Research map / Probabilistic and scientific modeling / Bayesian optimization
A Quantile-based Approach for Hyperparameter Transfer Learning
David Salinas, Huibin Shen, Valerio Perrone
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 Gaussian Copula normalization and a parametric prior in a Gaussian Copula Process | Hyperparameter optimization · DeepAR | DTM normalized over random search: 0.00 -> 0.73 | Improved |
| Use Gaussian Copula normalization and a parametric prior in a Gaussian Copula Process | Hyperparameter optimization · FCNET | DTM normalized over random search: 0.00 -> 0.94 | Improved |
| Use Gaussian Copula normalization and a parametric prior in a Gaussian Copula Process | Hyperparameter optimization · XGBoost | DTM normalized over random search: 0.00 -> 0.37 | Improved |
| Use Gaussian Copula normalization and a parametric prior in a Gaussian Copula Process | Hyperparameter optimization · NAS-Bench-201 | DTM normalized over random search: 0.00 -> 0.94 | Improved |
| Add a parametric prior to Gaussian Copula Process (ablation) | GCP · DeepAR | DTM normalized over random search: 0.42 -> 0.73 | Improved |
| Add a parametric prior to Gaussian Copula Process (ablation) | GCP · FCNET | DTM normalized over random search: 0.79 -> 0.94 | Improved |
| Add a parametric prior to Gaussian Copula Process (ablation) | GCP · XGBoost | DTM normalized over random search: 0.31 -> 0.37 | Improved |
| Add a parametric prior to Gaussian Copula Process (ablation) | GCP · NAS-Bench-201 | DTM normalized over random search: 0.45 -> 0.94 | Improved |
| Use Gaussian Copula normalization instead of standardization with a parametric prior (ablation) | Thompson sampling · FCNET | DTM normalized over random search: -563.27 -> 0.83 | Improved |
| Use Gaussian Copula normalization instead of standardization with a parametric prior (ablation) | Gaussian process with parametric prior · FCNET | DTM normalized over random search: -166.64 -> 0.94 | Improved |
| Add a Gaussian Copula parametric prior to use transfer information | Copula-based hyperparameter optimization · DeepAR with a dissimilar target task | DTM normalized over random search | Mixed |
| Adapt the transfer prior with observations from the current task using GCP (ablation) | CTS · DeepAR and XGBoost | DTM | Improved |
Closest papers by meaning
- Bayesian Optimisation with Unknown Hyperparameters: Regret Bounds Logarithmically Closer to Optimal (NeurIPS 2024 · similarity 0.88)
- Cost-Sensitive Freeze-thaw Bayesian Optimization for Efficient Hyperparameter Tuning (NeurIPS 2025 · similarity 0.87)
- Transfer Learning Through Conditional Quantile Matching (NeurIPS 2026 · similarity 0.87)
- Efficient Hyperparameter Optimization with Adaptive Fidelity Identification (CVPR 2024 · similarity 0.86)
- Optimizing Posterior Samples for Bayesian Optimization via Rootfinding (ICLR 2025 · similarity 0.86)
- Modeling All Response Surfaces in One for Conditional Search Spaces (AAAI 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.
