ConceptGaussian Process Lower Confidence Bound
GP-LCB
An acquisition function for Bayesian optimization based on the lower confidence bound of a Gaussian process, balancing exploration and exploitation. It constructs a confidence lower bound from the predictive mean and variance, and selects the point with the smallest lower confidence bound as the next sampling point. The formulation is suited to minimization problems.
Timeline (last 90 days)
Nov 3
When combined with the GP-LCB acquisition function, BO-pro-c-tr reduces average simple regret by 2.6% and computational overhead by 34.9% relative to methods based on independent local GPs and trust regions.
Unverified50%