Oversee AI/ML Model QA/QC (e.g., Configuration Management, Model Performance) PMI-CPMAI Practice Questions (Page 1)

Questions 1–5

1. 1foundation · easy

   Which activity is part of a peer review process for an AI/ML model?

   • AWriting the final project report for management
   • BReviewing the model's code for logic errors and adherence to best practices
   • CSelecting the model's hyperparameters
   • DSetting up the cloud infrastructure for model training
   Reveal Solution Discussion Select an answer first
2. 2application · medium

   A team is using a configuration file to manage hyperparameters for their model. They need to track which configuration was used for each experiment. What is the best approach?

   • ASave a copy of the configuration file in the experiment folder with a timestamp
   • BUse a version control system to track changes to the configuration file and tag each experiment
   • CRecord the configuration in a comment in the training script
   • DKeep the configuration file in a shared drive and overwrite it for each experiment
   Reveal Solution Discussion Select an answer first
4. 3foundation · easy

   Why is it important to monitor performance metrics during model development and testing?

   • ATo document the model's business requirements
   • BTo ensure the model is deployed to production automatically
   • CTo reduce the size of the training dataset
   • DTo detect regressions in model performance early
   Reveal Solution Discussion Select an answer first
5. 4application · medium

   A data science team is developing a churn prediction model. They have a test set with a fixed label distribution, but a new data engineer accidentally added a few thousand duplicate rows to the test set. The team runs the model and sees a slight improvement in accuracy. What should the QA lead do first?

   • ARetrain the model on the augmented test set to incorporate the new data
   • BCompare the new accuracy to the previous accuracy and if higher, approve the model
   • CInvestigate the test set integrity and re-run the evaluation on the original, clean test set
   • DAccept the improvement and document it in the model report
   Reveal Solution Discussion Select an answer first
6. 5foundation · easy

   Which practice is considered a best practice in AI/ML model development?

   • AAvoiding documentation to save time
   • BUsing a single training dataset without validation
   • CVersioning the model code, data, and parameters
   • DTesting the model only after deployment
   Reveal Solution Discussion Select an answer first