Research map / Probabilistic and scientific modeling / Dynamical systems
Mesh Field Theory: Port–Hamiltonian Formulation of Mesh-Based Physics
Satoshi Noguchi, Yoshinobu Kawahara
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 |
|---|---|---|---|
| Fix incidence-wired conservative coupling and learn only metric and dissipation (MeshFT-Net vs MGN) | graph mesh simulator · Plane-wave regular grids | normalized energy drift: 25.9 -> 1.3 × 10−4 | Improved |
| Use MeshFT-Net instead of MGN | graph mesh simulator · Dissipative benchmark | one-step MSE: 1.2 × 10−7 vs 5.2 × 10−8 (MGN) | Hurt |
| Use MeshFT-Net instead of MGN | graph mesh simulator · Physics-consistency, momentum change | momentum change: 3.90 × 10−1 -> 4.89 × 10−8 | Improved |
| Use MeshFT-Net instead of baselines under OOD shifts | graph mesh simulator · OOD frequency, wave speed, resolution | energy drift: e.g. MGN >100 vs 5.1e-3 on frequency | Improved |
| Use MeshFT-Net on heterogeneous elasticity | graph mesh simulator · Irregular heterogeneous elasticity | TSMSE: 2.01 × 10−1 -> 4.91 × 10−3 | Improved |
| Restrict metric to spatially uniform (remove learned heterogeneous metric) (ablation) | MeshFT-Net · Heterogeneous elasticity | TSMSE: 4.83 × 10−3 -> 1.50 × 10−2 | Hurt |
| Add topology-fixed signed-incidence wiring (ablation) | MeshFT-Net · Heterogeneous elasticity | momentum variation: 1.26 × 10−2 -> 2.67 × 10−7 | Improved |
| Allow state-dependent J instead of fixed J (ablation) | MeshFT-Net · Controlled state-dependent coupling | TSMSE: 4.52 × 10−5 -> 6.77 × 10−6 | Improved |
| Use full cross-channel G instead of diagonal G with fixed J (ablation) | MeshFT-Net · Controlled state-dependent coupling | TSMSE: 4.52 × 10−5 -> 3.28 × 10−5 | Mixed |
| Add Hamiltonian penalty to MGN (MGN-HP) | MGN · Plane-wave regular grids | one-step MSE: 1.6 × 10−7 -> 5.7 × 10−4 | Hurt |
Closest papers by meaning
- A meshfree exterior calculus for generalizable and data-efficient learning of physics from point clouds (NeurIPS 2026 · similarity 0.87)
- JEPAWG: Interpretable Hypernetworks for Weight-Space Physics (NeurIPS 2026 · similarity 0.85)
- UGM2N: An Unsupervised and Generalizable Mesh Movement Network via M-Uniform Loss (NeurIPS 2025 · similarity 0.84)
- Riemann Tensor Neural Networks: Learning Conservative Systems with Physics-Constrained Networks (ICML 2025 · similarity 0.84)
- PIORF: Physics-Informed Ollivier-Ricci Flow for Long–Range Interactions in Mesh Graph Neural Networks (ICLR 2025 · similarity 0.83)
- Physics-informed Reduced Order Modeling of Time-dependent PDEs via Differentiable Solvers (NeurIPS 2025 · similarity 0.83)
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.
