Random axis-aligned subspaces let a simple expected-improvement rule generate large parallel batches without extra hyperparameters, and the method wins on most CEC2017 benchmarks.
Mockus, Application of bayesian approach to numerical methods of global and stochastic optimization, Journal of Global Optimization 4 (1994) 347–365
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Scalable Batch Bayesian Optimization Via Subspace Acquisition Functions
Random axis-aligned subspaces let a simple expected-improvement rule generate large parallel batches without extra hyperparameters, and the method wins on most CEC2017 benchmarks.