Unveiling the nuclear matter EoS from neutron star properties: a supervised machine learning approach

TL;DR

This paper employs supervised machine learning to accurately map neutron star observables to nuclear matter equations of state, enabling improved understanding of nuclear physics through astrophysical data.

Contribution

It introduces a machine learning framework that captures complex non-linear relationships between neutron star properties and nuclear matter parameters, surpassing previous linear approximation methods.

Findings

High-accuracy non-linear maps between NS observables and EoS

Identification of key nuclear matter parameters affecting NS properties

Predictions of star constraints based on learned models

Abstract

We explore supervised machine learning methods in extracting the non-linear maps between neutron stars (NS) observables and the equation of state (EoS) of nuclear matter. Using a Taylor expansion around saturation density, we have generated a set of model independent EoS describing stellar matter constrained by nuclear matter parameters that are thermodynamically consistent, causal, and consistent with astrophysical observations. From this set, the full non-linear dependencies of the NS tidal deformability and radius on the nuclear matter parameters were learned using two distinct machine learning methods. Due to the high accuracy of the learned non-linear maps, we were able to analyze the impact of each nuclear matter parameter on the NS observables, identify dependencies on the EoS properties beyond linear correlations and predict which stars allow us to draw strong constraints.