1.2Standard
Production receives an unseen input type.
Do not auto-confirm uncertain predictions.
Choose two controls.
× Incorrect
Monitor distribution, distance, confidence, and missingness and route low-confidence or out-of-distribution inputs to rejection, alternatives, or human review.
Detailed explanation
IncorrectAdd unknown inputs to training immediately without validation
Errors or attacks can be learned.
CorrectMonitor distribution, distance, confidence, and missingness to detect inputs outside training scope
Uncertain cases become visible.
IncorrectHigh confidence proves an unseen input is correct
Calibration may fail out of distribution.
CorrectReject, fall back, request more data, or review low-confidence inputs
Unsafe automation is reduced.
Check official documentation
Compare the following points with the official documentation.
AWS公式AIF-C01 Domain 1.2・3.3の不確実性、分布外検知、人的監督を確認する。
Official references
What to confirm
モデルが学習していない入力を検出し、自動確定から安全に切り替えられる。
Key points
- Out of distribution
- Uncertainty
- Reject
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
Out-of-distribution input
An input whose characteristics differ materially from the training or validation population.
Confidence score
A model or system estimate associated with a prediction or response.
Question IDAWS-AIF-492
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