Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Efficient Computation of Gaussian Process Regression for Large Spatial Data Sets by Patching Local Gaussian Processes
Chiwoo Park, Jianhua Z. Huang
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 patched GP instead of GMRF | GMP baseline: GMRF · TCO.L2 (N=182,591) | computation time (s): 5702.7 -> 807.4 | Improved |
| use patched GP instead of GMRF | GMRF · TCO.L2 | NLPD: 5.26 -> 3.70 | Improved |
| use patched GP instead of GMRF | GMRF · TCO.L2 | RMSE: 9.24 -> 6.74 | Improved |
| use patched GP (best setting) instead of GMRF | GMRF · TCO | RMSE: 2.78 -> 2.85 | Hurt |
| use patched GP instead of GMRF | GMRF · ICETHICK | NLPD: 7.08 -> 6.12 | Improved |
| refine finite-element mesh (h to zero) in patched GP (ablation) | patched GP · synthetic-1d | mean squared difference from full GP: about e-6 | Improved |
| vary number of local regions S (ablation) | patched GP · synthetic-1d, synthetic-2d, TCO, ICETHICK | prediction accuracy | No clear effect |
| use larger S in patched GP (ablation) | patched GP · TCO, ICETHICK | computation cost | Improved |
| use patched GP instead of DDM with longer computation time | DDM · TCO, TCO.L2, ICETHICK | MSE | Improved |
| use patched GP instead of DDM | DDM · TCO, TCO.L2 | NLPD | Hurt |
| use patched GP instead of DDM | DDM · ICETHICK | NLPD | Improved |
| use patched GP boundary estimation instead of DDM | DDM · synthetic-2d boundary (S=16) | mean squared difference to full GP | Improved |
Closest papers by meaning
- Patchwork Kriging for Large-scale Gaussian Process Regression (JMLR 2018 · similarity 0.95)
- Jump Gaussian Process Model for Estimating Piecewise Continuous Regression Functions (JMLR 2022 · similarity 0.88)
- Scalable Gaussian-process regression and variable selection using Vecchia approximations (JMLR 2022 · similarity 0.88)
- Learning from Summarized Data: Gaussian Process Regression with Sample Quasi-Likelihood (AAAI 2025 · similarity 0.87)
- Robust and Conjugate Spatio-Temporal Gaussian Processes (ICML 2025 · similarity 0.87)
- Turbocharging Gaussian Process Inference with Approximate Sketch-and-Project (NeurIPS 2025 · 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.
