Nonparametric Representation of Neutron Star Equation of State Using Variational Autoencoder

Abstract We introduce a new nonparametric representation of the neutron star (NS) equation state (EOS) by using variational autoencoder (VAE). As deep neural network, VAE is frequently used for dimensionality reduction since it can compress input data to low-dimensional latent space encoder component and then reconstruct decoder component. Once trained, one take as generator. employ 100,000 EOSs that are generated method based on Han et al. training set try different settings we get an EOS generator (the trained VAE’s decoder) with four parameters. use mass–tidal-deformability binary NS merger event GW170817, mass–radius PSR J0030+0451, J0740+6620, J0437-4715, 4U 1702-429, nuclear constraints perform Bayesian inference. The overall results analysis includes all observations R 1.4 = 12.59 − 0.42 + 0.36 km , ${{\rm{\Lambda }}}_{1.4}={489}_{-110}^{+114}$?> Λ 489 110 114 ${M}_{\max }={2.20}_{-0.19}^{+0.37}\,{M}_{\odot }$?> M max 2.20 0.19 0.37 ⊙ (90% credible levels), where R 1.4 /Λ radius/tidal deformability canonical M ⊙ NS, maximum mass nonrotating NS. indicate implementation these techniques obtain reasonable results, while accelerating calculation factor ∼3–10 or more, compared original method.