Thanks to Deep Learning, Computers Are Learning to be Creative

Human beings are naturally creative creatures. Before we learned how to write, we learned how to create art. Most notably, we learned how to tell stories through art, e.g. the paintings and etchings that can be found in ancient caves. But what if computers learned how to create art as well?

Thanks to a byproduct of Machine Learning known as Deep Learning, neural networks are beginning to learn to not only visualize objects, but to encapsulate it all and to generate creative works of art as a result.

The TED Talk provided below presents Blaise Agüera y Arcas, speaking about his work with Google in developing machines that are capable of creativity. These machines – technically neural networks – use Deep Learning to quantify images into data, only to then reverse engineer them with perceptual awareness of what they’re looking at.

As a result, Arcas and his team are now able to teach these machines how to express their perceptions in ways of which we would’ve never thought possible. Many today might even deny that computers are able to be creative, claiming that only humans are uniquely capable of such.

Though, as Arcas noted in his talk, “perception and creativity are by no means uniquely human. We start to have computer models that can do exactly these sorts of things. And that ought to be unsurprising; the brain is computational.”

Sorry, humans. It looks like there’s a new player in town and its creative works so far are on fleek!