Inferring the Equation of State from Neutron Star Observables via Machine Learning
N. K. Patra, Tuhin Malik, Helena Pais, Kai Zhou, B. K. Agrawal, Constança Providência · 2025
Read Inferring the Equation of State from Neutron Star Observables via Machine Learning by N. K. Patra, Tuhin Malik, Helena Pais, Kai Zhou, B. K. Agrawal, Constança Providência (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2502-20226.txt.
Introduction
Patra and five co-authors use machine learning to read neutron star interiors in a paper from two thousand twenty-five. The equation of state of ultra-dense matter is a famous unknown. Telescopes, though, measure masses, radii, and tidal squeezing of neutron stars.
Symbolic regression across many candidate equations found tight links between those traits and the hidden physics. A star's maximum mass tracks pressure and density at one key energy level. Within the common mass range, the equation of state can be written from radius and tidal deformability alone. The authors offer this as a fast decoding frame: precise star measurements in, dense matter physics out. Models with exotic particles were included in the search.
Catalog record
| CREATOR | N. K. Patra, Tuhin Malik, Helena Pais, Kai Zhou, B. K. Agrawal, Constança Providência |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 32 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2502-20226
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Inferring the Equation of State from Neutron Star Observables via Machine Learning N. K. Patra, Tuhin Malik, Helena Pais, Kai Zhou, B. K. Agrawal, Constança Providência Abstract We have conducted an extensive study using a diverse set of equations of state (EoSs) to uncover strong relationships between neutron star (NS) observables and the underlying EoS parameters using symbolic regression method. These EoS models, derived from a mix of agnostic and physics-based approaches, considered neutron stars composed of nucleons, hyperons, and other exotic degrees of freedom in beta equilibrium. The maximum mass of a NS is found to be strongly correlated with the pressure and baryon density at an energy density of approximately 800 MeV.fm -3 . We have also demonstrated that the EoS can be expressed as a function of radius and tidal deformability within the NS mass range 1-2 M ⊙ M_{\odot} . These insights offer a promising and efficient framework to decode the dense matter EoS directly from the accurate knowledge of NS observables. Journal: Physics Letters B 1 Introduction Neutron stars (NSs) are one of the most fascinating and mysterious objects in the universe [ 1 , 2 , 3 ] . They provide a unique window into the physics of extremely dense matter. These remnants of supernova explosions are incredibly dense and compact, having mass upto about twice that of the Sun with the radii \sim 10 km [ 4 , 5 , 6 , 7 ] . The astrophysical observations of NS properties are crucial, as they offer a way to constrain the equation of state (EoS) of dense matter [ 8 , 9 , 10 , 11 , 12 , 13 , 14 ] . The EoS can be expressed as a relation between the pressure and energy density of NS matter. It is a critical input for understanding many astrophysical phenomena, such as supernovae and neutron star mergers [ 15 , 16 ] . The time evolution of supernova gravitational waves strongly depends on the EoS, as demonstrated in Ref. [ 17 ] . Recent advances in observational technology have provided new insights into NS properties, driving intense scientific interest and activity [ 18 , 19 , 20 , 21 , 22 , 23 , 24 ] . Recycled millisecond pulsars undergoing starquakes show a sudden increase in gravitational wave amplitude, providing a unique signature of NS matter at high densities [ 25 , 26 , 27 , 28 ] . These breakthroughs have inspired innovative research spanning multiple disciplines, leveraging collaborative and interdisciplinary approaches development. In particular, the intricate direct connections between EoS parameters and NS observables have become a focal point of study. Sophisticated computational techniques, such as machine learning, are increasingly employed to uncover hidden correlations within complex datasets, enabling researchers to explore relationships among EoS and NS properties with unprecedented precision [ 29 , 30 ] . A number of machine learning (ML) techniques, particularly neural networks, have been applied in recent years to unravel the physics of neutron stars [ 31 , 32 , 12 , 33 ] . An analysis of simulated observations of the radius and tidal deformability of NSs has been used to determine the internal composition of NSs using Bayesian Neural Networks (BNNs) [ 34 ] . Using these models, the dense matter equation of state has also been inferred from various astrophysical and gravitational wave measurements [ 35 ] . In a model-independent approach, ML algorithms have been used to predict the binding energy of atomic nuclei from AME2016 datasets, yielding an outer crust EoS for nuclear matter [ 36 ] . The equation of state and non-monotonic behaviour of speed of sound has also been studied using ML in NS [ 37 ] . As well as traditional machine learning models, feed-forward artificial neural networks were used to extrapolate the separation energy of hypernuclei [ 38 ] . It has also been shown that neural networks can handle complex non-linear relationships between observables and the underlying physics with respect to dense matter EoS reconstructions based on mass-radius (M-R) observations and other parameters of NSs [ 39 , 40 ] . By utilizing the non-linear potential of neural networks, this innovative approach creates intricate mappings between observations (input) and the EoS table (output) [ 41 ] . As a result of its application to NS physics, ML has not only provided insights [ 42 , 43 ] on par with or even exceeding traditional physics-based models but has also demonstrated a significant potential to reduce computational costs. In addition to these methods, symbolic regression can provide a powerful alternative to uncover hidden correlations among different features in the datasets. Symbolic regression method (SRM), in particular, has emerged as a potent technique for identifying human-readable equations or relations hidden within datasets [ 44 , 45 , 46 , 47 ] . By utilizing the symbolic regression method, scientists can formulate comprehensible mathematical relationships, enhancing our understanding of the underlying physics of dense neutron star matter. Recent studies in physics have effectively leveraged symbolic regression techniques to derive analytical formulas and uncover underlying mathematical expressions from complex data [ 48 ] . From deriving particle kinematics in collider phenomenology to improving success rates in finding symbolic expressions related to physics problems, the symbolic regression method has demonstrated its power and adaptability [ 49 ] . Additionally, it has been instrumental in modelling astrophysical phenomena such as assembly bias, as well as rediscovering known physical laws such as the Gell-Mann–Okubo formula [ 50 ] . These applications exemplify the increasing collaboration between physics and machine learning and the great potential of the symbolic regression method to advance our understanding of the universe [ 51 , 52 ] . This paper investigates the potential of machine learning, particularly the symbolic regression method, to uncover the hidden relationships between neutron star properties and the equation of state. For that purpose, we use a variety of neutron star equations of state, encompassing different compositions and several modelling frameworks, including meta-models. The EoS collection includes non-linear (NL) nucleonic and hyperonic (NL-hyp) models, models taken from the CompOSE database (CompOSE), models constructed from the extrapolation of the speed of sound (CSE), piecewise-polytropic (PWP) models, density-dependent models obtained from Bayesian inference (DDB), and models from a spectral representation (SR), providing a comprehensive foundation for our investigation. Exploring the relationships between various NS properties through correlation analysis offers a compelling approach to understand the dense matter EoS [ 53 , 54 , 55 , 56 , 57 , 58 , 59 , 60 , 61 , 62 , 63 , 64 , 65 ] . Specifically, in this work, we applied a symbolic regression method to a wide-ranging dataset to investigate: (i) the relation between the NS maximum mass and EoS properties at different energy densities, and (ii) the interplay between neutron star properties and the central energy density, and pressure for the neutron star masses in the range of 1-2 M_{\odot} . The paper is organized as follows. In Section 2 , we provide a concise overview of the formalism used in the present study. The comprehensive analysis of the results and the subsequent discussions are presented in Section 3 . Finally,