Research map / Video / Video segmentation and tracking
Met2Net: A Decoupled Two-Stage Spatio-Temporal Forecasting Model for Complex Meteorological Systems
Shaohan Li, Hao Yang, Min Chen, Xiaolin Qin
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 Multiple Encoder and Decoder (MED) (ablation) | multivariate TAU baseline · Weather_L (T2M) | MSE: 2.2247 -> 1.6708 | Improved |
| add Multiple Encoder and Decoder (MED) (ablation) | multivariate TAU baseline · Weather_L (UV10) | MSE: 1.8347 -> 1.8327 | No clear effect |
| add Multiple Encoder and Decoder (MED) (ablation) | multivariate TAU baseline · Weather_L (TCC) | MSE: 0.0439 -> 0.0445 | Hurt |
| add Variable Attention (VA) (ablation) | Baseline+MED · Weather_L (UV10) | MSE: 1.8327 -> 1.6287 | Improved |
| add Variable Attention (VA) (ablation) | Baseline+MED · Weather_L (R) | MSE: 28.8003 -> 26.6849 | Improved |
| add Implicit Two-Stage training (ITS) (ablation) | Baseline+MED+VA · Weather_L (T2M) | MSE: 1.5432 -> 0.8247 | Improved |
| add Implicit Two-Stage training (ITS) (ablation) | Baseline+MED+VA · Weather_L (UV10) | MSE: 1.6287 -> 1.5055 | Improved |
| predict multiple variables jointly instead of single variable (ablation) | TAU · Weather_L (T2M) | MSE: 1.1620 -> 2.2247 | Hurt |
| scale to more variables (8 encoder-decoder pairs) | TAU · Weather_L (UV10) | MSE: 1.5925 -> 1.7345 | Hurt |
| scale to more variables (8 encoder-decoder pairs) | Met2Net · Weather_L (UV10) | MSE: 1.5055 -> 1.4740 | Improved |
| use TAU translator block instead of Moga in Met2Net (ablation) | Met2Net · Weather_L (T2M) | MSE: 1.0314 -> 0.8271 | Improved |
| use TAU translator block instead of ConvNext in Met2Net (ablation) | Met2Net · Weather_L (T2M) | MSE: 1.0238 -> 0.8271 | Improved |
Closest papers by meaning
- iTrendRNN: An Interpretable Trend-Aware RNN for Meteorological Spatiotemporal Prediction (AAAI 2024 · similarity 0.86)
- OmniCast: A Masked Latent Diffusion Model for Weather Forecasting Across Time Scales (NeurIPS 2025 · similarity 0.86)
- On the Integration of Spatial-Temporal Knowledge: A Lightweight Approach to Atmospheric Time Series Forecasting (NeurIPS 2025 · similarity 0.85)
- WeatherGFM: Learning a Weather Generalist Foundation Model via In-context Learning (ICLR 2025 · similarity 0.85)
- Integrating Multi-Source Data for Long Sequence Precipitation Forecasting (AAAI 2025 · similarity 0.85)
- EWMoE: An Effective Model for Global Weather Forecasting with Mixture-of-Experts (AAAI 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.
