Research map / Optimization / Optimizers
Factorization-in-Loop:Proximal Fill-in Minimization for Sparse Matrix Reordering
Ziwei Li, Shuzi Niu, Tao Yuan, Huiyuan Li, Wenjia Wu
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 pretrained spectral embedding (vs random init) (ablation) | PFM · SuiteSparse SP | fill-in ratio: 123.1 -> 45.32 | Improved |
| use factorization loss instead of pairwise cross entropy (ablation) | PFM · SuiteSparse SP+CFD | fill-in ratio: 54.46 -> 51.81 | Improved |
| use factorization loss instead of UDNO envelope loss (ablation) | PFM · SuiteSparse SP+CFD | fill-in ratio: 53.12 -> 51.81 | Improved |
| use MgGNN encoder instead of GraphUnet (ablation) | PFM · SuiteSparse SP+CFD | fill-in ratio: 89.56 -> 51.81 | Improved |
| PFM vs UDNO overall | UDNO · SuiteSparse all | fill-in ratio: 86.29 -> 86.14 | Improved |
| PFM vs UDNO overall | UDNO · SuiteSparse all | LU factorization time (s): 57.49 -> 48.80 | Improved |
| PFM vs UDNO on CFD | UDNO · SuiteSparse CFD | LU factorization time (s): 16.36 -> 17.59 | Hurt |
| PFM vs UDNO on Other matrices | UDNO · SuiteSparse Other | fill-in ratio: 154.82 -> 157.65 | Hurt |
| PFM vs Metis on Other matrices | Metis · SuiteSparse Other | fill-in ratio: 146.68 -> 157.65 | Hurt |
| PFM vs Metis overall | Metis · SuiteSparse all | LU factorization time (s): 123.52 -> 48.80 | Improved |
Closest papers by meaning
- Causal Structure Learning for Sparse Matrix Fill-in Reduction (ICML 2026 · similarity 0.87)
- On the Crucial Role of Initialization for Matrix Factorization (ICLR 2025 · similarity 0.86)
- Two Sparse Matrices are Better than One: Sparsifying Neural Networks with Double Sparse Factorization (ICLR 2025 · similarity 0.85)
- Provable Acceleration of Nesterov's Accelerated Gradient for Asymmetric Matrix Factorization and Linear Neural Networks (NeurIPS 2024 · similarity 0.85)
- In-depth Analysis of Low-rank Matrix Factorisation in a Federated Setting (AAAI 2025 · similarity 0.85)
- Symmetric Matrix Completion with ReLU Sampling (ICML 2024 · similarity 0.85)
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.
