SciML / SciML/NeuralOperators.jl

Implement Geo-FNO (Geometry-Aware Fourier Neural Operator)

Open
#106 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Julia
Stars
41
Forks
15
Avg merge
13h 14m
Merged PRs (30d)
12

Description

Summary

Implement Geo-FNO, which learns a deformation from arbitrary input geometries to a uniform latent grid where standard FFT can be applied.

Reference

  • Li et al., "Fourier Neural Operator with Learned Deformations for PDEs on General Geometries," JMLR, 2023. arXiv:2207.05209

Description

Standard FNO requires uniform grids for FFT. Geo-FNO adds a learnable deformation network that maps arbitrary geometries (point clouds, meshes, irregular grids) to a uniform latent space, applies standard FNO layers, then maps back. This enables FNO to handle complex geometries while retaining FFT efficiency. The paper reports 10^5x speedups over numerical solvers.

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

No implementation files, tests, or entry point are named. Start with the referenced JMLR paper or arXiv article and locate the repository's existing FNO implementation; completion should cover deformation to a uniform latent grid, FNO processing, and mapping back for arbitrary geometries.

Written by the indexing model from the issue text.

Assessment

Tech stack
julia
Domain
machine-learning
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
30/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.