Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Randomly Projected Additive Gaussian Processes for Regression
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 input-space ARD pre-scaling before projecting (ablation) | DPA-GP · UCI airfoil | RMSE: 0.44 ± 0.07 -> 0.31 ± 0.08 | Helped |
| Add input-space ARD pre-scaling before projecting (ablation) | DPA-GP · UCI housing | RMSE: 0.41 ± 0.18 -> 0.34 ± 0.13 | Helped |
| Choose diverse directions by minimizing projection redundancy (ablation) | RPA-GP-1 · UCI airfoil | RMSE: 0.44 ± 0.08 -> 0.44 ± 0.07 | Helped |
| Use 20 additive random projections rather than one random projection (ablation) | Single-RP · UCI concreteslump | RMSE: 0.99 ± 0.59 -> 0.10 ± 0.09 | Helped |
| Use SKI inference (ablation) | DPA-GP-ARD · UCI housing | RMSE: 0.34 ± 0.13 -> 0.38 ± 0.17 | No clear effect |
| Add more random projections and higher-degree sub-kernels (ablation) | RPA-GP · UCI benchmarks | RMSE | Helped |
| Increase the number of 1D projections (ablation) | RPA-GP · Regression tasks in the convergence study | RMSE | Helped |
| Apply ARD to prune irrelevant input features before projection (ablation) | DPA-GP · Synthetic regression with irrelevant features | RMSE | Helped |
| Use an axis-aligned fully additive kernel (ablation) | DPA-GP methods · Additive synthetic regression | RMSE | Mixed |
| Use DPA-GP-ARD rather than RBF-ARD or GAM-GP | High-dimensional Olivetti face orientation regression · Olivetti faces | RMSE | Helped |
| Exploit additive projections to use SKI instead of Cholesky inference | RPA-GP · Synthetic regression with 100 input dimensions | training runtime scaling | Helped |
Closest papers by meaning
- Log-Linear-Time Gaussian Processes Using Binary Tree Kernels (NeurIPS 2022 · similarity 0.89)
- High-Dimensional Gaussian Process Inference with Derivatives (ICML 2021 · similarity 0.89)
- Bezier Gaussian Processes for Tall and Wide Data (NeurIPS 2022 · similarity 0.88)
- Additive Gaussian Processes Revisited (ICML 2022 · similarity 0.88)
- Gaussian Process Neural Additive Models (AAAI 2024 · similarity 0.86)
- 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.