Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Conjugate Gradients for Kernel Machines
Simon Bartels, Philipp Hennig
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 KMCG kernel approximation (CG on K without noise, substituting approximated kernel for k_*) instead of plain CG | conjugate gradients · common regression datasets, SE kernel | average relative error of posterior mean | Improved |
| compare KMCG with FITC and VFE baselines at equal cost per step (M=sqrt(NP)) | KMCG · common regression datasets, SE kernel | relative error of mean, variance, evidence | Hurt |
| use KMCG instead of FITC/VFE when matrix-vector products are cheaper than O(N^2) | KMCG · artificial grid-structured datasets, SE kernel | relative error of posterior mean | Improved |
| use KMCG instead of FITC/VFE for posterior variance on grid datasets | KMCG · artificial grid-structured datasets, SE kernel | relative error of variance | Hurt |
| use KMCG instead of baselines for evidence on grid datasets | KMCG · artificial grid-structured datasets, SE kernel | relative error of evidence | No clear effect |
| use KMCG instead of baselines on a Toeplitz-like missing-data audio task | KMCG · Natural Sound Modeling (SOUND) | SMSE | Improved |
Closest papers by meaning
- Overfitting Behaviour of Gaussian Kernel Ridgeless Regression: Varying Bandwidth or Dimensionality (NeurIPS 2024 · similarity 0.87)
- A Comprehensive Analysis on the Learning Curve in Kernel Ridge Regression (NeurIPS 2024 · similarity 0.87)
- Learning Curves of Stochastic Gradient Descent in Kernel Regression (ICML 2025 · similarity 0.87)
- Kernel Regression in Structured Non-IID Settings: Theory and Implications for Denoising Score Learning (NeurIPS 2025 · similarity 0.87)
- Generalization in Kernel Regression Under Realistic Assumptions (ICML 2024 · similarity 0.86)
- Generalization error of spectral algorithms (ICLR 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.
