Explain the basic principles of Bayesian methods for hyperparameter tuning. How do these methods differ from random search and grid search? Provide a detailed explanation of how Bayesian optimization works, including the role of the surrogate model and the acquisition function in guiding the search process.