In the shared responsibility model for IaaS, which of the following is typically the customer's responsibility?

Physical security of the data center

Network infrastructure security

Patching the guest operating system

A security team is deploying an image classification model for a facial recognition system. They are concerned about adversarial examples that could cause misclassification. Which defensive technique is most effective at improving the model's robustness?

Adversarial training with perturbed samples

Removing all regularization techniques

Increasing the model's complexity without additional data

Using a larger batch size during training

A security administrator is concerned about a vulnerability in the hypervisor that could allow a VM to access the host's memory. Which control is most effective in mitigating this risk?

Use resource limits to cap VM memory usage

Disable unused virtual hardware devices

Enable VM snapshots for rollback

Apply hypervisor security patches promptly

A company trains a fraud detection model. To protect it from attackers who might try to manipulate the model's behavior, which practice is most effective?

Limit the number of API requests per user

Use a larger model architecture

Keep the model's training data secret

Regularly retrain the model with a mix of clean and adversarially perturbed samples

GIAC Security Essentials

Domain 6Objective 3

Virtualization, Cloud Security, and AI Essentials GSEC Practice Questions (Page 5)

Part of the Data Protection and Emerging Technologies domain, which makes up ~12% of our current practice bank. GIAC (SANS) does not publish an official question count, but from its 240-minute exam (~95–160 total, ~11–19 in this domain), expect 4–6 from this objective — we provide 53 practice questions to prepare you well beyond it. (estimate)

53questions here

11free pages

10concepts

Questions 21–25

21foundation · easy
In the shared responsibility model for IaaS, which of the following is typically the customer's responsibility?
Select an answer first
22application · medium
A security team is deploying an image classification model for a facial recognition system. They are concerned about adversarial examples that could cause misclassification. Which defensive technique is most effective at improving the model's robustness?
Select an answer first
23application · medium
A security administrator is concerned about a vulnerability in the hypervisor that could allow a VM to access the host's memory. Which control is most effective in mitigating this risk?
Select an answer first
24application · medium
A company trains a fraud detection model. To protect it from attackers who might try to manipulate the model's behavior, which practice is most effective?
Select an answer first
25foundation · easy
Which cloud service model provides a fully managed application that the customer accesses via a web browser or API, with no control over the underlying infrastructure?
Select an answer first

Finished these 5 questions?

Review the revealed explanations, or continue through the curriculum.

Next 5 questions →Back to GSEC

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 GIAC (SANS). “GSEC” is a trademark of its owner, used for identification only.