Describe the Impact of Predictive AI CERTIFIED-AI-EXPERT Practice Questions (Page 4)

Questions 16–20

1. 16foundation · easy

   Which limitation is associated with predictive AI models?

   • AThey can perpetuate biases present in the training data.
   • BThey are immune to changes in the underlying data distribution.
   • CThey are always fully interpretable by humans.
   • DThey require no data to make predictions.
   Reveal Solution Discussion Select an answer first
2. 17application · medium

   A data science team has trained a predictive model to forecast equipment failures in a manufacturing plant. They are about to deploy it to production. Which step is essential to include in the model lifecycle after deployment?

   • AContinuously monitor model performance and retrain when data drift is detected
   • BReplace the model with a generative AI model to create maintenance reports
   • CDelete the training data to save storage costs
   • DFreeze the model to ensure consistent predictions forever
   Reveal Solution Discussion Select an answer first
4. 18application · medium

   A data science team has developed a predictive model to forecast website traffic. They have deployed it to production. Which activity is part of the monitoring stage of the model lifecycle?

   • ASelecting the best algorithm for the model
   • BDocumenting the model's features for the data dictionary
   • CCollecting more training data to improve the model before deployment
   • DTracking model accuracy against actual traffic data and alerting on significant deviations
   Reveal Solution Discussion Select an answer first
5. 19expert · hard

   A manufacturing company wants to use predictive AI to reduce equipment downtime. They have sensor data from machines, but the data is noisy and has missing values. They also need to justify the investment to management by showing a clear return on investment (ROI). What is the most effective strategy?

   • AClean and impute the sensor data, train a predictive model for failure, and quantify cost savings from reduced downtime
   • BDeploy a model immediately with the noisy data to get quick results
   • CFocus on model interpretability rather than accuracy to gain management approval
   • DUse a generative model to create synthetic sensor data to replace missing values
   Reveal Solution Discussion Select an answer first
6. 20foundation · easy

   Which stage in the predictive AI model lifecycle involves tracking model performance over time?

   • AModel monitoring
   • BData preprocessing
   • CModel training
   • DModel deployment
   Reveal Solution Discussion Select an answer first