Research map / Probabilistic and scientific modeling / Differential equations and PDEs
Learning Neural PDE Solvers with Parameter-Guided Channel Attention
Makoto Takamoto, Francesco Alesiani, Mathias Niepert
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 CAPE to FNO | FNO · 1D Advection | nRMSE: 0.69 -> 0.03 | Improved |
| Add CAPE to FNO | FNO · 1D Burgers | nRMSE: 0.54 -> 0.09 | Improved |
| Add CAPE to FNO | FNO · 2D NS | nRMSE: 1.06 -> 0.80 | Improved |
| Add CAPE to FNO | FNO · 2D Burgers | nRMSE: 0.0249 -> 0.0267 | Hurt |
| Add PDE parameter channels as input | FNO · 1D Advection | nRMSE: 0.69 -> 0.05 | Improved |
| Add PINO loss | FNO · 1D Advection | nRMSE: 0.69 -> 0.70 | No clear effect |
| Use curriculum training instead of fully autoregressive training (ablation) | CAPE + FNO · 2D NS | nMSE: 1.3×10^{+0} -> 8.0 × 10^{-1} | Improved |
| Use curriculum training instead of pure teacher-forcing (ablation) | CAPE + FNO · 2D NS | nMSE: 3.2×10^{+0} -> 8.0 × 10^{-1} | Improved |
| Add LayerNorm to CAPE (ablation) | CAPE + FNO · 2D NS | nRMSE: 1.24 -> 0.80 | Improved |
| Add spectral convolution to CAPE (ablation) | CAPE + FNO · 1D Advection | nRMSE: 0.06 -> 0.04 | Improved |
| Remove LayerNorm from CAPE (ablation) | CAPE + FNO · 1D Advection | nRMSE: 0.04 -> 0.03 | Improved |
| Remove LayerNorm from CAPE (ablation) | CAPE + FNO · 1D Burgers | nRMSE: 0.13 -> 0.09 | Improved |
Closest papers by meaning
- Active Learning for Neural PDE Solvers (ICLR 2025 · similarity 0.89)
- P$^2$C$^2$Net: PDE-Preserved Coarse Correction Network for efficient prediction of spatiotemporal dynamics (NeurIPS 2024 · similarity 0.88)
- CALM-PDE: Continuous and Adaptive Convolutions for Latent Space Modeling of Time-dependent PDEs (NeurIPS 2025 · similarity 0.88)
- Learning a Neural Solver for Parametric PDEs to Enhance Physics-Informed Methods (ICLR 2025 · similarity 0.88)
- Data-Efficient Operator Learning via Unsupervised Pretraining and In-Context Learning (NeurIPS 2024 · similarity 0.87)
- Poseidon: Efficient Foundation Models for PDEs (NeurIPS 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.
