Differentiable Stellar Atmospheres with Physics-Informed Neural Networks

2025/07/08 by Li, Jiadong, Jian, Mingjie, Ting, Yuan-Sen +1
#Astrophysics of Galaxies (astro-ph.GA) #Earth and Planetary Astrophysics (astro-ph.EP) #FOS: Physical sciences #Instrumentation and Methods for Astrophysics (astro-ph.IM) #Solar and Stellar Astrophysics (astro-ph.SR)

paper · doi:10.48550/arxiv.2507.06357

Abstract

We present Kurucz-a1, a physics-informed neural network (PINN) that emulates 1D stellar atmosphere models under Local Thermodynamic Equilibrium (LTE), addressing a critical bottleneck in differentiable stellar spectroscopy. By incorporating hydrostatic equilibrium as a physical constraint during training, Kurucz-a1 creates a differentiable atmospheric structure solver that maintains physical consistency while achieving computational efficiency. Kurucz-a1 can achieve superior hydrostatic equilibrium and more consistent with the solar observed spectra compared to ATLAS-12 itself, demonstrating the advantages of modern optimization techniques. Combined with modern differentiable radiative transfer codes, this approach enables data-driven optimization of universal physical parameters across diverse stellar populations-a capability essential for next-generation stellar astrophysics.

Citations

• Korg: fitting, model atmosphere interpolation, and Brackett lines
• Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
• Hidden Fluid Mechanics: A Navier-Stokes Informed Deep Learning Framework\n for Assimilating Flow Visualization Data
• The Deep Ritz method: A deep learning-based numerical algorithm for solving variational problems
• Gaussian Error Linear Units (GELUs)
• New Grids of ATLAS9 Model Atmospheres
• Uncertainties in stellar abundance analyses
• Line formation in solar granulation: I. Fe line shapes, shifts and asymmetries
• Line formation in solar granulation: II. The photospheric Fe abundance

Related

• Data-driven radiative hydrodynamics simulations of the solar photosphere using physics-informed neural networks: proof of concept
• Generating stellar spectra using Neural Networks
• MaNN: Multiple Artificial Neural Networks for modelling the Interstellar Medium
• Approximating Rayleigh Scattering in Exoplanetary Atmospheres using Physics-informed Neural Networks (PINNs)
• Using Physics Informed Neural Networks for Supernova Radiative Transfer Simulation