← Papers
Graphs

DGN / LDGN

Sampling flow distributions on meshes with diffusion graph networks

Paper title Learning Distributions of Complex Fluid Simulations with Diffusion Graph Networks

Lino et al. (TUM and Google DeepMind) · Aero & CFD · Public: code, data, weights

Paper arXiv 2504.02843PDF

A diffusion model whose denoiser is a multi-scale graph network on the simulation mesh, optionally in a learned latent space, draws independent flow states from the statistically steady distribution for a given geometry and flow condition, so RMS and other statistics come from samples rather than a long run.

Published
ICLR

Share on LinkedIn, X or Bluesky, or or .

Know related work that belongs here?

Something wrong on this page?

More papers

All papers
  1. GraphsR-BiLoGo — Transfer learning for rib effects in sheet-forming surrogates
  2. GraphsRead-Write-Relax — Why neural PDE surrogates need both global and local processing
  3. GraphsGraph-LED — Learning Effective Dynamics across Spatio-Temporal Scales of Complex Flows