Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Sparse Gaussian Processes with Spherical Harmonic Features
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 spherical harmonic inducing features instead of standard inducing points | Sparse GP · Energy, Concrete, Power | MSE: Energy: 0.013 ± 0.00 -> 0.003 ± 0.00; Concrete: 0.128 ± 0.01 -> 0.122 ± 0.02; Power: 0.055 ± 0.00 -> 0.054 ± 0.00 | Mixed |
| Use spherical harmonic features instead of additive VFF | Sparse GP · UCI regression datasets | MSE: Concrete: 0.123 ± 0.01 -> 0.122 ± 0.02; Kin8mn: 0.561 ± 0.02 -> 0.219 ± 0.01; Power: 0.058 ± 0.00 -> 0.054 ± 0.00 | Helped |
| Increase spherical harmonic feature count from M=9 to M=225 or M=784 (ablation) | Spherical harmonics · Banana classification | ELBO | Helped |
| Increase feature count from M=210 to M=660 (ablation) | Spherical harmonics · Airline delay (N=5,929,413) | NLPD: 1.29 ± 0.002 -> 1.27 ± 0.002 | Helped |
| Increase feature count from M=210 to M=660 (ablation) | Spherical harmonics · Airline delay (N=5,929,413) | Time (seconds): 41.32 ± 0.81 -> 160.8 ± 3.80 | Hurt |
| Use spherical harmonic inducing features instead of SVGP inducing points | Sparse GP · Airline delay (N=5,929,413) | NLPD: 1.31 ± 0.002 -> 1.29 ± 0.002 | Helped |
| Use spherical harmonic inducing features instead of SVGP inducing points | Sparse GP · Airline delay (N=5,929,413) | Time (seconds): 918.77 ± 1.21 -> 41.32 ± 0.81 | Helped |
| Use spherical harmonic inducing features instead of SVGP with an SE kernel | Sparse GP classifier · SUSY | AuC: 0.852 ± 0.002 -> 0.859 ± 0.001 | Helped |
| Use spherical harmonic GP instead of a five-layer neural network | SUSY classifier · SUSY | AuC: 0.867 ± 0.002 -> 0.859 ± 0.001 | Hurt |
Closest papers by meaning
- Spherical Inducing Features for Orthogonally-Decoupled Gaussian Processes (ICML 2023 · similarity 0.92)
- Sparse Gaussian Processes: Structured Approximations and Power-EP Revisited (NeurIPS 2025 · similarity 0.90)
- Sparse within Sparse Gaussian Processes using Neighbor Information (ICML 2021 · similarity 0.90)
- Orthogonally Decoupled Variational Gaussian Processes (NeurIPS 2018 · similarity 0.90)
- Spatio-Temporal Variational Gaussian Processes (NeurIPS 2021 · similarity 0.89)
- New Bounds for Sparse Variational Gaussian Processes (ICML 2025 · similarity 0.89)
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.