Towards A Rigorous Science of Interpretable Machine Learning

   Abstract