An ML engineer trained an ML model on Amazon SageMaker to detect automobile accidents from dosed-circuit TV footage. The ML engineer used SageMaker Data Wrangler to create a training dataset of images

Question

Options

• ACollect more images from all the cameras. Use Data Wrangler to prepare a new training dataset.
• BRecreate the training dataset by using the Data Wrangler corrupt image transform. Specify the
• CRecreate the training dataset by using the Data Wrangler enhance image contrast transform.
• DRecreate the training dataset by using the Data Wrangler resize image transform. Crop all images

• A
  7% (3)
• B
  2% (1)
• C
  80% (37)
• D
  11% (5)

Explanation

Option C is correct because the root cause of the production failures is image quality variation across different cameras - specifically differences in brightness and contrast. Applying SageMaker Data Wrangler's enhance image contrast transform normalizes these inconsistencies across the training dataset, closing the gap between training distribution and production data, which directly targets the identified problem in the least amount of time.

Why the distractors fail:

• A requires collecting and labeling new images from all cameras, which is time-consuming and doesn't guarantee quality normalization - making it the slowest option and not the root fix.
• B (corrupt image transform) intentionally degrades images to simulate noise, which is the opposite of what's needed; you want to improve consistency, not introduce more degradation.
• D (resize/crop) addresses dimensional inconsistency, not quality/contrast variation - resizing won't fix exposure or brightness differences between cameras.

Memory tip: Match the fix to the symptom. "Quality variation" = contrast/brightness problem → contrast enhancement. If the question said "different resolutions," D would be the candidate. If it said "need more data," A would apply. The keyword quality points directly to contrast transforms.

No community discussion yet for this question.