Machine Learning Interpretability—Validation of Self-Learning Models

How can risk practitioners integrate self-learning technologies across risk functions and ensure appropriate model-validation practices?