A new study has successfully extracted the amplitudes of the deeply virtual Compton scattering (DVCS) process from both the proton and the neutron. This advance was achieved through a global analysis of experimental data, utilizing a novel extraction framework integrated into the PARTONS software ecosystem. DVCS amplitudes are a crucial intermediate quantity that links generalized parton distributions (GPDs) to experimental observables, providing a unique window into the three-dimensional structure of nucleons.

The analysis employed two distinct modeling approaches for the DVCS amplitudes, both relying on machine learning techniques: one model-agnostic and the other theory-augmented. For the first time, the extraction incorporated helicity-flip amplitudes, which are particularly sensitive to higher-twist effects. Furthermore, the extracted amplitudes were compared to timelike Compton scattering (TCS) data using established relations between DVCS and TCS, serving as an important test of GPD universality.

This work also extracted the DVCS subtraction constant, for which ab initio predictions from lattice QCD have been obtained. This marks the first time these theoretical predictions can be directly confronted with experimental DVCS data, initiating a new virtuous cycle between first-principles theory and phenomenology. This progress is fundamental for refining our understanding of the internal structure of nucleons and the dynamics of quarks and gluons within them.