MLA-C01 Question #236: Real Exam Question with Answer & Explanation

Question

A company is creating an ML model to identify defects in a product. The company has gathered a dataset and has stored the dataset in TIFF format in Amazon S3. The dataset contains 200 images in which the most common defects are visible. The dataset also contains 1,800 images in which there is no defect visible. An ML engineer trains the model and notices poor performance in some classes. The ML engineer identifies a class imbalance problem in the dataset. What should the ML engineer do to solves this problem?

Options

• AUse a few hundred images and Amazon Rekognition Custom Labels to train a new model.
• BUndersample the 200 images in which the most common defects are visible.
• COversample the 200 images in which the most common defects are visible.
• DUse all 2,000 images and Amazon Rekognition Custom Labels to train a new model.

Explanation

{"question_number": 9, "correct_answer": "C", "explanation": "The dataset has a severe class imbalance: 200 defect images (minority class) vs. 1,800 non-defect images (majority class), a 9:1 ratio. The model performs poorly on the defect class because it has far fewer examples to learn from. The standard remedy for this scenario is oversampling the minority class - generating additional training samples (via duplication or augmentation techniques like rotation, flipping, brightness changes) from the 200 defect images until the classes are more balanced. Option B (undersampling the 200 defect images) would further reduce the minority class and worsen the problem. Options A and D (Amazon Rekognition Custom Labels) are valid managed training services, but they do not inherently fix the underlying class imbalance - the same skewed ratio would persist, and the model would still underperform on the defect class.", "generated_by": "claude-sonnet", "llm_judge_score": 4}

No community discussion yet for this question.