6.4 Explain AI (generative and Predictive) and Machine Learning in Network Operations 200-301 Practice Questions (Page 4)

Questions 16–20

1. 16application · medium

   A network automation team is developing a system that automatically adjusts routing metrics based on network congestion. The system receives feedback on whether its adjustments improved or worsened performance and uses that feedback to improve future decisions. Which type of machine learning is this?

   • ASupervised learning
   • BUnsupervised learning
   • CGenerative AI
   • DReinforcement learning
   Reveal Solution Discussion Select an answer first
2. 17application · medium

   A network team is considering using a generative AI tool to automatically generate configuration changes for all routers. What is a key limitation they should consider before implementation?

   • AThe AI may produce configurations that are syntactically correct but semantically incorrect for the network context
   • BThe AI will always require more time than manual configuration
   • CThe AI cannot handle standard configuration templates
   • DThe AI will replace the need for network engineers
   Reveal Solution Discussion Select an answer first
4. 18expert · hard

   A network team wants to predict when a router's CPU utilization will exceed 90% based on historical data. They have collected data for six months, but the data includes maintenance windows that caused spikes. What is the best way to handle this before training a predictive model?

   • AUse a generative AI model to create a report of the spikes
   • BRemove or label maintenance window data so the model learns normal patterns
   • CInclude all data as-is to capture all patterns
   • DUse unsupervised learning to cluster the spikes
   Reveal Solution Discussion Select an answer first
5. 19expert · hard

   A network engineer wants to use generative AI to create a configuration for a new site. The AI has access to the company's configuration templates and change history. What is the most reliable way to ensure the generated configuration is correct?

   • AAsk the AI to generate multiple versions and choose the one with the most commands
   • BUse the AI to generate a draft, then validate it against the templates and test in a lab
   • CUse the AI only for documentation, not for configuration
   • DTrust the AI output because it has access to the templates
   Reveal Solution Discussion Select an answer first
6. 20application · medium

   A network engineer is troubleshooting a complex routing issue and asks a generative AI assistant for suggestions. The AI provides a list of possible causes and recommended commands to run. What is the most appropriate way for the engineer to use this information?

   • AExecute all recommended commands immediately and apply any suggested configuration changes.
   • BUse the suggestions as a starting point and verify each command's output before making changes.
   • CIgnore the AI suggestions because they are generated, not based on real data.
   • DAsk the AI to automatically apply the changes to the router.
   Reveal Solution Discussion Select an answer first