanding the internal structure of nucleons and other hadrons; however, they cannot be accessed directly. The DVCS amplitude encodes information on the Compton Form Factors (CFFs), which are convolutions of the GPDs with a hard perturbative piece. By measuring the DVCS cross sections, we gain access to the CFFs, not the GPDs. Theoretical models for GPDs have been developed that describe CFFs . Alternatively, QCD lattice correlation functions can obtain Gravitational Form Factors (GFFs), which, through Mellin moments can be inverted to give the GPDs. This inversion is non-trivial. Here, we describe an alternative approach to reconstruct the GPDs via sets of orthogonal polynomials - the Bernstein, Jacobi or Chebyshev polynomials. We begin to employ machine learning to incorporate these GPD extraction techniques into a whole.