LIME (Local Interpretable Model-Agnostic Explanations): A method for explaining the predictions of any classifier by learning an interpretable model locally around the prediction.