Arts & Literature

Desktop Da Vinci

Machines are learning the visual language of emotion.

Comments (0)
Thursday, December 06, 2012

When I was an undergrad, I had an art history professor with little patience for what he saw as the useless prattle induced by the question, “What is art?”

He sliced right through the Gordian Knot of debates about illustration versus “fine art” and advertising versus art for its own sake with a simple formula. “If people make it,” he said, “it’s art.”

He rightly pointed out that everything we make is, in some sense, designed, even if that design is constrained or completely driven by function. That’s certainly a generous approach, and a recipe for democratizing art, and gaining inspiration from the humblest phone cord or coffee cup.

Much as I admire that approach, it does leave related questions unadressed. What about the paintings done by elephants or chimpanzees? And what about the artistic imaginings of the machines that draw ever closer to artificial intelligence? Continuing all that generosity of spirit would dictate that those aren’t conundrums at all, that anything made by an elephant is elephant art, anything made by a computer is computer art.

Seeing what machines can and will devise ought to prove intriguing at the least. And according to research recently discussed in New Scientist, computers are destined to be far more capable of evoking human emotion than it might at first appear. In what may merely be further proof of European devotion to matters cultural, European scientists employed “mechanical vision” to analyze hundreds of works of abstract art at the Museum of Modern and Contemporary Art of Trento and Rovereto. In the New Scientist article, Hal Hodson writes, “The system measured how color is distributed across each work, as well as the occurrence of different shapes or outlines. Using data on how 100 people responded to the paintings, the system then worked out what emotional impact these elements had. For example, black, spiky features tended to correspond to the bleaker end of the emotional spectrum, whereas bright, smooth features were more feel-good.”

Part two of the study involved the machine attempting to predict viewers’ emotional responses, something it managed to do a stunning 80 percent of the time.

The article goes on to discuss the work of The Painting Fool, which is no fool, but rather a computer program and aspiring artist devised by Dr. Simon Colton of the Imperial College, London.

Visiting The Painting Fool’s site feels oddly futuristic. Whereas Windows and online banking systems jar with the occasional contention that documents or other content are “mine” rather than “yours,” The Painting Fool goes full-on first person. The website itself is talking to you when it says, “The aim of this project is for me to be taken seriously—one day—as a creative artist in my own right. I have been built to exhibit behaviours that might be deemed as skilful, appreciative and imaginative.”

Shades of Kubrick’s Hal 9000 are inevitable. It’s almost spooky. The artwork is remarkable in its embrace of quite human themes, but it also exhibits traits that might well be described as the first stirrings of machine style. Maybe it was because I knew I was beholding computer art, but something about The Painting Fool’s creation of images from nearly identical building blocks immediately marked it as the product of the kind of process at which computers excel, repetitive and formulaic at a micro level, yet sophisticated and complex when viewed as a whole. Colton told New Scientist that, indeed, his program was good at doing things that humans simply cannot, like “sampling every tweet on Twitter for inspiration.” 

It might be preconceived notion to say that the Fool’s work does seem a bit lacking in emotional content, dazzling though it sometimes is. But the emotional content the European machine learned to divine might well be employed by machines like the Fool in short order. What happens then ought to be especially interesting—will machines produce artwork equal in its sophistication to that of humans?

In a way, that’s the wrong question. The ability of an elephant or a computer to, if you’ll pardon the term, ape human art is only of passing interest. Art’s power to manifest and incorporate the artist’s inner world is, after all, what makes it so powerful and so useful. 

If computers learn to evoke emotion in human viewers, how long will it be before the machines can use art the same way we do, before they start revealing the contours of the landscape of machine thought? That’s when things ought to get really interesting.•

 

Comment:

Name:

Password:

New User/Guest?

Find it Here:
keyword:
search type:
search in:

« Previous   |   Next »
Print Email RSS feed

Steaming Up a Phantasy
Stagestruck: Revisiting Vengeance
The Visit at WTF: A revenge drama with music
Fade to Gold
Zing, a small studio in Westfield, makes big records.
Catching a Coyote
A Q&A with Valley and fellow Publishing Genius Press author Christy Crutchfield
Bigger, Bolder, Darker, Deeper
Anselm Kiefer turns to the dark side.
Like A Cannonball
Northampton poet Mike Young’s new book, Sprezzatura, lands with a splash
The Space Between
Technicolor and Chalk