Research map / Probabilistic and scientific modeling / Bayesian optimization
Batch Bayesian Optimization on Permutations using the Acquisition Weighted Kernel
Changyong Oh, Roberto Bondesan, Efstratios Gavves, Max Welling
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 |
|---|---|---|---|
| add acquisition weights to the DPP kernel (LAW-EST vs DPP-MAX-EST) (ablation) | LAW2ORDER · TSP-att48 | tour cost: 39539 -> 19846 | Improved |
| add acquisition weights to the DPP kernel (LAW-EST vs DPP-MAX-EST) (ablation) | LAW2ORDER · QAP-chr12a | cost: 14732 -> 12067 | Improved |
| add acquisition weights to the DPP kernel (LAW-EST vs DPP-MAX-EST) (ablation) | LAW2ORDER · QAP-esc32a | cost: 276.5 -> 191.7 | Improved |
| add acquisition weights to DPP with prior-covariance diversity gauge (ablation) | LAW-PRIOR variants · permutation benchmarks | cost | Improved |
| use posterior covariance instead of prior covariance as diversity gauge (ablation) | LAW · permutation benchmarks | cost | Improved |
| LAW-EST vs q-EST/q-EI on QAP | LAW-EST · QAP benchmarks | cost: 11790 -> 12067 (chr12a) | Hurt |
| LAW-EST vs best baselines on FSP and TSP | LAW-EST · TSP-bayg29 | tour cost: 2060 -> 2038 | Improved |
| LAW-EST vs q-EI/q-EST on structure learning | LAW-EST · Alarm | negative log NML excess: 2739.77 -> 1409.27 | Improved |
| LAW-EST vs q-EI on Child structure learning | LAW-EST · Child | negative log NML excess: 864.85 -> 866.64 | No clear effect |
| LAW-EST (620 evals) vs GA with twice the evaluations | LAW-EST · Alarm | negative log NML excess: 4114.97 -> 1409.27 | Improved |
| LAW-EST vs GA(620) and GA(1240) on Sachs | LAW-EST · Sachs | negative log NML excess: 31.90 -> 29.58 | Mixed |
Closest papers by meaning
- Random Exploration in Bayesian Optimization: Order-Optimal Regret and Computational Efficiency (ICML 2024 · similarity 0.88)
- Omnipresent Yet Overlooked: Heat Kernels in Combinatorial Bayesian Optimization (NeurIPS 2025 · similarity 0.86)
- Training Greedy Policy for Proposal Batch Selection in Expensive Multi-Objective Combinatorial Optimization (ICML 2024 · similarity 0.86)
- Bayesian Optimization of Functions over Node Subsets in Graphs (NeurIPS 2024 · similarity 0.86)
- Batched Energy-Entropy acquisition for Bayesian Optimization (NeurIPS 2024 · similarity 0.85)
- Stopping Bayesian Optimization with Probabilistic Regret Bounds (NeurIPS 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.
