SciML / SciML/NeuralOperators.jl

Implement Multi-Fidelity DeepONet

Open
#102 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 Multi-Fidelity DeepONet for learning operators from datasets with mixed fidelity levels (e.g., coarse and fine simulations).

Reference

  • Howard et al., "Multifidelity deep operator networks for data-driven and physics-informed problems," Journal of Computational Physics, 2023. DOI: 10.1016/j.jcp.2023.112462

Description

Multi-Fidelity DeepONet uses a composite architecture with separate DeepONet components for low-fidelity and high-fidelity mappings. The low-fidelity component learns from abundant cheap data, while the high-fidelity component learns the correction/residual from sparse expensive data. This is particularly useful in engineering applications where high-fidelity simulations are costly.

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 by reading the existing DeepONet implementation in NeuralOperators.jl and the cited Howard et al. paper to identify how low- and high-fidelity data should be represented. Define the architecture and validation criteria for learning a low-fidelity mapping and a high-fidelity correction, then add tests showing the combined model handles mixed-fidelity datasets.

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.