RIKEN astrophysicists have developed an improved model for the interior structure of neutron stars that agrees well with observations. Unlike previous models, it can be extended to consider what happens when two neutron stars merge.

The collapsed remnants of giant stars, neutron stars are fascinating objects. They are a mere 20−30 kilometers in diameter but are nearly 400,000–600,000 times more massive than the Earth, which makes them incredibly dense.

Neutron stars are not uniform agglomerations of neutrons—like the astrophysical equivalent of a giant atomic nucleus containing only neutrons. Rather they have an onion-like structure. Theorists have been busy trying to model this internal structure based on quantum mechanics and data from observations.

Strong lensing of the galaxy, which can be seen as arc-like features, is a powerful probe of the small-scale DM halos. The populations of small-scale DM halo give us hints about its particle properties. We need to manage huge parameter spaces (e.g. redshift distribution of the source galaxies, lensing galaxies, mass functions of perturbing subhalos and so on) to determine the subhalo signatures from the strong-lensing image data using likelihood ratio test. The machine-learning based techniques of the reduced likelihood ratio estimator enable us to derive the parameters of subhalo mass function, which are key quantities to access the nature of DM, in an efficient way. The importance of this technique increases for the coming era of large-sized lensing image data. In the near future, we should probe the parameters of the subhalo mass function hence the DM properties from galaxy-galaxy lensing. Furthermore, the method is so flexible that encourages us to consider much wider applications in DM search.