AI Model Training, Testing, and Validation AAIR Practice Questions (Page 1)

Questions 1–5

1. 1foundation · easy

   What is the primary role of the validation dataset in model development?

   • ATo train the model's parameters
   • BTo tune hyperparameters and select the best model variant
   • CTo provide a final unbiased estimate of model performance
   • DTo augment the training data with additional samples
   Reveal Solution Discussion Select an answer first
2. 2foundation · easy

   Why is the test set kept separate from the training and validation sets?

   • ATo increase the amount of data available for training
   • BTo serve as a backup in case the training data is lost
   • CTo provide an unbiased evaluation of the final model's performance
   • DTo allow for continuous hyperparameter tuning after deployment
   Reveal Solution Discussion Select an answer first
4. 3application · medium

   A facial-recognition system used for building access is being evaluated. The security team wants to test whether an attacker could bypass the system by wearing a printed mask of an authorized employee's face. Which testing approach directly addresses this concern?

   • AStandard accuracy testing on a large dataset of employee photos
   • BCross-validation on different demographic subgroups to check fairness
   • CA/B testing the model against a different facial-recognition vendor
   • DAdversarial testing with physical-world perturbations such as printed masks and patterned glasses
   Reveal Solution Discussion Select an answer first
5. 4foundation · easy

   What should be included in an audit trail for AI model development?

   • AThe names of all employees who worked on the project
   • BVersion control of datasets, code, and model artifacts
   • CThe cost of cloud resources used for training
   • DOnly the final model's accuracy score
   Reveal Solution Discussion Select an answer first
6. 5foundation · easy

   In a binary classification problem with imbalanced classes, which metric is generally more informative than accuracy?

   • AMean squared error
   • BR-squared
   • CF1-score
   • DAccuracy
   Reveal Solution Discussion Select an answer first