DARCI: A Computer With Great Taste


Friday, July 06, 2012

DARCI DARCI (Philip Graitcer)

To make art, a computer first needs to understand what art is. 

A group of computer scientists at Brigham Young University is attempting this by feeding their program images by the thousands and describing those images. Digital Artist Communicating Intent (she goes by DARCI) recognizes about 2,000 adjectives so far, including terms like peaceful, scary, and dark. The goal is to teach DARCI to pick out those visual qualities in artwork — and ultimately, to write algorithms modeling creativity for artificial intelligence. 

Last year, the team took DARCI out for a spin at the Conference on Creativity and Cognition in Atlanta. They invited artists to put their work on a thumb drive, upload to the program, and be judged by DARCI. The program scored works according to simple criteria, which were kept secret; the accepted work was displayed in a temporary exhibition at the High Museum.

Several experienced artists had their work rejected; so did George, age six, who was skeptical of the algorithmic curation. “I can’t believe a computer can do it, because there's never good art or bad art — there's different types of art,” he says.

BYU professor Dan Ventura says their effort wasn’t much different from the normal operation of the art world. “Whenever you enter art into a juried show, you're often not told what the criteria are,” he reasons. “Somebody is going to make a judgment about your art and decides whether you get in or not. And DARCI is that someone right now.”

(Originally aired: December 16, 2011)

→ Teach DARCI how to associate adjectives with images


Slideshow: DARCI evaluates art

Philip Graitcer

Brigham Young University computer scientists Dan Ventura, David Norton, and Derrall Heath (from left to right) are developing a computer program that analyzes artwork — Digital Artist Communicating Intent (DARCI). By feeding DARCI thousands of images and adjectives, the programmers are teaching the computer program to recognize specific visual qualities.

Philip Graitcer

Last month, the team brought DARCI to the Conference on Creativity and Cognition at the High Museum in Atlanta. They invited artists to upload their images to be judged by DARCI. The program scored works according to simple but secret criteria; images scoring over 70 out of a possible 100 were "accepted" while lower scorers were "rejected." The accepted work was displayed in a temporary exhibition at the museum. The image on the laptop screen, submitted by Kellam Mattie, scored a 73.

Mattie Kellam

DARCI gave another piece by Kellam Mattie a 78.

Gina Deininger

DARCI gave artist Gina Deininger's triptych of female figures a 20, making it ineligible for the exhibition.

Gina Deininger

Deininger's abstract map was also rejected by DARCI.

George Mattie

Reporter Philip Graitcer's six-year-old grandson, George, scored a 24 for his paper collage portrait.

Philip Graitcer

Programmers fed DARCI an image of Pablo Picasso's "Les Demoiselles d'Avignon" (1907) — it was also rejected with a score of 26.

    Music Playlist
  • Op Hop Detonation
    Artist: Stereolab
    Album: Cobra And Phases Group Play Voltage In The Milky Night
    Label: Elektra / Wea
    Purchase: Amazon


Philip Graitcer

Comments [3]

the whole premise of this show is total complete and utter denial of the fundamental facts of what art is. scientifically proven facts of art realized as long ago as by aristotle who describes factually and accurately the fundamental design feature of ART in whatever form it takes: first of all it is social. It is discourse. Objects of art are artifacts which lose all their meaning taking out of contexts. The letter "C" in Turkish is pronounce [dj] but in English [k] or [s]. cacaphony and cistern for example but turkish "hoca" is pronounced [hoja]. the Iroquois Gospel of Handsome Lake is a bunch of pretty beads on a string if you don't have one of the half dozen authorised reciters who tell you what each bead says. A crucifix depicts the worst humans do to each other [until you believe in the rest of the story] Five times a day in Arab countries people shout out the time out the window [unless you know it MEANS come to pray] the absurdity of treating art as object with no soul mean hopelessly stupidly designing "Artificial Intelligence" deprived of the key fact of what you need to know what art is. The Call to Prayer was invented as an art form by a musician and not by the Prophet [saws] 99% of human art started out as prayer, or as some sort of spiritual discipline. Theater in classical greece was part of religious rituals. whether you call it magical spiritual religious philosophical art throughout human history has always had meaning and meaning is social, discourse in a discourse community. not something dead hanging on a wall like a trophy fish. this whole program rejects the fundamental hard facts of what art is. no wonder you can't program it. It is like you are eating cookbooks and wondering why they don't taste good.

Jul. 09 2012 03:02 AM

ameobas are creative. actually. in ways computers never can be. our creativity serves in a system of values i maintain for and about myself. emotional intelligence emotional maturity is the foundation of creativity. art resovles cognitive dissonances that are unresolvable any other way. there is a fundamental necessary dissonance when self awareness serves self preservation. all human creativity acts on that problem which is like goeddels laws unresolvable. it starts with ameobic and other unicellular community. creativity is social. call it attachment theory :-) creativity is resolution. computers are not in conflict. creativity is existential. give computers a reason to be creative and that's a creativity problem they will solve.

Jul. 08 2012 01:01 PM
Isa Kocher from Kucukcekmece

creativity as an outcome is meaningless if there is no consequence. these machines neither gain nor lose regardless. in evolution there are real world consequences and creativity is NOT an outcome. creativity is functional.

all of these projects bear no relationship to creativity in any real way. the computer has to have a stake in the outcome before it can ever be "creative." there clearly is an evolutionary process going on here in that creators endow machines with more of themselves the more the machine engages in creativity producing activities. until the computer can itself choose for its own sake can it ever be creative in any real sense of the word. then computers will be able to organize and reorganize itself to achieve its goals and obtain rewards it values.

humans have very practical value systems: it i don't stick my foot in where it doesn't belong i don't lose my toes. if i give what i can't eat to others, the chances are others will reciprocate for mutual benefit. computers don't care because they have no choices.

as we build in choice and value into computers at some point we'll build in awareness.

Jul. 08 2012 11:54 AM

Leave a Comment

Email addresses are required but never displayed.