Research map / Graph learning / GNN architectures
Graph Rewiring based on Flow Alignment for Improving Fluid Simulation
Zenong Li, Wei Xian Lim, Wai Lee Chan, Adams Wai-Kin Kong
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 |
|---|---|---|---|
| rewire the mesh graph with flow-aligned 2-hop connections (FLARE) versus no rewiring | MGN · CylinderFlow | full-rollout RMSE: 40.35 -> 23.38 | Improved |
| use FLARE instead of PIORF | MGN · CylinderFlow | full-rollout RMSE: 33.59 -> 23.38 | Improved |
| use FLARE instead of non-physics 2-HOP-ALL | MGN · CylinderFlow | full-rollout RMSE: 29.40 -> 23.38 | Improved |
| connect 2-hop neighbors against the flow direction (2-HOP-OPPOSITE) instead of along it (ablation) | FLARE · CylinderFlow | full-rollout RMSE: 23.38 -> 55.48 | Hurt |
| use opposite-direction connections instead of flow-aligned ones (ablation) | FLARE · Airfoil density | full-rollout RMSE: 90.58 -> 96.39 | Hurt |
| extend flow-aligned rewiring from 2-hop to 3-hop (ablation) | FLARE · CylinderFlow | full-rollout RMSE: 23.38 -> 28.67 | Hurt |
| extend flow-aligned rewiring from 2-hop to 4-hop (ablation) | FLARE · CylinderFlow | full-rollout RMSE: 23.38 -> 34.87 | Hurt |
| add 10% density-specific 2-hop rewiring to FLARE (ablation) | FLARE (MGN) · Airfoil density | full-rollout RMSE: 90.58 -> 85.66 | Improved |
| add two message-passing blocks before Transolver++ | Transolver++ · CylinderFlow | full-rollout RMSE: 38.12 -> 32.77 | Improved |
| PIORF rewiring versus no rewiring on density | BSMS-GNN · Airfoil density | full-rollout RMSE: 126.78 -> 99.75 | Improved |
| random 2-hop rewiring instead of no rewiring | BSMS-GNN · Airfoil density | full-rollout RMSE: 126.78 -> 195.92 | Hurt |
| use FLARE instead of 2-HOP-ALL on Transolver++ Tandem-Airfoil | Transolver++ · Tandem-Airfoil-Cruise | MSE: 0.66 -> 0.68 | No clear effect |
Closest papers by meaning
- PIORF: Physics-Informed Ollivier-Ricci Flow for Long–Range Interactions in Mesh Graph Neural Networks (ICLR 2025 · similarity 0.92)
- MeshMask: Physics-Based Simulations with Masked Graph Neural Networks (ICLR 2025 · similarity 0.87)
- Learning Distributions of Complex Fluid Simulations with Diffusion Graph Networks (ICLR 2025 · similarity 0.85)
- Finite Volume Features, Global Geometry Representations, and Residual Training for Deep Learning-based CFD Simulation (ICML 2024 · similarity 0.85)
- Deep Learning-based Algebraic Reynolds Stress Closures for RANS simulations of Turbulent Flows (NeurIPS 2026 · similarity 0.85)
- Gradient Alignment in Physics-informed Neural Networks: A Second-Order Optimization Perspective (NeurIPS 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.
