Interpretable Machine Learning -- A Brief History, State-of-the-Art and Challenges