Nuclear energy density functionals from machine learning

Machine learning is employed to build an energy density functional for self-bound nuclear systems the first time. By kinetic as a of nucleon alone, robust and accurate orbital-free nuclei established. Self-consistent calculations that bypass Kohn-Sham equations provide ground-state densities, total energies, root-mean-square radii with high accuracy in comparison solutions. No existing theory comes close this performance nuclei. Therefore, it provides new promising way future developments functionals whole chart.