Research map / Probabilistic and scientific modeling / Bayesian optimization
Triangulation candidates for Bayesian optimization
Robert B. Gramacy, Annie Sauer, Nathan Wycoff
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 |
|---|---|---|---|
| replace multi-start L-BFGS-B search of EI with tricands candidates | EI · Goldstein-Price | best observed value / EI evaluations: about 1700 -> about 9000 evaluations | Improved |
| replace LHS candidates with tricands for TS | Thompson sampling · Goldstein-Price | criterion evaluations: 5\% | Mixed |
| replace multi-start EI with tricands at n_end | EI · ATO | BOV: p-value 0.0037 | Improved |
| hybrid: tricands best candidate initializes local numerical EI | EI-hyb · ATO | BOV: 85 and 88 out of 100 | Improved |
| replace LHS candidates with tricands | TGP/DGP hybrid EI · Gramacy & Lee | BOV: 97/100 | Improved |
| replace LHS candidates with tricands | DGP EI · Levy 5d | BOV: 81/100 | Improved |
| add numerical local search hybrid on top of tricands (ablation) | TGP/DGP EI · Levy 5d | BOV | No clear effect |
| replace LHS candidates with tricands | TGP/DGP EI · Michaelwicz 4d, n=75 | BOV: 75/100 | Improved |
| replace multi-start EI with tricands in early acquisitions | EI · Hartmann 6 | BOV | Hurt |
| replace multi-start EI with tricands at end of search | EI · Hartmann 6 | BOV | Improved |
| use DGP surrogate instead of TGP | tricands EI · Levy 5d (smooth changes) | BOV | Improved |
| use TGP surrogate instead of DGP | tricands EI · Gramacy & Lee and Michaelwicz (abrupt changes) | BOV | Improved |
Closest papers by meaning
- Local Bayesian optimization via maximizing probability of descent (NeurIPS 2022 · similarity 0.87)
- An Empirical Study of Bayesian Optimization: Acquisition Versus Partition (JMLR 2021 · similarity 0.87)
- Understanding High-Dimensional Bayesian Optimization (ICML 2025 · similarity 0.86)
- Boundary Exploration for Bayesian Optimization With Unknown Physical Constraints (ICML 2024 · similarity 0.86)
- Relaxing the Additivity Constraints in Decentralized No-Regret High-Dimensional Bayesian Optimization (ICLR 2024 · similarity 0.85)
- Vanilla Bayesian Optimization Performs Great in High Dimensions (ICML 2024 · similarity 0.85)
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.
