Research map / Optimization / Optimizers
Preconditioning Kernel Matrices
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 low-rank (Nystrom, FITC, PITC) preconditioner to CG for solving kernel systems | PCG for GP kernel systems · UCI regression (Concrete, Power Plant, Protein) | CG iterations to converge: significant reductions in iterations, often ~an order of magnitude | Helped |
| Add Block Jacobi preconditioner | PCG for GP kernel systems · UCI regression (Concrete, Power Plant, Protein) | CG iterations to converge | No clear effect |
| Add regularization (diagonal offset delta) preconditioner with flexible CG | PCG for GP kernel systems · UCI regression (Concrete, Power Plant, Protein) | CG iterations to converge | No clear effect |
| Add SKI-based Kronecker preconditioner | PCG for GP kernel systems · UCI regression (Concrete, Power Plant) | CG iterations to converge | No clear effect |
| Use Nystrom preconditioner with m=4 sqrt(n) random inducing points (vs plain CG) in stochastic-gradient GP hyperparameter learning | GP regression and classification hyperparameter learning · GP regression/classification (Spam, EEG, Power Plant, Protein) | RMSE / NLL / error rate over time | Helped |
| Preconditioning Nystrom-type approximations vs exact Cholesky for modest n | GP learning with PCG · modestly sized datasets | accuracy vs computation | No clear effect |
Closest papers by meaning
- Preconditioning for Scalable Gaussian Process Hyperparameter Optimization (ICML 2022 · similarity 0.91)
- Kernel Methods Through the Roof: Handling Billions of Points Efficiently (NeurIPS 2020 · similarity 0.90)
- Fast Training of Large Kernel Models with Delayed Projections (NeurIPS 2025 · similarity 0.89)
- Randomized Block-Diagonal Preconditioning for Parallel Learning (ICML 2020 · similarity 0.89)
- Learning Sparse Approximate Inverse Preconditioners for Conjugate Gradient Solvers on GPUs (NeurIPS 2025 · similarity 0.88)
- Spectral Preconditioning for Gradient Methods on Graded Non-convex Functions (ICML 2024 · 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.