#machine learning
May 2025
Surrogate Modeling of 3D Rayleigh-Benard Convection with Equivariant Autoencoders
An end-to-end equivariant surrogate model consisting of an equivariant convolutional autoencoder and an equivariant convolutional LSTM using $G$-steerable kernels is presented, demonstrating significant gains in sample and parameter efficiency, as well as a better scaling to more complex dynamics.
Fynn Fromme, Christine Allen-Blanchette, Hans Harder et al.
· arXiv.org · 3 citations