ISTQB Certified Tester: AI Testing Practice Mode

Community Discussion

Written by other learners. Comments can disagree with the marked answer — check the official documentation when in doubt.

• Chose B0

  - This is an adversarial example — an input crafted with a small, often imperceptible perturbation that causes a model to misclassify. FGSM (Fast Gradient Sign Method) and PGD (Projected Gradient Descent) are common adversarial attack algorithms used in robustness testing. A metamorphic relation here would have asserted the prediction stays 'cat' under such perturbation; this test failure demonstrates lack of robustness.