AWS Certified Machine Learning Engineer Associate · Free Practice Question Medium
Question 40
A team is training a neural network model in Amazon SageMaker. During training, they notice that the model's performance has plateaued, and suspect a vanishing gradient issue. How can they leverage SageMaker Debugger to diagnose and address this problem?
-
A
Set up custom debug rules to capture disappearing gradients and adjust the learning rate as needed.
-
B
Enable SageMaker Clarify to detect vanishing gradients during training.
-
C
Use cross-validation to rotate training data, preventing gradient loss.
-
D
Increase the training epochs until SageMaker Debugger resolves the issue.
Reveal correct answer
Correct answer: A
A.
SageMaker Debugger can capture disappearing gradients through debug rules, allowing users to diagnose gradient issues and adjust parameters like the learning rate.
B.
SageMaker Clarify is focused on bias detection and explainability, not gradient issues.
C.
Cross-validation is used for generalization, not to address vanishing gradients.
D.
Increasing epochs alone won’t resolve vanishing gradient problems; adjustments in model configuration or learning rate are typically needed.
Discussion
Think the marked answer is wrong, or have a better explanation? Share it below — comments appear after review.
