Research map / Probabilistic and scientific modeling / Bayesian optimization
Preferential Bayesian Optimization
Javier González, Zhenwen Dai, Andreas Damianou, Neil D. Lawrence
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 GP-based preferential Bayesian optimization with dueling-Thompson sampling | Preferential optimization on Forrester, Six-Hump Camel, Gold-Stein, and Levy · Four objective-function benchmarks | true objective value at the current Condorcet winner | Improved |
| Model correlations across arms with a GP rather than using Sparring | Six-Hump Camel · Six-Hump Camel | speed of finding the optimum | Improved |
| Use interactive Bayesian optimization with a greedy acquisition | Interactive Bayesian optimization (IBO) · Four objective-function benchmarks | true objective value at the current Condorcet winner | Hurt |
| Choose duels uniformly at random | Random policy · Four objective-function benchmarks | true objective value at the current Condorcet winner | Hurt |
| Use bandit methods that do not model correlations across arms | Sparring · Discrete benchmark grids | evaluations needed to converge | Hurt |
| Use pure exploration instead of dueling-Thompson sampling (ablation) | PBO · Four objective-function benchmarks | true objective value at the current Condorcet winner | Hurt |
| Use Copeland Expected Improvement instead of dueling-Thompson sampling (ablation) | PBO · Forrester | true objective value at the current Condorcet winner | Hurt |
| Use dueling-Thompson sampling rather than pure exploration or Copeland Expected Improvement (ablation) | PBO · Forrester, Six-Hump Camel, Gold-Stein, and Levy | true objective value at the current Condorcet winner | Improved |
| Use Bernoulli outcome variance as the exploration signal rather than latent-function uncertainty (ablation) | PBO pure-exploration acquisition · Duel space | ability to avoid revisiting observed duels | Hurt |
Closest papers by meaning
- Principled Preferential Bayesian Optimization (ICML 2024 · similarity 0.89)
- Bayesian Optimization from Human Feedback: Near-Optimal Regret Bounds (ICML 2025 · similarity 0.89)
- PABBO: Preferential Amortized Black-Box Optimization (ICLR 2025 · similarity 0.88)
- DT-PBO: an Interpretable Tree-based Surrogate Model for Preferential Bayesian Optimization (NeurIPS 2026 · similarity 0.88)
- User Preference Meets Pareto-Optimality in Multi-Objective Bayesian Optimization (AAAI 2025 · similarity 0.87)
- Direct Preference-Based Evolutionary Multi-Objective Optimization with Dueling Bandits (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.
