Research map / Probabilistic and scientific modeling / Bayesian optimization
Trading Convergence Rate with Computational Budget in High Dimensional Bayesian Optimization
Hung Tran-The, Sunil Gupta, Santu Rana, Svetha Venkatesh
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 |
|---|---|---|---|
| Optimize the acquisition function over multiple sampled low-dimensional subspaces instead of the full space | MS-UCB · High-dimensional benchmark functions | log distance to true optimum | Improved |
| Use MS-UCB rather than the compared methods | Neural network parameter search on MNIST · MNIST validation loss, 100 and 500 dimensions | validation loss | Improved |
| Use MS-UCB rather than the compared methods | Ramp-loss classification on Gisette · Gisette, varying C | loss | Improved |
| Increase the number of sampled subspaces by increasing N0 or alpha (ablation) | MS-UCB · Ackley and Hyper-Ellipsoid, D=100 | log regret | Improved |
| Set alpha to 2 instead of 0 with N0=1 (ablation) | MS-UCB · Ackley and Hyper-Ellipsoid, D=100 | log regret | Improved |
| Increase subspace dimension d (ablation) | MS-UCB · Ackley and Hyper-Ellipsoid, D=100 | log regret | Improved |
| Increase subspace dimension d (ablation) | MS-UCB · Ackley and Hyper-Ellipsoid, D=100 | computational cost | Hurt |
| Use d=5-dimensional subspaces rather than full-space acquisition optimization | MS-UCB · Ackley, Levy, Hyper-Ellipsoid, and Camelback | log distance to true optimum | Improved |
| Use MS-UCB rather than SRE | High-dimensional benchmark optimization · Camelback, D=50 | accuracy | Improved |
| Use MS-UCB rather than SRE | High-dimensional benchmark optimization · Camelback, D=100 | accuracy | Improved |
| Increase the number of subspaces as alpha grows (ablation) | MS-UCB · Cumulative regret bound | cumulative regret: R_T ≤ O*(√(dTγ_T) + D) if α ≥ D-d-1 | Improved |
Closest papers by meaning
- Bayesian Optimisation with Unknown Hyperparameters: Regret Bounds Logarithmically Closer to Optimal (NeurIPS 2024 · similarity 0.91)
- Understanding High-Dimensional Bayesian Optimization (ICML 2025 · similarity 0.91)
- BOIDS: High-Dimensional Bayesian Optimization via Incumbent-Guided Direction Lines and Subspace Embeddings (AAAI 2025 · similarity 0.90)
- Vanilla Bayesian Optimization Performs Great in High Dimensions (ICML 2024 · similarity 0.90)
- Relaxing the Additivity Constraints in Decentralized No-Regret High-Dimensional Bayesian Optimization (ICLR 2024 · similarity 0.89)
- Standard Gaussian Process is All You Need for High-Dimensional Bayesian Optimization (ICLR 2025 · similarity 0.89)
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.
