Research map / Probabilistic and scientific modeling / Differential equations and PDEs
Physics-Informed Inference Time Scaling for Solving High-Dimensional Partial Differential Equations
Zexi Fan, Yan Sun, Shihao Yang, Yiping Lu
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 simulation-based defect correction (SCaSML) to the PINN surrogate | PINN surrogate · LCD 10d | relative L2 error: 5.20E-02 -> 2.74E-02 | Improved |
| add SCaSML correction | PINN surrogate · LCD 60d | relative L2 error: 3.13E-01 -> 1.32E-01 | Improved |
| add SCaSML correction | PINN surrogate · VB-PINN 20d | relative L2 error: 1.17E-02 -> 4.03E-03 | Improved |
| add SCaSML correction | GP surrogate · VB-GP 20d | relative L2 error: 1.47E-01 -> 6.23E-02 | Improved |
| add SCaSML correction | GP surrogate · VB-GP 80d | relative L2 error: 2.66E-01 -> 1.52E-01 | Improved |
| add SCaSML correction | PINN surrogate · LQG 100d-160d | relative L2 error: 7.97E-02 -> 5.53E-02 | Improved |
| use SCaSML instead of naive MLP solver | naive MLP · LQG 100d | relative L2 error: 5.63E+00 -> 5.53E-02 | Improved |
| add SCaSML correction to an already accurate PINN | PINN surrogate · Diffusion-reaction 100d-160d | relative L2 error: 1.41E-02 -> 1.11E-02 | Improved |
| add SCaSML correction | PINN surrogate · Diffusion-reaction 160d | relative L2 error: 3.45E-02 -> 3.22E-02 | Improved |
| add inference-time simulation correction (runtime cost) | PINN surrogate · LCD 10d | total runtime (s): 0.45 -> 13.31 | Hurt |
| add SCaSML to GP surrogate for convergence rate | GP surrogate · Burgers d=20 | convergence rate gamma: 0.37 -> 0.55 | Improved |
| add SCaSML to GP surrogate for convergence rate | GP surrogate · Burgers d=80 | convergence rate gamma: 0.35 -> 0.50 | Improved |
Closest papers by meaning
- Compute-optimal data scaling for neural surrogates via multi-fidelity training (NeurIPS 2026 · similarity 0.90)
- Learning a Neural Solver for Parametric PDEs to Enhance Physics-Informed Methods (ICLR 2025 · similarity 0.88)
- Ultra Fast PDE Solving via Physics Guided Few-step Diffusion (NeurIPS 2026 · similarity 0.88)
- SBNO : Schrödinger Bridge Neural Operator for the Forward and Inverse Problems under Missing Information (NeurIPS 2026 · similarity 0.88)
- Calibrated Physics-Informed Uncertainty Quantification (ICML 2025 · similarity 0.87)
- P$^2$C$^2$Net: PDE-Preserved Coarse Correction Network for efficient prediction of spatiotemporal dynamics (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.
