A team is building a voice assistant and wants to test if it can be tricked by inaudible commands. Which type of adversarial example generation is most relevant?

Generate perturbations in the audio waveform that are imperceptible to humans but recognized as commands by the model.

Use a different microphone to record the command.

Add background music to the audio.

Change the speaker's voice pitch.

A security team is testing a facial recognition system. They have white-box access to the model (weights and gradients). They want to create a physical adversarial example (e.g., eyeglass frames) that fools the system in the real world. Which approach is most effective?

Print a high-resolution photo of the target person and hold it up to the camera.

Use a black-box attack that queries the system many times.

Apply a random noise pattern to the eyeglass frames.

Use the gradient information to optimize a perturbation that is robust to different angles and lighting conditions.

A financial institution uses an AI model to detect fraudulent transactions. They want to defend against an attacker who has access to the model's training data and can poison it. Which defensive strategy is most effective?

Robust model design, such as using anomaly detection and outlier removal during training.

Using a black-box model to hide the architecture.

Regularly retraining the model with the same data.

Limiting the number of features in the model.

EC-CouncilCertified Offensive AI Security Professional

Domain 4Objective 1

Adversarial Machine Learning Techniques COASP Practice Questions (Page 8)

Part of the Adversarial Machine Learning and Model Privacy Attacks domain, which makes up ~13% of our current practice bank.

38questions here

8free pages

3concepts

Questions 36–38

36application · medium
A team is building a voice assistant and wants to test if it can be tricked by inaudible commands. Which type of adversarial example generation is most relevant?
Select an answer first
37expert · hard
A security team is testing a facial recognition system. They have white-box access to the model (weights and gradients). They want to create a physical adversarial example (e.g., eyeglass frames) that fools the system in the real world. Which approach is most effective?
Select an answer first
38application · medium
A financial institution uses an AI model to detect fraudulent transactions. They want to defend against an attacker who has access to the model's training data and can poison it. Which defensive strategy is most effective?
Select an answer first

Finished these 3 questions?

Review the revealed explanations, or continue through the curriculum.

No more pagesBack to COASP

Free Basic Practice is a study aid with revealable answers — not a scored exam. Examers.io is independent and not affiliated with or endorsed by EC-Council. “COASP” is a trademark of its owner, used for identification only.

Adversarial Machine Learning Techniques COASP Practice Questions (Page 8)

Questions 36–38

A team is building a voice assistant and wants to test if it can be tricked by inaudible commands. Which type of adversarial example generation is most relevant?

A security team is testing a facial recognition system. They have white-box access to the model (weights and gradients). They want to create a physical adversarial example (e.g., eyeglass frames) that fools the system in the real world. Which approach is most effective?

A financial institution uses an AI model to detect fraudulent transactions. They want to defend against an attacker who has access to the model's training data and can poison it. Which defensive strategy is most effective?

Keep studying — Adversarial Machine Learning and Model Privacy Attacks