Research map / Architectures and efficiency / RNNs and normalization
Neural-HSS: Hierarchical Semi-Separable Neural PDE Solver
Pietro Sittoni, Emanuele Zangrando, Angelo Alberto Casulli, Nicola Guglielmi, Francesco Tudisco
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 |
|---|---|---|---|
| use Neural-HSS instead of FNO | FNO · Heat equation (1D) | test error: 8 x 10^-3 -> 3 x 10^-3 | Improved |
| use Neural-HSS instead of FNO | FNO · Poisson equation (2D) | test error: 7 x 10^-6 -> 7 x 10^-8 | Improved |
| use Neural-HSS instead of FNO | FNO · Burgers' equation (1D) | test error: 0.26 -> 0.12 | Improved |
| use Neural-HSS instead of ResNet | ResNet · Gray-Scott forward (2D) | test error: 0.273 -> 0.294 | Hurt |
| use Neural-HSS instead of ResNet | ResNet · Gray-Scott inverse (2D) | test error: 0.193 -> 0.203 | Hurt |
| use Neural-HSS instead of FNO | FNO · Navier-Stokes forward (2D) | test error: 0.483 -> 0.123 | Improved |
| use Neural-HSS instead of FNO | FNO · Navier-Stokes inverse (2D) | test error: 0.19 -> 0.208 | Hurt |
| train Neural-HSS with 16 samples vs FNO with 64 samples | Neural-HSS · Poisson 3D data efficiency | test error: 16 samples matches FNO at 64 and ResNet at 128 | Improved |
| use full-rank lifting and projection (ablation) | Neural-HSS · Burgers' equation (1D) | test error | Improved |
| shift Burgers' equation toward elliptic (smaller beta) | Neural-HSS · Modified Burgers' equation | trajectory test loss | Improved |
| vary rank r, outer rank r_out and levels t over grid (ablation) | Neural-HSS · Gray-Scott | test error | No clear effect |
| use non-rectangular or mixed-BC domains | Neural-HSS · Poisson L-shape domain | test error: 2 x 10^-1 -> 2 x 10^-2 | Improved |
Closest papers by meaning
- MC^2: Monte Carlo Correction for Fast Elliptic PDE Solving (NeurIPS 2026 · similarity 0.90)
- Unisolver: PDE-Conditional Transformers Towards Universal Neural PDE Solvers (ICML 2025 · similarity 0.89)
- Mechanistic PDE Networks for Discovery of Governing Equations (ICML 2025 · similarity 0.89)
- BENO: Boundary-embedded Neural Operators for Elliptic PDEs (ICLR 2024 · similarity 0.89)
- Breakeven complexity: A new perspective on neural partial differential equation solvers (NeurIPS 2026 · similarity 0.89)
- UGrid: An Efficient-And-Rigorous Neural Multigrid Solver for Linear PDEs (ICML 2024 · similarity 0.88)
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.
