Explain Core AI Principles and Terminology CY0-001 Practice Questions (Page 8)

Questions 36–39

1. 36expert · hard

   A security team is building a model to detect advanced persistent threats (APTs) that evolve over time. They have a large dataset of network traffic. They are considering using a deep neural network. What is a key trade-off they must consider?

   • ADeep learning models can capture complex patterns but are often less interpretable, making it harder to explain why a specific alert was raised
   • BDeep learning models are easier to deploy in real-time because they are lightweight
   • CDeep learning models require less data than traditional machine learning, which is a benefit
   • DDeep learning models are always more accurate than traditional machine learning, so there is no trade-off
   Reveal Solution Discussion Select an answer first
2. 37application · medium

   A data scientist is explaining to a security analyst how a neural network detects malicious network traffic. She says the network has an input layer, several hidden layers, and an output layer. What is the primary function of the hidden layers?

   • ATo reduce the number of input features to exactly one
   • BTo provide the final classification decision
   • CTo store the training data so the network can recall it during classification
   • DTo transform the input data into increasingly abstract representations that help the network learn complex patterns
   Reveal Solution Discussion Select an answer first
4. 38foundation · easy

   Which cybersecurity application commonly uses machine learning?

   • AAnomaly detection in network traffic to identify unusual behavior.
   • BEncrypting data using the AES algorithm.
   • CManual review of security logs by analysts.
   • DConfiguring firewall rules based on vendor documentation.
   Reveal Solution Discussion Select an answer first
5. 39foundation · easy

   Why is deep learning particularly suited for complex cybersecurity tasks like malware classification?

   • AIt does not require any labeled data.
   • BIt requires less data than traditional machine learning.
   • CIt can automatically extract relevant features from raw data such as binary files or network packets.
   • DIt is faster to train than traditional machine learning models.
   Reveal Solution Discussion Select an answer first