Research map / Probabilistic and scientific modeling / Dynamical systems
PMGC-SimVP: Parametric Multi-scale Gated Convolution for Global Ionospheric TEC Prediction
Yu Xia, Yingkui Gong, Hao Zhang
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 parameter conditioning (space weather drivers) (ablation) | PMGC-SimVP · Global TEC overall | RMSE: 3.118 -> 3.051 | Improved |
| add dual-baseline residual prediction (ablation) | PMGC-SimVP · Global TEC overall | RMSE: 3.161 -> 3.051 | Improved |
| use Charbonnier loss (ablation) | PMGC-SimVP · Global TEC overall | RMSE: 3.164 -> 3.051 | Improved |
| add PMSC multi-scale module (ablation) | PMGC-SimVP · Global TEC overall | RMSE: 3.132 -> 3.051 | Improved |
| add spatial aggregation (SA) module (ablation) | PMGC-SimVP · Global TEC overall | RMSE: 3.092 -> 3.051 | Improved |
| zero out parameter inputs at test time (ablation) | PMGC-SimVP · Global TEC overall | RMSE: 3.051 -> 3.074 | Hurt |
| shuffle parameter inputs across batch at test time (ablation) | PMGC-SimVP · Global TEC overall | RMSE: 3.051 -> 3.066 | Hurt |
| replace WaST with PMGC-SimVP on stormy 2023 data | PMGC-SimVP · Global TEC 2023 stormy | RMSE: 4.439 -> 4.159 | Improved |
| use PMGC backbone without parameters on TaxiBJ vs WaST | PMGC-SimVP · TaxiBJ | SSIM: 0.984 -> 0.980 | Mixed |
| use low-rank temporal adapter | PMGC-SimVP · Global TEC | params/FPS: 195M -> 9.89M params; 989.8 -> 1745.2 FPS vs SimVP | Improved |
Closest papers by meaning
- DAWP: A framework for global observation forecasting via Data Assimilation and Weather Prediction in satellite observation space (NeurIPS 2025 · similarity 0.83)
- CALM-PDE: Continuous and Adaptive Convolutions for Latent Space Modeling of Time-dependent PDEs (NeurIPS 2025 · similarity 0.82)
- Tensor-Var: Efficient Four-Dimensional Variational Data Assimilation (ICML 2025 · similarity 0.82)
- Deep Random Features for Scalable Interpolation of Spatiotemporal Data (ICLR 2025 · similarity 0.82)
- Mechanism-Aware Ensemble Conditioning for Data-Limited Emulation of Extreme Events (NeurIPS 2026 · similarity 0.81)
- Predicting partially observable dynamical systems via diffusion models with a multiscale inference scheme (NeurIPS 2025 · similarity 0.81)
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.
