Databricks Machine Learning Associate Question 207
Select 2You are training a machine learning model on Databricks and want to tune its hyperparameters using Bayesian optimization. Which of the following statements about Bayesian methods for hyperparameter tuning are correct?
- A
Bayesian optimization builds a probabilistic model of the objective function to guide the search for optimal hyperparameters.
- B
Bayesian methods typically require more iterations than grid search to find optimal hyperparameters.
- C
Bayesian optimization uses prior knowledge from previous evaluations to decide the next set of hyperparameters to evaluate.
- D
Bayesian optimization is well-suited for high-dimensional hyperparameter spaces compared to random search.
- E
Bayesian optimization cannot handle non-continuous hyperparameter spaces.
Show answer and explanation
Correct answers: A, C
Explanation
Bayesian optimization is an efficient approach for tuning hyperparameters as it builds a probabilistic model of the objective function and uses prior evaluations to guide its search. This makes it faster and more efficient than exhaustive methods like grid search, but it is not ideal for very high-dimensional spaces. Additionally, it can handle both continuous and non-continuous hyperparameter spaces.
- A. Correct.
Correct: Bayesian optimization uses a probabilistic model, such as a Gaussian Process, to approximate the objective function and guide the search for optimal hyperparameters.
- B. Incorrect.
Incorrect: Bayesian optimization typically requires fewer iterations than methods like grid search, as it uses a more informed approach to selecting hyperparameter values.
- C. Correct.
Correct: Bayesian optimization leverages prior evaluations to update its probabilistic model, enabling it to make informed decisions about the next hyperparameters to test.
- D. Incorrect.
Incorrect: While Bayesian optimization can handle high-dimensional spaces, its performance degrades as the number of dimensions increases, making it less efficient in such cases compared to random search.
- E. Incorrect.
Incorrect: Bayesian optimization can handle non-continuous hyperparameter spaces, such as categorical variables, by using appropriate surrogate models or techniques.