AIF-C01 Guidelines for Responsible AI Practice Question

Exhibit

Refer to the exhibit.
```
{
  "smclarify_bias_report": {
    "pre_training": {
      "ClassImbalance": {
        "value": 1.5,
        "description": "Ratio of majority to minority class counts"
      }
    },
    "post_training": {
      "DPPL": {
        "value": 0.15,
        "description": "Difference in Positive Proportions in Predicted Labels"
      }
    }
  }
}
```

Refer to the exhibit. A data scientist runs an Amazon SageMaker Clarify bias analysis on a binary classifier. The pre-training ClassImbalance is 1.5 and the post-training DPPL is 0.15. What should the data scientist conclude?

⚠ Common exam trap

In AWS AI Practitioner exams, a common misconception is that a low pre-training imbalance automatically means the model is fair, but the post-training DPPL metric directly measures prediction bias and can reveal unfairness even when the data appears balanced.

Answer choices

Why each option matters

Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.

Correct answer & explanation

✓

The data has a mild class imbalance, but the model shows a noticeable bias in predictions.

The pre-training ClassImbalance metric of 1.5 indicates a mild class imbalance (values close to 1.0 indicate balance, while values significantly above 1.0 indicate imbalance). The post-training DPPL (Difference in Positive Proportions in Labels) metric of 0.15 exceeds the commonly accepted fairness threshold of 0.10, indicating a noticeable bias in the model's predictions. Therefore, the data has a mild imbalance, but the model exhibits a bias that warrants further investigation.

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

✗
The data is highly imbalanced and the model is unbiased.
Why it's wrong here
DPPL of 0.15 exceeds the 0.1 threshold, so the model is biased, not unbiased. The ClassImbalance of 1.5 does indicate imbalance, making the first half tempting, but the conclusion about fairness inverts the post-training metric, which measures disparity in positive predictions between facets.
✓
The data has a mild class imbalance, but the model shows a noticeable bias in predictions.
Why this is correct
ClassImbalance of 1.5 sits just above the balanced threshold, indicating only mild skew in the training data. DPPL of 0.15 exceeds the typical 0.1 tolerance, meaning predicted positive rates differ noticeably between groups, so the model exhibits real predictive bias.
✗
The pre-training metric indicates a fairness issue, but the post-training metric is acceptable.
Why it's wrong here
ClassImbalance of 1.5 sits within the acceptable range, so the pre-training data shows no fairness issue; DPPL of 0.15 exceeds the 0.1 threshold, so the model is biased. The option is tempting because it correctly flags a post-training problem, but it swaps which metric fails.
✗
The data is perfectly balanced and the model is fair.
Why it's wrong here
A ClassImbalance of 1.5 indicates imbalance rather than perfect balance, and DPPL of 0.15 exceeds the 0.1 fairness threshold, so the model is not fair. The option is tempting because balanced data and fair outcomes are the desired end state, but neither metric here supports that conclusion.

Go deeper

AIF-C01 Guidelines for Responsible AI Practice Question

Exhibit

Why each option matters

Option-by-option breakdown

Related to this question