MLA-C01 ML Model Development Practice Question
A data scientist is using SageMaker Automatic Model Tuning with Hyperband. They want to stop poorly performing trials early to save resources. Which strategy does Hyperband use?
⚠ Common exam trap
MLA-C01 often tests the association between Hyperband and Successive Halving; candidates may confuse it with Bayesian optimization because both are advanced tuning methods.
Answer choices
Why each option matters
Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.
Correct answer & explanation
✓
Successive Halving
Hyperband is a hyperparameter tuning strategy that uses Successive Halving to allocate resources efficiently. It starts many trials with small resource budgets, then iteratively promotes the best-performing trials to larger budgets while stopping poor performers early, saving compute.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Grid search
Why it's wrong here
Grid search exhaustively evaluates every hyperparameter combination, so no trial is stopped early; Hyperband instead allocates successive halving rounds, discarding the worst-performing trials. Grid search is tempting because it guarantees coverage of a discrete search space, and it would be the correct choice for a small, low-dimensional tuning job where exhaustive evaluation is affordable.
- ✗
Random search
Why it's wrong here
Random search samples hyperparameters independently and runs every trial to completion, providing no mechanism to terminate poor trials early; Hyperband's successive halving does. Random search is tempting because it explores the space cheaply, and it would be correct for high-dimensional spaces where only a few hyperparameters matter.
- ✓
Successive Halving
Why this is correct
Successive Halving runs many configurations with small resource budgets, then repeatedly discards the worst half and doubles resources for survivors. Hyperband wraps this in multiple brackets with varying budgets, so poor trials terminate early and compute is redirected to promising candidates.
- ✗
Bayesian optimization
Why it's wrong here
Bayesian optimisation builds a surrogate model to pick the next hyperparameters, but it does not itself terminate underperforming trials; Hyperband's successive halving does that. Bayesian optimisation is tempting because it is the default SageMaker tuner, and it would be correct when evaluations are expensive and you want sample-efficient search rather than early stopping.
Go deeper
Related to this question
About these practice questions
One of 665 original MLA-C01 practice questions on Courseiva, each with a full explanation and wrong-answer analysis — not exam dumps or protected exam content. Learn why practice questions differ from exam dumps →
JA
Written and reviewed by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
Last reviewed September 2026 · checked against the official Amazon Web Services exam blueprint
This MLA-C01 practice question is part of Courseiva's free Amazon Web Services certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the MLA-C01 exam.