← Papers
Transformers

OFormer

Transformer for Partial Differential Equations' Operator Learning

Paper title Transformer for Partial Differential Equations' Operator Learning

Li et al. (Carnegie Mellon University) · Aero & CFD · Electromagnetics · General PDEs · Public: code

Paper arXiv 2205.13671PDF

Treats the input field as a set of points with values and coordinates: self-attention (linear-cost) encodes them, then each query point reads the encoding through cross-attention, so input and output can sit on different, irregular or sparse sets of points. Time-dependent problems are rolled forward in a latent space before decoding.

Published
TMLR

Share on LinkedIn, X or Bluesky, or or .

Know related work that belongs here?

Something wrong on this page?

More papers

All papers
  1. TransformersMNC — Retrofitting foundation models to unstructured meshes
  2. TransformersPSP grounding — Correcting a CFD-trained wing surrogate with wind-tunnel pressure paint
  3. TransformersTransolver without attention — Does Transolver really need a Transformer?