4.2Standard
A model's group gap grows after launch.
Find the cause.
Choose two practices.
× Incorrect
Compare time, group, distribution, threshold, and human-decision metrics regularly and define investigation, notice, stop, and reassessment.
Detailed explanation
CorrectCompare group metrics, distribution, thresholds, and human decisions by period
Operational change is visible.
IncorrectHide small-group metrics in aggregates
Gaps disappear.
IncorrectFairness at launch needs no later monitoring
Use and distributions change.
CorrectDefine investigation, notice, stop, reassessment, and stakeholder explanation when gaps grow
Remedy is continuous.
Check official documentation
Compare the following points with the official documentation.
AWS公式AIF-C01 Domain 4.2・3.3の公平性監視、ドリフト、緩和を確認する。
Official references
What to confirm
モデルと利用環境の変化で公平性が悪化する可能性を運用へ組み込める。
Key points
- Fairness monitoring
- Distribution change
- Alert
Notes
- Environment: AWS公式AIF-C01試験ガイドとAWS公式ドキュメントの確認
- Compare the definition with the AIF-C01 objective and the service documentation before applying it.
- Keep the concept separate from a particular answer choice or implementation step.
Foundation review
Fairness monitoring
Ongoing measurement of group outcomes and error differences after deployment.
Fairness drift
A change in group disparities caused by data, behavior, policy, or model changes.
Question IDAWS-AIF-515
EnvironmentAWS公式AIF-C01試験ガイドとAWS公式ドキュメントの確認
Last technical review2026-08-20
Report an error or rights issue