AWS Certified Machine Learning – Specialty — Question 309

A data scientist is trying to improve the accuracy of a neural network classification model. The data scientist wants to run a large hyperparameter tuning job in Amazon SageMaker. However, previous smaller tuning jobs on the same model often ran for several weeks. The ML specialist wants to reduce the computation time required to run the tuning job.

Which actions will MOST reduce the computation time for the hyperparameter tuning job? (Choose two.)

Answer options

• A. Use the Hyperband tuning strategy.
• B. Increase the number of hyperparameters.
• C. Set a lower value for the MaxNumberOfTrainingJobs parameter.
• D. Use the grid search tuning strategy.
• E. Set a lower value for the MaxParallelTrainingJobs parameter.

Correct answer: A, C

Explanation

Utilizing the Hyperband tuning strategy reduces computation time by stopping poorly performing training jobs early using a multi-fidelity approach. Additionally, reducing the MaxNumberOfTrainingJobs parameter directly limits the total number of training runs executed, thereby shortening the overall duration. Other options, such as increasing hyperparameters or using grid search, would increase computational overhead, while lowering MaxParallelTrainingJobs would increase the overall wall-clock time by reducing concurrency.