A Partially Supervised Latent-Space Variational Autoencoder for Generating Neutron Star Equations of State
Abstract
We develop a partially supervised latent-space variational autoencoder (PSL-VAE) framework to reconstruct and generate neutron star (NS) equations of state (EOS). The PSL-VAE consists of an encoder network that maps high-dimensional EOS data into a lower-dimensional latent space and a decoder network that reconstructs the full EOS from the latent representation. The latent space includes supervised NS observables derived from the training EOS data, as well as variational latent variables that capture additional EOS features learned automatically. Using a PSL-VAE trained on a Skyrme EOS dataset, we find that a latent space consisting of two supervised observables, the maximum mass Mmax and the canonical radius R1.4, together with a single variational latent variable associated mainly with the EOS near the crust-core transition, is sufficient to reconstruct Skyrme EOSs with high fidelity. The decoder reconstructed EOSs reproduce Mmax and R1.4 with mean absolute percentage errors within 0.08%. Sampling the latent space generates new EOSs that are causal, thermodynamically stable, and consistent with imposed constraints on the supervised observables. The framework therefore provides a compact and physically interpretable parameterization of the NS EOS that is well suited for Bayesian inference with multimessenger observations, including pulsar mass-radius measurements and gravitational wave data.
Versions
Related
- PISCES: Physics-Informed Solar-wind Convolutional autoEncoder for Space-weather Anomaly Detection and Early Warning
- Parameter Estimation of Ringdown Quasinormal Modes with Autoencoder
- Unresolved Binary Systems in the Rubin Era I: An Autoencoder Framework for Binary Identification Applied to 47 Tucanae
- Generating radio continuum survey maps of an arbitrary size with latent diffusion models
- Revisiting damping of Alfvén waves in partially ionized plasmas
BibTeX
@misc{zhao2026partially,
title = {{A Partially Supervised Latent-Space Variational Autoencoder for Generating Neutron Star Equations of State}},
author = {Tianqi Zhao and Fanglida Yan and Alex Ross and James M. Lattimer},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2605.27562},
eprint = {2605.27562},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2605.27562},
}
RIS
TY - UNPB
TI - A Partially Supervised Latent-Space Variational Autoencoder for Generating Neutron Star Equations of State
AU - Tianqi Zhao
AU - Fanglida Yan
AU - Alex Ross
AU - James M. Lattimer
PY - 2026
DA - 2026/09/15
PB - arXiv
DO - 10.48550/arxiv.2605.27562
UR - https://arxiv.org/abs/2605.27562
AB - We develop a partially supervised latent-space variational autoencoder (PSL-VAE) framework to reconstruct and generate neutron star (NS) equations of state (EOS). The PSL-VAE consists of an encoder network that maps high-dimensional EOS data into a lower-dimensional latent space and a decoder network that reconstructs the full EOS from the latent representation. The latent space includes supervised NS observables derived from the training EOS data, as well as variational latent variables that capture additional EOS features learned automatically. Using a PSL-VAE trained on a Skyrme EOS dataset, we find that a latent space consisting of two supervised observables, the maximum mass M_max and the canonical radius R_1.4, together with a single variational latent variable associated mainly with the EOS near the crust-core transition, is sufficient to reconstruct Skyrme EOSs with high fidelity. The decoder reconstructed EOSs reproduce M_max and R_1.4 with mean absolute percentage errors within 0.08%. Sampling the latent space generates new EOSs that are causal, thermodynamically stable, and consistent with imposed constraints on the supervised observables. The framework therefore provides a compact and physically interpretable parameterization of the NS EOS that is well suited for Bayesian inference with multimessenger observations, including pulsar mass-radius measurements and gravitational wave data.
ER -