Bayesian Optimization uses a probabilistic model to find out the relationship between hyperparammeters and the model performance, that guides the model to balance exploration (using new values) and exploitation (evaluating old values)