Machine learning interpretability: the ability to understand how a machine learning model arrives at its predictions and decisions.