An ML engineer needs to train a supervised deep learning model. The available dataset is a large number of unlabeled images that only employees should access. The ML engineer needs to implement a solu

Question

Options

• AUse Amazon Rekognition to automatically label the dataset.
• BTrain the deep learning model directly on the raw data. Let the model infer the labels by itself.
• CUse Amazon SageMaker Ground Truth to create an annotation job that specifies the labeling task
• DSet up workforce teams to access a private workforce to run and review the annotation job
• EUse Amazon Mechanical Turk to complete the annotation job created by Amazon SageMaker

• A
  13% (5)
• B
  8% (3)
• C
  74% (29)
• E
  5% (2)

Explanation

C and D work together to solve both requirements simultaneously: SageMaker Ground Truth (C) provides the managed annotation workflow for creating labeling jobs with built-in quality control mechanisms, while a private workforce (D) restricts access to employees only - satisfying the security constraint that only employees should access the images. Together, they maximize labeling accuracy through human review with workforce verification.

Why the distractors fail:

• A (Rekognition): Automated labeling trades accuracy for speed; the question explicitly requires the highest possible accuracy, which demands human review.
• B (raw data / self-inferred labels): Supervised learning by definition requires labeled training data - you cannot train a supervised model on unlabeled data and have it "infer" its own labels; that's closer to self-supervised or unsupervised learning.
• E (Mechanical Turk): MTurk uses the public crowd, which violates the requirement that only employees access the images. A private workforce is the correct choice for restricted data.

Memory tip: Think of it as a two-part lock - Ground Truth is the tool (C), and private workforce is the key (D). Whenever you see "only employees should access," that's your signal to eliminate any public crowd option (MTurk) and choose private workforce instead.

No community discussion yet for this question.