Check your understanding of adversarial attacks before moving to the next section.
Section Quiz
Q1. An evasion attack targets which phase of the AI lifecycle?
Q2. Which of the following is the most direct defence against model extraction attacks?
Q3. A membership inference attack is particularly concerning when AI models are trained on:
Q4. Zou et al. (2023) demonstrated that gibberish-looking adversarial suffixes could:
Q5. How does a model inversion attack reconstruct information about training data?
Q6. Which defence is aimed specifically at model inversion and membership inference, by preventing a model from memorising individual records?
