Eliminating the Hidden Black Boxes in Machine Learning Models