💡 Main Idea

We propose to build an educational and onboarding demo project within the Anemoi framework by reproducing the model presented in Keisler 2022. This demo project will provide a practical, step-by-step pipeline: from dataset preparation to model verification and visualization, specifically tailored for teaching and training purposes.

In short:

• Create an accessible and fully documented demo repository demonstrating the complete MLWP pipeline using Anemoi.
• Enhance understanding of the scientific principles underlying graph-based MLWP models.
• Simplify the onboarding experience for new users by showcasing a clear, real-world example.
• Provide educational material suitable for lectures and practical exercises for students learning MLWP

⛩️ Philosophy

• The project must offer a pleasant introduction to MLWP and invite users to think more deeply about the MLWP problem statement and the choices made when building a MLWP-model.
• Bridge the gap between anemoi-docs and anemoi-configs on the one hand and the different anemoi packages on the other hand.
• Explain the scientific and technical choices made
• Interact as much as possible with anemoi-docs

🧑‍💻 Code base

https://github.com/mpvginde/anemoi-demo-keisler2022

:slack: Slack channel

#anemoi-demo-keisler2022