Research map / Label-efficient and robust learning / Semi-supervised learning
ExtremeWeather: A large-scale climate dataset for semi-supervised detection, localization, and understanding of extreme weather events
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 reconstruction training on unlabeled timesteps (λ=1) (ablation) | 3D CNN · ExtremeWeather-small validation | mAP at IoU=0.1: 51.00 -> 52.92 | Helped |
| add reconstruction training on unlabeled timesteps (λ=1) (ablation) | 3D CNN · ExtremeWeather-small validation | mAP at IoU=0.5: 11.60 -> 7.31 | Hurt |
| add temporal convolutions (ablation) | supervised CNN · ExtremeWeather-small validation | mAP at IoU=0.5: 16.98 -> 11.60 | Hurt |
| add temporal convolutions (ablation) | supervised CNN · ExtremeWeather-small validation | AP for ETC and TC: ETC: 21.92; 14.42 -> 22.65; 15.53; TC: 52.26; 9.23 -> 50.01; 9.12 | Mixed |
| add reconstruction training on unlabeled timesteps (ablation) | 3D CNN · ExtremeWeather-small validation | AP for ETC and TC at IoU=0.1: ETC: 22.65 -> 24.74; TC: 50.01 -> 56.40 | Helped |
| add reconstruction training on unlabeled timesteps (ablation) | 2D CNN, 66.53M parameters · ExtremeWeather-small validation | mAP at IoU=0.1: 51.42 -> 51.11 | Hurt |
| increase reconstruction-loss weight from λ=1 to λ=10 (ablation) | 2D CNN, 66.53M parameters · ExtremeWeather-small validation | mAP at IoU=0.1: 51.11 -> 48.85 | Hurt |
| increase the 2D model from 16.68M to 66.53M parameters (ablation) | supervised CNN · ExtremeWeather-small validation | mAP at IoU=0.1: 49.21 -> 51.42 | Helped |
Closest papers by meaning
- HR-Extreme: A High-Resolution Dataset for Extreme Weather Forecasting (ICLR 2025 · similarity 0.88)
- Identifying Spatio-Temporal Drivers of Extreme Events (NeurIPS 2024 · similarity 0.86)
- WeatherGFM: Learning a Weather Generalist Foundation Model via In-context Learning (ICLR 2025 · similarity 0.85)
- ERF: A Benchmark Dataset for Robust Semantic Segmentation Under Extreme Rainfall Conditions (AAAI 2025 · similarity 0.85)
- OneForecast: A Universal Framework for Global and Regional Weather Forecasting (ICML 2025 · similarity 0.84)
- Scaling transformer neural networks for skillful and reliable medium-range weather forecasting (NeurIPS 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.