Research map / Graph learning / Graph algorithms and clustering
Combinatorial Bayesian Optimization using the Graph Cartesian Product
Changyong Oh, Jakub Tomczak, Efstratios Gavves, Max Welling
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 |
|---|---|---|---|
| replace SA/SMAC/TPE/BOCS with COMBO | combinatorial BO · Contamination control lambda=0 | objective (lower better): 21.37±0.03 -> 21.28±0.03 | Improved |
| use COMBO instead of SA / BOCS-SDP | Ising sparsification · Ising sparsification lambda=1e-4 | KL objective (lower better): 0.059±0.013 -> 0.081±0.028 | Hurt |
| use COMBO instead of SA | Branin · Branin (ordinal) | objective (lower better): 0.4659±0.0211 -> 0.4113±0.0012 | Improved |
| use COMBO instead of SA | Pest Control · Pest Control | objective (lower better): 12.7154±0.0918 -> 12.0012±0.0033 | Improved |
| use COMBO instead of SA | wMaxSAT · wMaxSAT43 | negated minimum (lower better): -75.76±2.30 -> -85.02±2.14 | Improved |
| model arbitrarily high-order interactions with GP (COMBO) vs low-order BOCS-SDP | wMaxSAT · wMaxSAT43 | negated minimum (lower better): -51.13±1.69 -> -85.02±2.14 | Improved |
| use Horseshoe sparsity prior instead of non-sparsity priors (ablation) | COMBO · wMaxSAT60 | negated minimum | Improved |
| COMBO vs Regularized Evolution at 260 vs 500 evaluations | NAS · NAS CIFAR10 binary space | objective | Improved |
| BOCS-SDP (second-order) vs random search | NAS · NAS CIFAR10 binary space | objective | No clear effect |
Closest papers by meaning
- Omnipresent Yet Overlooked: Heat Kernels in Combinatorial Bayesian Optimization (NeurIPS 2025 · similarity 0.91)
- Mercer Features for Efficient Combinatorial Bayesian Optimization (AAAI 2021 · similarity 0.89)
- Bayesian Optimization of Functions over Node Subsets in Graphs (NeurIPS 2024 · similarity 0.88)
- Framework and Benchmarks for Combinatorial and Mixed-variable Bayesian Optimization (NeurIPS 2023 · similarity 0.88)
- Understanding High-Dimensional Bayesian Optimization (ICML 2025 · similarity 0.86)
- BayeSQP: Bayesian Optimization through Sequential Quadratic Programming (NeurIPS 2025 · 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.
