It is suggested that deep learning approaches could be the vehicle for translating big biomedical data into improved human health and develop holistic and meaningful interpretable architectures to bridge deep learning models and human interpretability.