Research map / Probabilistic and scientific modeling / Differential equations and PDEs
Learning-Guided Integration Contours Construction for Fast Large-Scale Generalized Eigensolvers
Yeqiu Chen, Ziyan Liu, Hong Wang, Lei Liu
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 ENO (FNO) instead of MLP spectral predictor (ablation) | Deepcontour · Kirchhoff-Love Plate (N=50000) | missed eigenvalues: 32.4 -> 0 | Improved |
| add ENO (FNO) instead of MLP spectral predictor (ablation) | Deepcontour · Kirchhoff-Love Plate (N=50000) | CIRR solve time (s): 27.4 -> 25.2 | Improved |
| add KDE adaptive contouring instead of naive interval expanding (ablation) | Deepcontour · Kirchhoff-Love Plate (N=50000) | CIRR solve time (s): 44.5 -> 25.2 | Improved |
| add KDE adaptive contouring (coverage) (ablation) | Deepcontour · Kirchhoff-Love Plate (N=50000) | missed eigenvalues: 0 -> 0 | No clear effect |
| use Deepcontour instead of Arnoldi scouting | CIRR solver · Kirchhoff-Love Plate (tol 1e-2) | end-to-end speedup: 5.63x | Improved |
| tighten tolerance from 1e-2 to 1e-12 | Deepcontour vs Arnoldi · Kirchhoff-Love Plate | end-to-end speedup: 5.63 -> 3.45 | Hurt |
| increase matrix size | Deepcontour · Kirchhoff-Love Plate and EGFR Electronic | speedup vs Arnoldi and GD | Improved |
Closest papers by meaning
- STNet: Spectral Transformation Network for Solving Operator Eigenvalue Problem (NeurIPS 2025 · similarity 0.84)
- Eigendecomposition-Free Training of Deep Networks for Linear Least-Square Problems (TPAMI 2020 · similarity 0.82)
- Free Decompression with Algebraic Spectral Curves (NeurIPS 2026 · similarity 0.82)
- Eigen Analysis of Conjugate Kernel and Neural Tangent Kernel (ICML 2025 · similarity 0.81)
- Kronecker-Factored Approximate Curvature for Physics-Informed Neural Networks (NeurIPS 2024 · similarity 0.81)
- Learning to Discover Iterative Spectral Algorithms (NeurIPS 2026 · similarity 0.81)
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.
