A convolutional VAE trained on projected dark matter density slices produces synthetic fields whose power spectra roughly match the training simulation at intermediate scales, with small-scale smoothing and validation only against the same simulation.
Enzo: An adaptive mesh refinement code for astrophysics
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
astro-ph.CO 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Variational autoencoder for generating realistic $N$-body simulations for dark matter halos
A convolutional VAE trained on projected dark matter density slices produces synthetic fields whose power spectra roughly match the training simulation at intermediate scales, with small-scale smoothing and validation only against the same simulation.