Research map / Learning theory and algorithms / Matrix and tensor methods
Revisiting the Nyström Method for Improved Large-scale Machine Learning
Alex Gittens, Michael W. Mahoney
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 leverage-score sampling instead of uniform Nystrom | SPSD sketch · Enron, k=60 (l=k ln n) | spectral norm error ratio: 1.386 -> 1.042 | Improved |
| use leverage-score sampling instead of uniform Nystrom | SPSD sketch · AbaloneD sigma=.15, k=20 (l=k ln n) | spectral norm error ratio: 2.204 -> 0.908 | Improved |
| use leverage-score sampling instead of uniform Nystrom | SPSD sketch · Protein, k=10 (l=k ln n) | spectral norm error ratio: 1.350 -> 0.994 | Improved |
| use leverage-score sampling instead of uniform Nystrom | SPSD sketch · Protein, k=10 (l=k ln n) | trace norm error ratio: 0.764 -> 0.765 | No clear effect |
| use Gaussian projection instead of uniform Nystrom | SPSD sketch · AbaloneD sigma=.15, k=20 (l=k ln n) | spectral norm error ratio: 2.204 -> 1.822 | Improved |
| increase samples from l=k+8 to l=k ln n for leverage sketch | leverage sketch · Enron, k=60 | spectral norm error ratio: 1.381 -> 1.042 | Improved |
| increase samples to l=k ln n for uniform Nystrom | uniform Nystrom · Enron, k=60 | spectral norm error ratio: 1.386 -> 1.386 | No clear effect |
| use Gaussian-based bounds vs prior bounds for predicting error | Gaussian sketch bounds · Enron, k=60 | pred./obs. spectral error: 20.1 vs 3041.0 | Improved |
Closest papers by meaning
- Towards More Efficient SPSD Matrix Approximation and CUR Matrix Decomposition (JMLR 2016 · similarity 0.90)
- Improved guarantees and a multiple-descent curve for Column Subset Selection and the Nystrom method (NeurIPS 2020 · similarity 0.90)
- Recursive Sampling for the Nystrom Method (NeurIPS 2017 · similarity 0.90)
- Entrywise error bounds for low-rank approximations of kernel matrices (NeurIPS 2024 · similarity 0.88)
- Spectral Perturbation Bounds for Low-Rank Approximation with Applications to Privacy (NeurIPS 2025 · similarity 0.87)
- Symmetric Matrix Completion with ReLU Sampling (ICML 2024 · 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.
