Research map / Probabilistic and scientific modeling / Differential equations and PDEs
Interpretability and Generalization Bounds for Learning Spatial Physics
Alejandro Francisco Queiruga, Theo Gutman-Solo, Shuai Jiang
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 higher polynomial degree p to training data | finite-difference parameter fit · 1D Poisson parameter recovery | relative parameter error | Hurt |
| add higher polynomial degree p to training data | linear model u=Wf · 1D Poisson OOD generalization | test MSE | Improved |
| replace polynomial training data with random piecewise linear (FEM) functions | linear model · Green's function recovery | recovery of true operator | Improved |
| add physics-informed loss | DeepONet · 1D Poisson OOD generalization | subspace generalization: 10^{-6} | No clear effect |
| add measurement noise to training data | linear model · 1D Poisson OOD generalization | lowest MSE: 10^{-20} -> 10^{-9} | Hurt |
| switch from linear model to MLP | black-box operator · 1D Poisson OOD generalization | subspace generalization | Hurt |
| switch from polynomial to FEM piecewise linear test data | finite difference model · 1D Poisson OOD generalization | test MSE | Hurt |
Closest papers by meaning
- Understanding Generalization in Physics Informed Models through Affine Variety Dimensions (NeurIPS 2025 · similarity 0.86)
- Refined generalization analysis of the Deep Ritz Method and Physics-Informed Neural Networks (ICML 2025 · similarity 0.85)
- Generalization error of spectral algorithms (ICLR 2024 · similarity 0.84)
- Towards Cross Domain Generalization of Hamiltonian Representation via Meta Learning (ICLR 2024 · similarity 0.83)
- Physics-Informed Diffusion Models (ICLR 2025 · similarity 0.83)
- Out-of-Domain Generalization in Dynamical Systems Reconstruction (ICML 2024 · 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.
