MIT Taxonomy Helps Build Explainability Into the Components of Machine-Learning Models