QuestionQ80

AI Governance and Risk

An IS auditor is reviewing the change-management documentation for an AI model. Which of the following presents the GREATEST risk to the model?

Explanation

Hyperparameter tuning should be performed using a validation set that is separate from the training data. Tuning on the training set can overfit both the model and the tuning decisions to that data, producing performance results that may not generalize to unseen data.

Community Discussion

No comments yet. Be the first to start the discussion!