The Neutron Star Outer Crust Equation of State: A Machine Learning approach

Abstract

Constructing the outer crust of the neutron stars requires the knowledge of the Binding Energy (BE) of the atomic nuclei. Although the BE of a lot of the nuclei is experimentally determined and can be obtained from the AME data table, for the others we need to depend on theoretical models. There exist a lot of physical theories to predict the BE, each with its own strengths and weaknesses. In this paper, we apply Machine Learning (ML) algorithms on AME2016 data set to predict the Binding Energy {of atomic nuclei}. The novel feature of our work is that it is model independent. We do not assume or use any nuclear physics model but use only ML algorithms directly on the AME2016 data set. Our results are further refined by using another ML algorithm to train the errors of the first algorithm, and repeating this process iteratively. Our best algorithm gives ${\sigma }_{\mathrm{rms}}\approx 0.58$ MeV for Binding Energy on randomized testing sets. This is comparable to all physics models or ML improved physics models studied in literature till date. Using the predictions of our Machine Learning algorithm, we construct the outer crust equation of state (EoS) of a neutron star and show that our model is comparable to existing models. This work also demonstrates the use of various ML algorithms and a detailed analysis on how we arrived at our best algorithm. It will help the physics community in understanding how to choose an ML algorithm which would be suited for their data set. Our algorithms and best fit model is also made publicly available for the use of the community.

27 pages, 8 figures, github link and neutron star outer crust EoS

References in corpus (14)

• Deep Learning in Neural Networks: An Overview
• Relativistic Continuum Hartree Bogoliubov Theory for Ground State Properties of Exotic Nuclei
• Neutron-Rich Nuclei in Heaven and Earth
• Surface diffuseness correction in global mass formula
• Further explorations of Skyrme-Hartree-Fock-Bogoliubov mass formulas. XII: Stiffness and stability of neutron-star matter
• Nuclear mass predictions based on Bayesian neural network approach with pairing and shell effects
• Further improvements on a global nuclear mass model
• Spin-isospin resonances: A self-consistent covariant description
• Nuclear charge radii: Density functional theory meets Bayesian neural networks
• Impact of the symmetry energy on the outer crust of non-accreting neutron stars
• Decoding Beta-Decay Systematics: A Global Statistical Model for Beta^- Halflives
• Validating neural-network refinements of nuclear mass models
• Improved radial basis function approach with the odd-even corrections
• Deep Learning for Real-time Gravitational Wave Detection and Parameter Estimation with LIGO Data

Cited by in corpus (8)

• Physically Interpretable Machine Learning for nuclear masses
• Decoding Neutron Star Observations: Revealing Composition through Bayesian Neural Networks
• Towards Uncovering Dark Matter Effects on Neutron Star Properties: A Machine Learning Approach
• Nonparametric model for the equations of state of neutron star from deep neural network
• Deep learning inference of the neutron star equation of state
• Realizing the potential of deep neural network for analyzing neutron star observables and dense matter equation of state
• Inferring the Equation of State from Neutron Star Observables via Machine Learning
• Virialized equation of state for warm and dense stellar plasmas in proto-neutron stars and Supernova matter