AWS Certified Machine Learning Engineer Associate · Free Practice Question Medium
Question 20
A data scientist needs to set up monitoring for a deployed machine learning model using Amazon SageMaker. They are primarily concerned with detecting when the incoming data used for predictions becomes significantly different from the training data. What type of drift should they monitor for in this case?
-
A
Bias Drift
-
B
Data Drift
-
C
Feature Attribution Drift
-
D
Model Quality Drift
Reveal correct answer
Correct answer: B
A.
Bias drift monitors if the model’s predictions are unfairly favoring certain groups or categories.
B.
Data drift occurs when the incoming data that the model sees during inference starts to differ significantly from the data it was trained on. This is a crucial metric to monitor as changes in data distribution can degrade model performance over time.
C.
Feature attribution drift tracks the importance of specific features in model predictions over time.
D.
Model quality drift refers to the degradation of model performance (e.g., accuracy) in production.
Discussion
Think the marked answer is wrong, or have a better explanation? Share it below — comments appear after review.
