Research map / Learning theory and algorithms / Matrix and tensor methods
The Variational Nystrom method for large-scale spectral problems
Max Vladymyrov, Miguel Carreira-Perpinan
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 Variational Nystrom instead of Nystrom with 25 s runtime limit | spectral clustering · BSDS500 image segmentation | closeness to exact segmentation | Improved |
| apply approximation directly to graph Laplacian instead of custom normalized out-of-sample kernel (ablation) | Nystrom, CS, VN · Swiss roll 5000 points | error vs exact LE | Hurt |
| Sum normalization of out-of-sample kernel for VN (ablation) | Variational Nystrom · Swiss roll 5000 points | error vs exact LE | Improved |
| CC normalization (sum all columns of C) for column sampling (ablation) | Column Sampling · Swiss roll 5000 points | error vs exact LE | Improved |
| different normalizations for Nystrom (ablation) | Nystrom · Swiss roll 5000 points | error vs exact LE | No clear effect |
| more power iterations q in random projections (ablation) | RP · Swiss roll / MNIST | error vs exact LE | Improved |
| k-means landmark selection instead of random (ablation) | Variational Nystrom · MNIST 20000 | error | No clear effect |
| VN vs Nystrom/CS/LLL ranking by error | landmark methods · MNIST 20000, d=10 | error vs exact LE | Improved |
Closest papers by meaning
- A Novel Normalized-Cut Solver With Nearest Neighbor Hierarchical Initialization (TPAMI 2023 · similarity 0.89)
- Randomized Clustered Nystrom for Large-Scale Kernel Machines (AAAI 2018 · similarity 0.88)
- Improved guarantees and a multiple-descent curve for Column Subset Selection and the Nystrom method (NeurIPS 2020 · similarity 0.88)
- Revisiting the Nyström Method for Improved Large-scale Machine Learning (JMLR 2016 · similarity 0.88)
- Grassmannian Manifold Optimization Assisted Sparse Spectral Clustering (CVPR 2017 · similarity 0.88)
- Towards More Efficient SPSD Matrix Approximation and CUR Matrix Decomposition (JMLR 2016 · 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.
