Research map / Probabilistic and scientific modeling / Differential equations and PDEs
Accelerated Training of Physics-Informed Neural Networks (PINNs) using Meshless Discretizations
Ramansh Sharma, Varun Shankar
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 |
|---|---|---|---|
| replace autograd spatial derivatives with RBF-FD discretization (p>2) and fp64 | vanilla-PINN · Linear Poisson, unit disk | training time: max 4x speedup | Improved |
| RBF-FD discretization for nonlinear Poisson | vanilla-PINN · Nonlinear Poisson, unit disk | training time | Improved |
| RBF-FD for spatial derivatives, autograd for time | vanilla-PINN · Heat equation | training time: 2-4x speedups | Improved |
| use RBF-FD order p=2 instead of p>2 (ablation) | DT-PINN · Linear Poisson relative l2 error | relative error | Hurt |
| use fp32 instead of fp64 for DT-PINN (ablation) | DT-PINN · Linear Poisson relative l2 error | relative error | Hurt |
| use fp64 instead of fp32 for vanilla-PINN (ablation) | vanilla-PINN · Linear Poisson training epochs | training speed | No clear effect |
| RBF-FD instead of autograd for derivative evaluation (p=4,5) (ablation) | Laplacian of random MLP · error vs fp64 autograd | relative error | Improved |
| RBF-FD instead of autograd for biharmonic operator (ablation) | random MLP · error vs fp64 autograd | relative error | Hurt |
Closest papers by meaning
- PINNsFormer: A Transformer-Based Framework For Physics-Informed Neural Networks (ICLR 2024 · similarity 0.90)
- Efficient Error Certification for Physics-Informed Neural Networks (ICML 2024 · similarity 0.90)
- Dual Cone Gradient Descent for Training Physics-Informed Neural Networks (NeurIPS 2024 · similarity 0.90)
- RoPINN: Region Optimized Physics-Informed Neural Networks (NeurIPS 2024 · similarity 0.90)
- Sub-Sequential Physics-Informed Learning with State Space Model (ICML 2025 · similarity 0.90)
- Challenges in Training PINNs: A Loss Landscape Perspective (ICML 2024 · similarity 0.89)
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.
