Research map / Probabilistic and scientific modeling / Bayesian optimization
Unexpected Improvements to Expected Improvement for Bayesian Optimization
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 |
|---|---|---|---|
| compute log of EI stably (LogEI) instead of naive EI | analytic EI · 10D Sum-of-Squares | regret | Helped |
| use LogEI instead of EI | analytic EI · Ackley (2D, 8D, 16D) | best observed value | Helped |
| use LogEI instead of EI | analytic EI · Michalewicz | best observed value | Helped |
| use LogCEI instead of CEI | constrained EI · four engineering design problems with black-box constraints | best feasible value | Helped |
| use qLogEI instead of qEI | parallel EI · Ackley 16D, q=4,16,32 | best observed value | Helped |
| joint batch optimization instead of sequential greedy with qLogEI (ablation) | qLogEI · Ackley 16D | best observed value | Helped |
| use qLogEI with SAAS prior or standard GP instead of qEI | qEI · 100D embedded Hartmann6 | best observed value | Helped |
| use qLogEI instead of qEI | qEI · 103D SVM and 100D Rover | test RMSE / reward | No clear effect |
| use qLogEHVI instead of qEHVI | qEHVI · Cell Network 30D and DTLZ2 16D | hypervolume | Helped |
| LogEI vs JES on Ackley and Michalewicz | JES · Ackley/Michalewicz | best observed value | Mixed |
Closest papers by meaning
- Expected Hypervolume Improvement Is a Particular Hypervolume Improvement (AAAI 2025 · similarity 0.91)
- A Unified Framework for Entropy Search and Expected Improvement in Bayesian Optimization (ICML 2025 · similarity 0.89)
- Probability Distribution of Hypervolume Improvement in Bi-objective Bayesian Optimization (ICML 2024 · similarity 0.88)
- Simple Extensions of Single-Objective Acquisition Functions and Hedge Strategies for Multi-Objective Bayesian Optimization (NeurIPS 2026 · similarity 0.88)
- Convergence Rates of Constrained Expected Improvement (NeurIPS 2025 · similarity 0.88)
- Bayesian Optimisation with Unknown Hyperparameters: Regret Bounds Logarithmically Closer to Optimal (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.