Research map / Probabilistic and scientific modeling / Differential equations and PDEs
M2N: Mesh Movement Networks for PDE Solvers
Wenbin Song, Mingrui Zhang, Joseph Gregory Wallwork, Junpeng Gao, Zheng Tian, Fanglei Sun, Matthew D Piggott, Junqing Chen, Zuoqiang Shi, Xiang Chen, Jun Wang
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 |
|---|---|---|---|
| use neural spline deformer instead of MLP with clipping (ablation) | M2N-Spline · Poisson square | error reduction (%): 16.74 -> 20.82 | Improved |
| use neural spline deformer instead of MLP with clipping (ablation) | M2N-Spline · Poisson square | element inversion (%): 1.60 -> 0.00 | Improved |
| use GAT deformer instead of ordinary GAT block with clipping (ablation) | M2N-GAT · Poisson square | element inversion (%): 3.11 -> 0.00 | Improved |
| use GAT deformer instead of ordinary GAT block with clipping (ablation) | M2N-GAT · Poisson square | error reduction (%): 19.95 -> 20.38 | No clear effect |
| use GAT deformer instead of ordinary GAT block with clipping (ablation) | M2N-GAT · Poisson heptagon | error reduction (%): 22.40 -> 22.39 | No clear effect |
| use GAT deformer instead of ordinary GAT block with clipping (ablation) | M2N-GAT · Poisson heptagon | element inversion (%): 0.68 -> 0.00 | Improved |
| use neural spline deformer instead of MLP with clipping (ablation) | M2N-Spline · Poisson heptagon | error reduction (%): 16.49 -> 16.15 | No clear effect |
| use GAT deformer instead of ordinary GAT block with clipping (ablation) | M2N-GAT · Burgers | error reduction (%): 51.69 -> 57.75 | Improved |
| use neural spline deformer instead of MLP with clipping (ablation) | M2N-Spline · Burgers | error reduction (%): 43.53 -> 48.92 | Improved |
| use GAT-based instead of neural-spline mesh deformer on irregular domain | M2N-GAT vs M2N-Spline · Poisson heptagon | error reduction (%): 16.15 -> 22.39 | Improved |
| replace Monge-Ampere with learned mesh movement (M2N-Spline) | M2N-Spline · Poisson square | time (ms): 5220.99 -> 5.55 | Improved |
| replace Monge-Ampere with learned M2N-GAT | M2N-GAT · Burgers | error reduction (%): 60.24 -> 57.75 | Hurt |
Closest papers by meaning
- Towards Universal Mesh Movement Networks (NeurIPS 2024 · similarity 0.93)
- Better Neural PDE Solvers Through Data-Free Mesh Movers (ICLR 2024 · similarity 0.91)
- UGM2N: An Unsupervised and Generalizable Mesh Movement Network via M-Uniform Loss (NeurIPS 2025 · similarity 0.89)
- PhyMPGN: Physics-encoded Message Passing Graph Network for spatiotemporal PDE systems (ICLR 2025 · similarity 0.88)
- P$^2$C$^2$Net: PDE-Preserved Coarse Correction Network for efficient prediction of spatiotemporal dynamics (NeurIPS 2024 · similarity 0.88)
- Graph Neural PDE Solvers with Conservation and Similarity-Equivariance (ICML 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.
