QuestionQ57

VerifiedTesting AI-Specific Quality Characteristics

Which ONE of the following tests is MOST likely to be useful for detecting different types of bias in an ML pipeline?

Explanation

Model evaluation can detect disparate predictions or performance across protected or relevant groups using complementary fairness metrics, exposing bias caused by either the data or the learned model behavior. AWS guidance recommends evaluating group-level performance and using multiple fairness metrics because each can expose different forms of bias.

Learn more

Community Discussion

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