Research map / Learning theory and algorithms / Matrix and tensor methods
Towards More Efficient SPSD Matrix Approximation and CUR Matrix Decomposition
Shusen Wang, Zhihua Zhang, Tong Zhang
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 fast model with sketch size s = 2c instead of Nystrom method | SPSD kernel approximation · RBF kernel approximation (LIBSVM datasets) | Frobenius approximation error | Improved |
| increase sketch size s up to 0.2n (ablation) | fast SPSD model · RBF kernel approximation (LIBSVM datasets) | Frobenius approximation error | Improved |
| use leverage score sampling instead of uniform sampling for sketch matrix S (ablation) | fast SPSD model · RBF kernel approximation (LIBSVM datasets) | Frobenius approximation error | No clear effect |
| select C by uniform+adaptive^2 sampling instead of uniform sampling (ablation) | SPSD approximation · RBF kernel approximation (LIBSVM datasets) | Frobenius approximation error | Improved |
| use fast model instead of Nystrom under same time or memory | approximate KPCA · KPCA eigenvector misalignment (k=3) | misalignment | Improved |
| use fast model instead of Nystrom under same time | KPCA + KNN classification · classification datasets (k=3 and k=10) | classification error: one to two percent of classification error | Improved |
| use fast model instead of Nystrom or prototype under same time | approximate spectral clustering · spectral clustering datasets | NMI | No clear effect |
Closest papers by meaning
- SPSD Matrix Approximation vis Column Selection: Theories, Algorithms, and Extensions (JMLR 2016 · similarity 0.91)
- Revisiting the Nyström Method for Improved Large-scale Machine Learning (JMLR 2016 · similarity 0.90)
- The Variational Nystrom method for large-scale spectral problems (ICML 2016 · similarity 0.88)
- Fast Deterministic CUR Matrix Decomposition with Accuracy Assurance (ICML 2020 · similarity 0.87)
- Entrywise error bounds for low-rank approximations of kernel matrices (NeurIPS 2024 · similarity 0.87)
- Fast exact recovery of noisy matrix from few entries: the infinity norm approach (NeurIPS 2025 · similarity 0.84)
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.
