AI Model Training, Testing, and Validation AAIR Practice Questions (Page 3)

Questions 11–15

1. 11application · medium

   A cybersecurity firm is testing an email-spam classifier. The team wants to evaluate how the model responds to adversarial inputs designed to evade detection, such as obfuscated text or specially crafted emails. Which testing strategy is most appropriate?

   • ATest the model on a large, diverse dataset of real emails to measure overall accuracy
   • BGenerate adversarial examples by perturbing input emails and test the model's classification on them
   • CCompare the model's precision and recall on the test set
   • DUse cross-validation to ensure the model performs consistently across different email samples
   Reveal Solution Discussion Select an answer first
2. 12application · medium

   A manufacturing company is building a predictive-maintenance model. The team accidentally included the maintenance outcome label in the feature set during training. The model achieves near-perfect accuracy on the test set. Which risk does this scenario most directly illustrate?

   • AConcept drift, where the data distribution changes over time
   • BData leakage, where information from the target variable is used as a feature
   • CUnderfitting, where the model is too simple to capture patterns
   • DOverfitting, where the model memorizes the training data
   Reveal Solution Discussion Select an answer first
4. 13foundation · easy

   Which metric is commonly used to assess fairness in classification models?

   • ADemographic parity
   • BF1-score
   • CMean absolute error
   • DR-squared
   Reveal Solution Discussion Select an answer first
5. 14application · medium

   A government agency is deploying an AI model for benefit eligibility determination. An external auditor will review the model's development process. Which documentation practice is most critical for supporting the audit?

   • ARecord the exact training, validation, and test dataset versions and their split methodology
   • BStore the model's source code without additional context
   • CKeep meeting notes from the development team's discussions
   • DDocument only the final model's performance metrics on the test set
   Reveal Solution Discussion Select an answer first
6. 15foundation · easy

   Which activity is part of the data preparation stage in AI model training?

   • AEvaluating the model on a held-out test set
   • BAdjusting the learning rate during the training loop
   • CChoosing the loss function for the model
   • DCleaning and transforming raw data into a format suitable for the model
   Reveal Solution Discussion Select an answer first