Research map / Architectures and efficiency / CNN design
Deep Learning for Precipitation Nowcasting: A Benchmark and A New Model
Xingjian Shi, Zhihan Gao, Leonard Lausen, Hao Wang, Dit-Yan Yeung, Wai-kin Wong, Wang-chun WOO
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 |
|---|---|---|---|
| replace fixed convolutional recurrent connections with learned, location-varying trajectories | ConvGRU → TrajGRU · HKO-7 offline | B-MSE: 5951 -> 5816 | Improved |
| replace fixed convolutional recurrent connections with learned, location-varying trajectories | ConvGRU → TrajGRU · HKO-7 online, r ≥ 30 | CSI: 0.1837 -> 0.1933 | Improved |
| train ConvGRU with balanced losses rather than original MSE and MAE (ablation) | ConvGRU · HKO-7 offline | B-MSE: 9087 -> 5951 | Improved |
| train ConvGRU with balanced losses rather than original MSE and MAE (ablation) | ConvGRU · HKO-7 offline, r ≥ 30 | CSI: 0.0712 -> 0.1776 | Improved |
| enable online fine-tuning | TrajGRU · HKO-7, r ≥ 30 | CSI: 0.1856 -> 0.1933 | Improved |
| increase learned recurrent links from 5 to 13 (ablation) | TrajGRU · MovingMNIST++ test set | Test MSE × 10⁻²: 1.351 -> 1.170 | Improved |
| use DFN local-filter prediction instead of TrajGRU | MovingMNIST++ video prediction · MovingMNIST++ test set | Test MSE × 10⁻²: 1.138 -> 1.461 | Hurt |
| use 3D rather than 2D CNN | CNN video prediction · HKO-7 offline | B-MSE: 7332 -> 7202 | Improved |
| use balanced losses instead of conventional unweighted losses | precipitation nowcasting loss evaluation · HKO-7 benchmark | correlation with CSI and HSS | Improved |
Closest papers by meaning
- PostCast: Generalizable Postprocessing for Precipitation Nowcasting via Unsupervised Blurriness Modeling (ICLR 2025 · similarity 0.87)
- Perceptually Constrained Precipitation Nowcasting Model (ICML 2025 · similarity 0.87)
- CasCast: Skillful High-resolution Precipitation Nowcasting via Cascaded Modelling (ICML 2024 · similarity 0.87)
- Data-driven Precipitation Nowcasting Using Satellite Imagery (AAAI 2025 · similarity 0.86)
- Fourier Amplitude and Correlation Loss: Beyond Using L2 Loss for Skillful Precipitation Nowcasting (NeurIPS 2024 · similarity 0.85)
- DiffCast: A Unified Framework via Residual Diffusion for Precipitation Nowcasting (CVPR 2024 · 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.
