Research map / Probabilistic and scientific modeling / Differential equations and PDEs
Fourier Neural Operators for Arbitrary Resolution Climate Data Downscaling
Qidong Yang, Alex Hernandez-Garcia, Paula Harder, Venkatesh Ramesh, Prasanna Sattigeri, Daniela Szwarcman, Campbell D. Watson, David Rolnick
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add softmax constraint layer (ablation) | DFNO · Navier-Stokes PDE 2x | MSE: 0.0015 -> 0.0011 | Improved |
| add softmax constraint layer (ablation) | DFNO · Navier-Stokes PDE 4x (zero-shot) | MSE: 0.0037 -> 0.0029 | Improved |
| add softmax constraint layer (ablation) | DFNO · ERA5 water content 2x | MSE: 0.2063 -> 0.1696 | Improved |
| add softmax constraint layer (ablation) | DFNO · ERA5 water content 4x (zero-shot) | MSE: 0.3628 -> 0.2779 | Improved |
| add softmax constraint layer (ablation) | CNN-2 (resized with bicubic) · Navier-Stokes PDE 4x | MSE: 0.0093 -> 0.0217 | Hurt |
| add softmax constraint layer (ablation) | GAN-2 (resized with bicubic) · Navier-Stokes PDE 4x | MSE: 0.0095 -> 0.0228 | Hurt |
| add softmax constraint layer | CNN-4 (resized with bicubic) · Navier-Stokes PDE 4x | MSE: 0.0070 -> 0.0063 | No clear effect |
| train DFNO on 4x data instead of zero-shot from 2x training | DFNO (constrained) · Navier-Stokes PDE integration 64x64 | MSE: 0.0018 -> 0.0007 | Improved |
| train DFNO on 2x data instead of 4x data | DFNO (constrained) · Navier-Stokes PDE integration 32x32 | MSE: 0.0012 -> 0.0004 | Improved |
| downscale low-resolution solver output with DFNO-2 instead of using FFNO-64 solver | DFNO-2 · Navier-Stokes PDE integration 64x64 | MSE: 0.0118 -> 0.0018 | Improved |
| replace CNN-4 with DFNO trained only on 2x | DFNO · ERA5 water content 4x (zero-shot, unconstrained) | MAE: 0.3010 -> 0.3067 | Hurt |
| replace CNN-2 with DFNO | DFNO · Navier-Stokes PDE 2x (unconstrained) | MSE: 0.0042 -> 0.0015 | Improved |
Closest papers by meaning
- FNP: Fourier Neural Processes for Arbitrary-Resolution Data Assimilation (NeurIPS 2024 · similarity 0.87)
- Precipitation Downscaling with Spatiotemporal Video Diffusion (NeurIPS 2024 · similarity 0.85)
- Neural Operators with Localized Integral and Differential Kernels (ICML 2024 · similarity 0.85)
- Probablistic Emulation of a Global Climate Model with Spherical DYffusion (NeurIPS 2024 · similarity 0.85)
- DiffDA: a Diffusion model for weather-scale Data Assimilation (ICML 2024 · similarity 0.84)
- Satellite Observations Guided Diffusion Model for Accurate Meteorological States at Arbitrary Resolution (CVPR 2025 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
