← Papers
Latent models

LED

Multiscale simulations of complex systems by learning their effective dynamics

Paper title Multiscale simulations of complex systems by learning their effective dynamics

Vlachas et al. (ETH Zurich, Harvard) · Aero & CFD · General PDEs · Public: code, data

Paper arXiv 2006.13431PDF

An autoencoder compresses the full simulation state to a few latent variables and an LSTM advances them in time (Latent-LED). Following the equation-free approach, the scheme can also alternate between the latent model and short runs of the original solver started from the decoded state, which corrects the accumulated error (Multiscale-LED).

Published
Nature Machine Intelligence

Share on LinkedIn, X or Bluesky, or or .

Know related work that belongs here?

Something wrong on this page?

More papers

All papers
  1. Latent modelsD-JEPA — One geometry latent, reused across flow conditions and physics
  2. Latent modelsCourant — A state-adaptive Perceiver surrogate with local support
  3. Latent modelsAeroJEPA — Semantic Latent Representations for 3D Aerodynamic Fields