Generative models that recover full physical fields from sparse, partial observations, for forward and inverse problems alike.
Website: https://generativepde.github.io
| Project | Venue | Paper | Page | Code |
|---|---|---|---|---|
| VideoPDE: Unified Generative PDE Solving via Video Inpainting Diffusion Models | arXiv 2025 | arXiv | page | VideoPDE |
| DiffusionPDE: Generative PDE-Solving Under Partial Observation | NeurIPS 2024 | arXiv | page | DiffusionPDE |
Starting a new project? Create your repository from project-template, then add it to the website.
Related: Neural Solvers, deterministic solvers that correct their own PDE residual.