Research map / Probabilistic and scientific modeling / Molecules and proteins
PreDiff: Precipitation Nowcasting with Latent Diffusion Models
Zhihan Gao, Xingjian Shi, Boran Han, Hao Wang, Xiaoyong Jin, Danielle C. Maddix, Yi Zhu, Mu Li, Bernie Wang
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 energy-conservation knowledge alignment to PreDiff (ablation) | PreDiff · N-body MNIST | E.MSE: 0.0226 -> 0.0039 | Improved |
| add energy-conservation knowledge alignment to PreDiff (ablation) | PreDiff · N-body MNIST | E.MAE: 0.1083 -> 0.0443 | Improved |
| add energy-conservation knowledge alignment to PreDiff (ablation) | PreDiff · N-body MNIST | MSE: 9.492 -> 21.90 | Hurt |
| add energy-conservation knowledge alignment to PreDiff (ablation) | PreDiff · N-body MNIST | FVD: 0.987 -> 4.063 | Hurt |
| add anticipated-intensity knowledge alignment to PreDiff (ablation) | PreDiff · SEVIR | FVD: 33.05 -> 34.18 | No clear effect |
| replace UNet backbone in LDM with Earthformer-UNet (ablation) | LDM · N-body MNIST | MSE: 46.29 -> 9.492 | Improved |
| replace UNet backbone in LDM with Earthformer-UNet (ablation) | LDM · N-body MNIST | FVD: 3.432 -> 0.987 | Improved |
| replace UNet backbone in LDM with Earthformer-UNet (ablation) | LDM · SEVIR | FVD: 133.0 -> 33.05 | Improved |
| use probabilistic PreDiff instead of deterministic Earthformer | Earthformer · SEVIR | CSI: 0.4419 -> 0.4100 | Hurt |
| use probabilistic PreDiff instead of deterministic Earthformer | Earthformer · SEVIR | CSI-pool16: 0.5005 -> 0.6244 | Improved |
| use probabilistic PreDiff instead of deterministic Earthformer | Earthformer · SEVIR | CRPS: 0.0304 -> 0.0246 | Improved |
Later papers that compared with this method
2 later accepted papers report PreDiff as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- DiffCast: A Unified Framework via Residual Diffusion for Precipitation Nowcasting (CVPR 2024 · better on 2 of 3 shared results)
- CasCast: Skillful High-resolution Precipitation Nowcasting via Cascaded Modelling (ICML 2024 · better on 1 of 2 shared results)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- OmniCast: A Masked Latent Diffusion Model for Weather Forecasting Across Time Scales (NeurIPS 2025 · similarity 0.89)
- PostCast: Generalizable Postprocessing for Precipitation Nowcasting via Unsupervised Blurriness Modeling (ICLR 2025 · similarity 0.88)
- Probablistic Emulation of a Global Climate Model with Spherical DYffusion (NeurIPS 2024 · similarity 0.88)
- DiffDA: a Diffusion model for weather-scale Data Assimilation (ICML 2024 · similarity 0.87)
- Elucidated Rolling Diffusion Models for Probabilistic Forecasting of Complex Dynamics (NeurIPS 2025 · similarity 0.87)
- CasCast: Skillful High-resolution Precipitation Nowcasting via Cascaded Modelling (ICML 2024 · similarity 0.87)
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.
