g., textual content or visuals), and it might instantly figure out the set of attributes which distinguish distinctive groups of data from one another. This eliminates a lot of the human intervention required and allows using huge amounts of facts. You could consider deep learning as "scalable