Allows machine learning models to take from different formats. Take text, image, video at the same time.