Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Sparse within Sparse Gaussian Processes using Neighbor Information
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 a sparsity-inducing prior to activate nearest inducing inputs for each data point | Sparse variational Gaussian processes · UCI regression and classification benchmarks | prediction quality | Helped |
| Use nearest-neighbor sparsity during training, rather than only selecting neighbors at prediction time | SVGP with M inducing inputs and H-neighbor prediction · UCI benchmarks | prediction quality | Helped |
| Use SWSGP with 4 neighbors instead of SVGP | EEG, M=256 · EEG | training iteration time (ms): 21.42 -> 26.18 | Hurt |
| Use SWSGP with 4 neighbors instead of SVGP | EEG · EEG | test time (ms): 1.43 -> 0.56 | Helped |
| Use SWSGP with 4 neighbors instead of SVGP | MNIST · MNIST odd/even classification | training iteration time (ms): 516 -> 233 | Helped |
| Use SWSGP with 4 neighbors instead of SVGP | MNIST · MNIST odd/even classification | error rate: 0.02 -> 0.016 | Helped |
| Use SWSGP with 4 neighbors instead of SVGP | MNIST · MNIST odd/even classification | MNLL: 0.066 -> 0.05 | Helped |
| Use SWSGP with a very large inducing set and fixed inducing locations | SWSGP · Airline | error rate: 34% -> 21% | Helped |
| Use SWSGP with a very large inducing set and fixed inducing locations | SWSGP · Airline | MNLL: 0.61 -> 0.48 | Helped |
| Increase active neighbors from 4 to 8 (ablation) | SWSGP-64 · Powerplant | RMSE: 4.27 -> 4.24 | Helped |
| Increase active neighbors from 4 to 8 (ablation) | SWSGP-64 · Kin | RMSE: 0.11 -> 0.10 | Helped |
| Increase active neighbors from 4 to 8 (ablation) | SWSGP-64 · Kin | MNLL: -1.27 -> -1.38 | Helped |
Closest papers by meaning
- Input Dependent Sparse Gaussian Processes (ICML 2022 · similarity 0.95)
- Variational nearest neighbor Gaussian process (ICML 2022 · similarity 0.91)
- Sparse Gaussian Processes: Structured Approximations and Power-EP Revisited (NeurIPS 2025 · similarity 0.91)
- New Bounds for Sparse Variational Gaussian Processes (ICML 2025 · similarity 0.91)
- Robust Gaussian Processes via Relevance Pursuit (NeurIPS 2024 · similarity 0.89)
- Sparse Variational Student-t Processes (AAAI 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.