Reproduce “Learned Surrogates of Three-Dimensional Magnetohydrodynamic Turbulence” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Semih Kacmaz, E. A. Huerta
The authors evaluate surrogates for three-dimensional resistive MHD turbulence, training a UNet and shifted-window transformer to predict and feed back states. In decaying, incompressible simulations with a weak magnetic field, both retain recognizable structures over 25 steps, about one turnover time. At Reynolds number 6000, velocity errors are 0.053 and 0.055; this does not establish longer-term fidelity or solver speedup.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.