Bridging Ab Initio Symmetries and Global Nuclear Masses with Interpretable Neural Networks

Phong Dang, Evander Espinoza, Xiaoliang Wan, Michela Negro, Jerry P. Draayer et al.|June 26, 2026arXiv

Key Takeaway

Physics-informed neural networks that encode fundamental symmetries can match state-of-the-art predictive performance while providing interpretable insights into which symmetry principles actually matter for nuclear binding—showing that Wigner's SU(4) symmetry carries real predictive power beyo...

Summary

This paper uses interpretable neural networks informed by nuclear symmetry principles (Wigner's SU(4) and Elliott's SU(3)) to predict nuclear binding energies across the entire nuclear chart.

architecture

Key Terms

casimir-operator su4-symmetry su3-symmetry physics-informed-neural-network uncertainty-quantification