Research map / Probabilistic and scientific modeling / Bayesian optimization
Minimum Regret Search for Single- and Multi-Task 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 |
|---|---|---|---|
| Use MRS rather than ES | Bayesian optimization on the synthetic single-task benchmark · Synthetic single-task benchmark | runs with simple regret > 10^-2: 10 for ES; 4 for MRS | Helped |
| Use MRS rather than ES | Bayesian optimization on the synthetic single-task benchmark · Synthetic single-task benchmark | median simple regret | No clear effect |
| Use MRS rather than ES | Bayesian optimization on the synthetic single-task benchmark · Synthetic single-task benchmark | mean simple regret | Helped |
| Use MRS rather than EI | Bayesian optimization on the synthetic single-task benchmark · Synthetic single-task benchmark | runs with simple regret > 10^-2: 4 for MRS; 4 for EI | No clear effect |
| Use MRS rather than ES | Bayesian optimization on the synthetic benchmark with model mismatch · Synthetic single-task benchmark with model mismatch | mean simple regret | Mixed |
| Use MRS rather than ES | Multi-task Bayesian optimization for robotic ball-throwing · Ball-throwing, one or two controllable joints | mean offline policy performance | No clear effect |
| Use MRS rather than ES | Multi-task Bayesian optimization for robotic ball-throwing · Ball-throwing, three or four controllable joints | mean offline policy performance: p < 0.01 after 200 episodes | Helped |
| Increase the number of controllable joints (ablation) | Multi-task Bayesian optimization for robotic ball-throwing · Ball-throwing | mean offline policy performance | Hurt |
| Replace expected-regret minimization with greedy GP-mean selection | Multi-task Bayesian optimization for robotic ball-throwing · Ball-throwing | mean offline policy performance | Hurt |
| Replace expected-regret minimization with random parameter selection | Multi-task Bayesian optimization for robotic ball-throwing · Ball-throwing | mean offline policy performance | Hurt |
| Use a single point estimate of the recommendation instead of marginalizing its uncertainty (ablation) | MRS · Synthetic single-task benchmark | mean simple regret | No clear effect |
| Use a single point estimate of the recommendation instead of marginalizing its uncertainty (ablation) | MRS · Synthetic single-task benchmark with model mismatch | mean simple regret | Hurt |
Closest papers by meaning
- Bayesian Optimization from Human Feedback: Near-Optimal Regret Bounds (ICML 2025 · similarity 0.89)
- Principled Preferential Bayesian Optimization (ICML 2024 · similarity 0.88)
- A Unified Framework for Entropy Search and Expected Improvement in Bayesian Optimization (ICML 2025 · similarity 0.87)
- Random Exploration in Bayesian Optimization: Order-Optimal Regret and Computational Efficiency (ICML 2024 · similarity 0.87)
- Posterior Sampling-Based Bayesian Optimization with Tighter Bayesian Regret Bounds (ICML 2024 · similarity 0.87)
- Efficient Exploration in Average-Reward Constrained Reinforcement Learning: Achieving Near-Optimal Regret With Posterior Sampling (ICML 2024 · 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.