SciML / SciML/NeuralOperators.jl

Implement GNOT (General Neural Operator Transformer)

Open
#117 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 GNOT, a general-purpose transformer-based neural operator with heterogeneous cross-attention and linear complexity.

Reference

  • Hao et al., "GNOT: A General Neural Operator Transformer for Operator Learning," ICML 2023. Paper

Description

GNOT uses heterogeneous normalized cross-attention to handle multiple types of input conditions (initial conditions, boundary conditions, forcing terms, PDE coefficients) within a single architecture. It achieves linear-time complexity through its attention mechanism, making it scalable to large grids. Designed as a general-purpose operator learning framework.

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

Start with the GNOT paper linked in the issue and inspect the NeuralOperators.jl repository to identify the relevant operator entry points. The issue names no files or tests, so the implementation scope and validation criteria need to be established before work begins; done should mean a working GNOT implementation matching the described heterogeneous attention and linear-complexity goals.

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
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.