Nuclear Mass Predictions for the Crustal Composition of Neutron Stars: A Bayesian Neural Network Approach

TL;DR

This paper introduces a Bayesian Neural Network approach to refine nuclear mass models, significantly improving their accuracy and providing reliable statistical errors, which enhances predictions of neutron star crust composition.

Contribution

The paper presents a novel BNN-based refinement method that improves existing nuclear mass models by about 40% and includes statistical error estimation, advancing astrophysical nuclear modeling.

Findings

40% improvement in mass prediction accuracy

Inclusion of statistical error estimates

Successful prediction of neutron star crust composition

Abstract

Besides their intrinsic nuclear-structure value, nuclear mass models are essential for astrophysical applications, such as r-process nucleosynthesis and neutron-star structure. To overcome the intrinsic limitations of existing "state-of-the-art" mass models, we propose a refinement based on a Bayesian Neural Network (BNN) formalism. A novel BNN approach is implemented with the goal of optimizing mass residuals between theory and experiment. A significant improvement (of about 40%) in the mass predictions of existing models is obtained after BNN refinement. Moreover, these improved results are now accompanied by proper statistical errors. Finally, by constructing a "world average" of these predictions, a mass model is obtained that is used to predict the composition of the outer crust of a neutron star. The power of the Bayesian neural network method has been successfully demonstrated by a systematic improvement in the accuracy of the predictions of nuclear masses. Extension to other nuclear observables is a natural next step that is currently under investigation.