Research map / Probabilistic and scientific modeling / Differential equations and PDEs
Deep Hidden Physics Models: Deep Learning of Nonlinear Partial Differential Equations
Maziar Raissi
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 more noise to training data (1%, 2%, 5% vs clean) (ablation) | Deep hidden physics model · Burgers | relative L2 error: 4.78e-03 -> 2.64e-02 -> 1.09e-01 -> 4.56e-01 | Hurt |
| raise highest derivative order from 1st to 2nd (ablation) | Deep hidden physics model · Burgers | relative L2 error: 1.14e+00 -> 1.29e-02 | Improved |
| raise highest derivative order from 2nd to 3rd (ablation) | Deep hidden physics model · Burgers | relative L2 error: 1.29e-02 -> 3.42e-02 | Hurt |
| raise highest derivative order from 2nd to 4th (ablation) | Deep hidden physics model · Burgers | relative L2 error: 1.29e-02 -> 5.54e-02 | Hurt |
| make N depend on space x (ablation) | Deep hidden physics model · Burgers generalization to new IC | relative L2 error: 7.33e-02 -> 4.25e-01 | Hurt |
| make N depend on time t (ablation) | Deep hidden physics model · Burgers generalization to new IC | relative L2 error: 7.33e-02 -> 2.58e-01 | Hurt |
| make N depend on both time and space (ablation) | Deep hidden physics model · Burgers generalization to new IC | relative L2 error: 7.33e-02 -> 1.46e+00 | Hurt |
| test on unseen initial condition instead of training data (Burgers) | Deep hidden physics model · Burgers | relative L2 error: 4.78e-03 -> 7.33e-02 | Mixed |
| test on unseen initial condition instead of extrapolation (KdV) | Deep hidden physics model · KdV | relative L2 error: 6.28e-02 -> 3.44e-02 | Improved |
| evaluate KdV on training portion versus extrapolated region | Deep hidden physics model · KdV | relative L2 error: 3.78e-02 -> 6.28e-02 | Mixed |
Closest papers by meaning
- Physics-Informed DeepONets for drift-diffusion on metric graphs: simulation and parameter identification (ICML 2025 · similarity 0.87)
- Learning a Neural Solver for Parametric PDEs to Enhance Physics-Informed Methods (ICLR 2025 · similarity 0.86)
- A Riemannian Framework for Learning Reduced-order Lagrangian Dynamics (ICLR 2025 · similarity 0.86)
- Learning from Integral Losses in Physics Informed Neural Networks (ICML 2024 · similarity 0.86)
- Mechanistic PDE Networks for Discovery of Governing Equations (ICML 2025 · similarity 0.86)
- Text2PDE: Latent Diffusion Models for Accessible Physics Simulation (ICLR 2025 · similarity 0.86)
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.
