Explainable AI
Explainable AI studies methods that make model behavior more interpretable through explanations, feature attributions, counterfactuals, or examples.
Definition
Explainable AI studies methods that make model behavior more interpretable through explanations, feature attributions, counterfactuals, or examples.
Technical Context
A post-hoc explanation is not a complete copy of the model's internal causal mechanism. In critical systems, explainability belongs alongside validation, calibration, and independent testing.