Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Scalable Log Determinants for Gaussian Process Kernel Learning
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 |
|---|---|---|---|
| Use stochastic Lanczos log-determinant estimation instead of scaled eigenvalues | GP with RBF kernel · Daily precipitation | MSE: 0.621 -> 0.613 | Helped |
| Use stochastic Lanczos log-determinant estimation instead of scaled eigenvalues | Log-Gaussian Cox process · Hickory | -log p(y|θ): 1851.69 -> 1828.07 | Helped |
| Use stochastic Lanczos log-determinant estimation instead of scaled eigenvalues | Log-Gaussian Cox process · Hickory | Time [s]: 2.5 -> 21.4 | Hurt |
| Use stochastic Lanczos log-determinant estimation instead of scaled eigenvalues | Log-Gaussian Cox process · Chicago assault | RMSE_test: 1.36 -> 1.33 | Helped |
| Use stochastic Lanczos log-determinant estimation instead of scaled eigenvalues | Deep kernel learning · Gas sensor | RMSE: 0.1045 ± 0.0228 -> 0.1053±0.0248 | No clear effect |
| Add diagonal correction to the SKI kernel approximation (ablation) | Lanczos and Chebyshev with SKI · Synthetic Matérn 3/2 prediction | predictive uncertainty agreement with FITC | Helped |
| Remove diagonal correction from the SKI kernel approximation (ablation) | Lanczos and Chebyshev with SKI · Synthetic Matérn 3/2 prediction | predictive uncertainty calibration | Hurt |
| Use Lanczos instead of Chebyshev for the log-determinant approximation (ablation) | Stochastic GP log-determinant estimation · RBF and Matérn kernel experiments | approximation accuracy | Helped |
Closest papers by meaning
- Log-Linear-Time Gaussian Processes Using Binary Tree Kernels (NeurIPS 2022 · similarity 0.91)
- Scalable Gaussian Processes with Grid-Structured Eigenfunctions (GP-GRIEF) (ICML 2018 · similarity 0.90)
- High-Dimensional Gaussian Process Inference with Derivatives (ICML 2021 · similarity 0.89)
- On the Identifiability and Interpretability of Gaussian Process Models (NeurIPS 2023 · similarity 0.89)
- Scalable Derivative Gaussian Processes via Exact Gradient Reduction (NeurIPS 2026 · similarity 0.87)
- Determinant Estimation under Memory Constraints and Neural Scaling Laws (ICML 2025 · similarity 0.87)
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.