Research map / Continual learning / Online and replay-based
Streaming Sparse Gaussian Process Approximations
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 streaming sparse variational free energy (SSGP) instead of GP, SGP, or SVI | Time-series regression · Pseudo periodic and TIMIT time-series | marginal log-likelihood | Helped |
| Use streaming sparse variational free energy (SSGP) instead of GP, SGP, or SVI | Time-series regression · Pseudo periodic and TIMIT time-series | RMSE | Helped |
| Increase the number of pseudo-points (ablation) | SSGP · Time-series regression | accuracy | Helped |
| Use SSGP instead of GP or SGP | Regression · OS Terrain 50 | RMSE | Helped |
| Use SSGP instead of GP or SGP | Regression · OS Terrain 50 | marginal log-likelihood | Hurt |
| Use SSGP instead of batch sparse variational inference | Sparse GP classification · banana | performance | No clear effect |
| Use higher alpha values instead of the variational limit (ablation) | Online alpha-divergence SSGP · Toy online regression | test log-likelihood | Hurt |
| Set alpha to 0.01 (ablation) | Online alpha-divergence SSGP · Pseudo periodic dataset | test log-likelihood: alpha=0.01 is virtually identical to VFE | No clear effect |
| Vary the SVI learning rate among 0.001, 0.01, and 0.1 (ablation) | SVI · Time-series regression | performance | No clear effect |
| Apply vanilla streaming variational Bayes for hyperparameter learning (ablation) | Streaming GP · Hyperparameter learning | performance | Hurt |
Closest papers by meaning
- Sparse Gaussian Process Hyperparameters: Optimize or Integrate? (NeurIPS 2022 · similarity 0.91)
- Sparse Gaussian Processes: Structured Approximations and Power-EP Revisited (NeurIPS 2025 · similarity 0.90)
- Computation-Aware Gaussian Processes: Model Selection And Linear-Time Inference (NeurIPS 2024 · similarity 0.89)
- Turbocharging Gaussian Process Inference with Approximate Sketch-and-Project (NeurIPS 2025 · similarity 0.89)
- New Bounds for Sparse Variational Gaussian Processes (ICML 2025 · similarity 0.88)
- Robust Gaussian Processes via Relevance Pursuit (NeurIPS 2024 · 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.