Research map / Probabilistic and scientific modeling / Differential equations and PDEs
Overcoming PINNs Failure Modes In High Dimension With Low-Rank Fourier Sum
Natan Kaminsky, Daniel Freedman, Kira Radinsky
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 neural network with low-rank separable Fourier sum and analytic loss | PINN · Convection PDE | rMAE: 0.0059 -> 8.2E-06 (vs PINN FP64) | Improved |
| replace neural network with LoRFS | PINN · Reaction PDE | rMAE: 0.0092 -> 0.0002 (vs PINNMamba) | Improved |
| replace neural network with LoRFS | PINN · Wave PDE | rMAE: 0.0080 -> 0.0033 (vs PINN FP64) | Improved |
| use LoRFS in high dimensions instead of PINN FP64 or PINNMamba | PINN solvers · High-dimensional Wave | rRMSE: 0.8584 -> 5.9E-05 (vs PINN FP64) | Improved |
| use LoRFS in high dimensions | PINN solvers · High-dimensional Reaction | rRMSE: 1.0001 -> 0.0082 (vs PINN FP64) | Improved |
| use analytic differentiation of Fourier series instead of stochastic or autodiff high-order estimators | biharmonic solver · Biharmonic equation | rMAE: 7.8E-03 -> 4.6E-04 (vs Forward Laplacian) | Improved |
| use stochastic Monte Carlo estimators (HTE, STDE) instead of deterministic Forward Laplacian | biharmonic solver · Biharmonic equation | rMAE: 7.8E-03 -> 1.97E-02 / 2.29E-02 | Hurt |
| increase rank s and Fourier degree K beyond a threshold (ablation) | LoRFS · Reaction PDE sensitivity | rRMSE: threshold K > 5 and s > 5 | No clear effect |
| use a slower PINN architecture (PINNMamba) within a fixed 1 hour budget in high dimension | PINNMamba · High-dimensional PDEs | rRMSE | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- PINNsFormer — PINNsFormer: A Transformer-Based Framework For Physics-Informed Neural Networks (ICLR 2024) · better on 2 of 2 shared results
- STDE — STDE++: Polynomial-Time Amortization for Linear Differential Operators (JMLR 2026) · better on 1 of 1 shared result
- RoPINN — RoPINN: Region Optimized Physics-Informed Neural Networks (NeurIPS 2024) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Efficient Error Certification for Physics-Informed Neural Networks (ICML 2024 · similarity 0.88)
- PINNacle: A Comprehensive Benchmark of Physics-Informed Neural Networks for Solving PDEs (NeurIPS 2024 · similarity 0.87)
- How does PDE order affect the convergence of PINNs? (NeurIPS 2024 · similarity 0.87)
- Sub-Sequential Physics-Informed Learning with State Space Model (ICML 2025 · similarity 0.86)
- S$^{2}$-PINN: Stochastic Separable Physics-Informed Neural Networks (NeurIPS 2026 · similarity 0.86)
- Per-Loss Adapters for Gradient Conflict in Physics-Informed Neural Networks (NeurIPS 2026 · 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.
