Generative modeling of nucleon-nucleon interactions

Developing high-precision nuclear interaction models remains one of the primary challenges in nuclear physics. It is key to a better understanding of theoretical uncertainties for many-body calculations and ab initio theories. Chiral Effective Theory has become the most widely used framework for constructing nuclear potentials due to its ability to generate contributions order-by-order in a well-defined expansion parameter. However, the uncertainties introduced by the dependence of chiral potentials on the resolution scale have not been fully estimated. Determining a chiral potential for a specific resolution scale parameter is computationally expensive. To address this issue, an effective method to generate chiral potentials over a continuous range of resolution scale parameters is needed. We propose the application of Glow model, a generative machine learning model that is capable of capturing salient features of samples, to generate chiral potentials across continuous resolution scale parameter space. Our results show that the potentials generated by a well-trained Glow model in a matter of seconds can reproduce high-quantity nucleon-nucleon scattering phase shifts.