Give an image model inputs with small brightness and noise changes.
Check whether a dangerous misclassification occurs.
What is the purpose of this test?
The test measures robustness to input variation and whether the system fails safely under abnormal or adversarial conditions.
Detailed explanation
A robust system maintains acceptable performance or fails safely.
Robustness and fairness are different properties.
Input perturbation tests are not a storage operation.
That leaves robustness risks unknown.
Check official documentation
Compare the following points with the official documentation.
AWS公式AIF-C01ガイドのDomain 4.2で堅牢性・安全な失敗・評価を確認する。
Official references
What to confirm
入力変化に対する性能と、安全に失敗できるかを堅牢性の評価として説明できる。
Key points
- Input variation
- Robustness
- Safe failure
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
Failure boundary
A system should define which degraded outcomes are acceptable and which require stopping.
Robustness evidence
Normal-case accuracy alone cannot demonstrate behavior under abnormal or adversarial inputs.