Constraining Nuclear Mass Models Using r-process Observables with Multi-objective Optimization

TL;DR

This paper presents a multi-objective optimization method using Pareto Fronts to improve nuclear mass models, enhancing their ability to predict r-process abundances consistent with observational data.

Contribution

It introduces a novel multi-objective optimization approach that leverages r-process observables to select more reliable machine learning nuclear mass models.

Findings

The Pareto Front algorithm effectively identifies models matching Solar and stellar r-process data.

The approach improves extrapolation accuracy for neutron-rich nuclear masses.

Results demonstrate enhanced predictive power of selected models.

Abstract

Predicting nuclear masses is a longstanding challenge. One path forward is machine learning (ML) which trains on experimental data, but can suffer large errors when extrapolating toward neutron-rich species. In nature, such masses shape observables for the rapid neutron capture process (r-process), which in principle could inform ML models. Here we introduce a multi-objective optimization approach using the Pareto Front algorithm. We show that this technique, capable of identifying models which generate r-process abundances aligning with both Solar and stellar data, is a promising method to select ML models with reliable extrapolation power.