adversarial example

Stub entry: single definition derived from the NIST trustworthy-AI glossary; sector-specific readings not yet documented — contributions welcome.

Machine learning input sample formed by applying a small but intentionally worst-case perturbation ... to a clean example, such that the perturbed input causes a learned model to output an incorrect answer.

NIST AIRC, The Language of Trustworthy AI glossary (public-domain compilation), entry 'adversarial example', definition attributed to NISTIR_8269_Draft