3.3Standard
Data and performance change over time.
Avoid blind retraining.
Choose two policies.
× Incorrect
Combine performance, drift, label quality, volume, fairness, and business changes and require independent testing, safety, cost, rollback, and approval.
Detailed explanation
IncorrectAutomation removes approval and audit needs
High-impact change needs control.
CorrectCombine performance, drift, label quality, volume, fairness, and business conditions
Need is evidence-based.
CorrectAfter retraining, run independent quality, safety, cost, and rollback tests
The new model is verified.
IncorrectRetrain immediately after every drift signal
Cause may be an attack or label change.
Check official documentation
Compare the following points with the official documentation.
AWS公式AIF-C01 Domain 1.3・3.3・4.3の再学習、監視、変更管理を確認する。
Official references
What to confirm
ドリフトを検知しただけで盲目的に再学習せず、品質と統制を含めて判断できる。
Key points
- Retraining trigger
- Independent test
- Approval
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
Retraining trigger
A condition that starts a review or new training workflow.
Independent test
An evaluation set kept separate from training and tuning.
Question IDAWS-AIF-527
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