← Papers
Transformers

GNOT

A General Neural Operator Transformer for Operator Learning

Paper title GNOT: A General Neural Operator Transformer for Operator Learning

Hao et al. (Tsinghua University) · Aero & CFD · Thermal · General PDEs · Public: code, data

Paper arXiv 2302.14376PDF

A transformer that reads the query points of an irregular mesh as tokens and lets them attend, with a linear-cost normalised attention, to any number of input functions (boundary shapes, material parameters, source terms), each encoded separately. A gate driven by the point coordinates mixes several expert feed-forward networks, which acts like a soft domain decomposition for multi-scale problems.

Published
ICML

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?