Fourier neural operators trained on 2D projected views of 3D astrophysical simulations can forecast subsequent projected density and velocity snapshots with 5-25% RMS error, though a constant hidden magnetic field means the latent-variable claim is not yet demonstrated.
Kovachki , author Z
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
astro-ph.GA 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Modeling turbulent and self-gravitating fluids with Fourier neural operators
Fourier neural operators trained on 2D projected views of 3D astrophysical simulations can forecast subsequent projected density and velocity snapshots with 5-25% RMS error, though a constant hidden magnetic field means the latent-variable claim is not yet demonstrated.