Towards Explainable Artificial Intelligence