BO with gpCAM

Bayesian Optimization and active learning have 3 core ingredients: 

(1) A surrogate modeler (predictions + UQ)

(2) A decision policy (the acquisition function)

(3) The deterministic optimizer operating on the acquisition function