Research map / Probabilistic and scientific modeling / Bayesian optimization
Bayesian Algorithm Execution: Estimating Computable Properties of Black-box Functions Using Mutual Information
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 |
|---|---|---|---|
| Choose queries to maximize mutual information with the algorithm output | InfoBAX on shortest-path estimation · Synthetic and California roads graphs | queries to recover the ground-truth shortest path: 5 to 547 times fewer queries | Helped |
| Choose queries to maximize mutual information with a local optimizer's output | InfoBAX on Bayesian local optimization · Two benchmark functions | function evaluations to match EvolutionStrategy performance: 8 to 20 times fewer function evaluations | Helped |
| Choose queries to maximize mutual information with the top-k set or its relevant subsequence | InfoBAX on top-k estimation · Top-k estimation | Jaccard distance: roughly 2 times fewer function evaluations | Helped |
| Use execution-path information rather than algorithm-output information (ablation) | InfoBAX · Top-k estimation | Jaccard distance | Hurt |
| Use execution-path information rather than algorithm-output or subsequence information (ablation) | InfoBAX · Shortest-path estimation, Bayesian local optimization, and top-k estimation | task-specific error | Hurt |
| Use algorithm-output information rather than execution-path information (ablation) | InfoBAX · Shortest-path estimation, Bayesian local optimization, and top-k estimation | task-specific error | Helped |
| Use a relevant execution-path subsequence's function values as the information target (ablation) | InfoBAX · Shortest-path estimation, Bayesian local optimization, and top-k estimation | task-specific error | Helped |
| Query for information about the function rather than the target property | UncertaintySampling and MI_f baselines · Shortest-path estimation and top-k estimation | task-specific error | Hurt |
Closest papers by meaning
- Practical Bayesian Algorithm Execution via Posterior Sampling (NeurIPS 2024 · similarity 0.91)
- Bayesian Optimization of Functions over Node Subsets in Graphs (NeurIPS 2024 · similarity 0.90)
- Cost-aware Bayesian Optimization via the Pandora's Box Gittins Index (NeurIPS 2024 · similarity 0.88)
- A General Framework for User-Guided Bayesian Optimization (ICLR 2024 · similarity 0.88)
- Minimizing UCB: a Better Local Search Strategy in Local Bayesian Optimization (NeurIPS 2024 · 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.