Hyperparameter Tuning MACHINE-LEARNING-ASSOCIATE Practice Questions (Page 2)

Questions 6–10

1. 6foundation · easy

   Which of the following is a key advantage of random search over grid search?

   • AIt is more efficient when only a few hyperparameters are important
   • BIt guarantees finding the global optimum
   • CIt always finds a better solution than grid search with the same number of evaluations
   • DIt requires fewer evaluations to cover the entire search space
   Reveal Solution Discussion Select an answer first
2. 7foundation · easy

   When using `fmin` with `algo=tpe.suggest`, what is the purpose of the `tpe.suggest` algorithm?

   • ATo perform a local search around the current best hyperparameters
   • BTo exhaustively evaluate all combinations of hyperparameters
   • CTo randomly sample hyperparameters from the search space
   • DTo use a tree-structured Parzen estimator to model the objective and propose promising hyperparameters
   Reveal Solution Discussion Select an answer first
4. 8foundation · easy

   Which of the following is a key benefit of using `SparkTrials` for hyperparameter tuning of single-node models?

   • AIt allows multiple trials to run in parallel, reducing wall-clock time
   • BIt automatically tunes the model's hyperparameters without user input
   • CIt reduces the total computational work of tuning
   • DIt enables training models on datasets that do not fit in memory
   Reveal Solution Discussion Select an answer first
5. 9foundation · easy

   In Bayesian optimization, what is the role of the surrogate model?

   • ATo approximate the objective function and guide the selection of the next hyperparameters
   • BTo store all evaluated hyperparameters and their losses
   • CTo replace the actual objective function for final model training
   • DTo calculate the gradient of the objective function
   Reveal Solution Discussion Select an answer first
6. 10foundation · easy

   After running `fmin`, how can you retrieve the loss value of the best trial?

   • ABy accessing the `best_loss` attribute of the returned dictionary
   • BBy re-running the objective function with the returned hyperparameters
   • CBy calling `fmin` with `return_loss=True`
   • DBy using the `trials` object to find the trial with the minimum loss
   Reveal Solution Discussion Select an answer first