Towards Robust Interpretability with Self-Explaining Neural Networks | Read Paper on Bytez