Research map / Probabilistic and scientific modeling / Bayesian optimization
Learning search spaces for Bayesian optimization: Another view of hyperparameter transfer learning
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 |
|---|---|---|---|
| Learn a bounding-box search space from related tasks | Random search, GP, and SMAC · SGD ridge regression | convergence speed | Helped |
| Learn an ellipsoidal or bounding-box search space from related tasks | Hyperband · SGD ridge regression | convergence speed | Helped |
| Use an ellipsoid instead of a bounding box (ablation) | Random search · OpenML binary classification | convergence speed | Helped |
| Add a learned bounding-box search space | Random search, SMAC, and GP · OpenML binary classification | convergence speed | Helped |
| Add a learned bounding-box search space (ablation) | GP warm-start · OpenML binary classification | convergence speed | Helped |
| Use slack variables to exclude outlier solutions (ablation) | Box and Ellipsoid search spaces · OpenML XGBoost | convergence speed | Helped |
| Use slack variables to exclude outlier solutions (ablation) | Box and Ellipsoid search spaces · OpenML SVM and random forest | convergence speed | No clear effect |
| Use a small historical sample to learn the search space (ablation) | Ellipsoid Random · OpenML binary classification | convergence speed | Helped |
| Use 9 related problems rather than only 3 (ablation) | Box Random and Ellipsoid Random · OpenML binary classification | convergence speed | Helped |
| Learn a bounding-box search space from related tasks | Random search, SMAC, and GP · 2-layer feed-forward neural network | convergence speed | Helped |
| Restrict optimization to the learned search space (ablation) | Search methods with 16 times as many iterations · OpenML binary classification and 2-layer feed-forward neural network | best-found performance | No clear effect |
| Use a log-scaled rather than unscaled search space (ablation) | Box and Ellipsoid transfer · OpenML binary classification | convergence speed | Mixed |
Closest papers by meaning
- Monte Carlo Tree Search based Space Transfer for Black Box Optimization (NeurIPS 2024 · similarity 0.91)
- Bayesian Optimisation with Unknown Hyperparameters: Regret Bounds Logarithmically Closer to Optimal (NeurIPS 2024 · similarity 0.91)
- Modeling All Response Surfaces in One for Conditional Search Spaces (AAAI 2025 · similarity 0.90)
- BONSAI: Bayesian Optimization with Natural Simplicity and Interpretability (NeurIPS 2026 · similarity 0.90)
- Large Language Models to Enhance Bayesian Optimization (ICLR 2024 · similarity 0.89)
- Posterior Inference with Diffusion Models for High-dimensional Black-box Optimization (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.