EPJ Web of Conferences 276, 06007 (2023)
https://doi.org/10.1051/epjconf/202327606007

A physics-based neural network reconstruction of the dense matter equation of state from neutron star observables

¹ Frankfurt Institute for Advanced Studies (FIAS), D-60438 Frankfurt am Main, Germany
² Institut für Theoretische Physik, Goethe Universität, D-60438 Frankfurt am Main, Germany
³ Xidian-FIAS International Joint Research Center, D-60438 Frankfurt am Main, Germany
⁴ Center for Nuclear Theory, Department of Physics and Astronomy, Stony Brook University, Stony Brook, New York 11794, USA
⁵ GSI Helmholtzzentrum für Schwerionenforschung GmbH, D-64291 Darmstadt, Germany

^* Presenter, e-mail: soma@fias.uni-frankfurt.de
^** e-mail: zhou@fias.uni-frankfurt.de

Published online: 1 March 2023

Abstract

We introduce a novel technique that utilizes a physics-driven deep learning method to reconstruct the dense matter equation of state from neutron star observables, particularly the masses and radii. The proposed framework involves two neural networks: one to optimize the EoS using Automatic Differentiation in the unsupervised learning scheme; and a pre-trained network to solve the Tolman–Oppenheimer–Volkoff (TOV) equations. The gradient-based optimization process incorporates a Bayesian picture into the proposed framework. The reconstructed EoS is proven to be consistent with the results from conventional methods. Furthermore, the resulting tidal deformation is in agreement with the limits obtained from the gravitational wave event, GW170817.