On artificial creativity
Computer-generated composition involving the concept "terrorism".
Creativity
What we call human creativity is the conceptual ability to be innovative, to combine existing objects and concepts in new ways, for new purposes. This is often a trial-and-error process, with the actual original idea often emerging as a side effect from something else (consider how the immensely popular Post-It notes actually emerged from a research trying to develop an extremely strong glue).
Creativity is not limited to art, it is merely our word for finding innovative solutions.
Artificial creativity tries to mimic this trial-and-error process in computer terms, tries to tell a computer algorithm how creativity works. Artificial creativity is closely related to something called emergence in the field of artificial intelligence. Emergent behaviour (in nature as well as in a computer program) appears when a number of simple entities develop more complex (often unexpected) behaviour as a group. Think about ants: a single ant has little or no intellectual or creative capacities: it just follows trails of scent and nothing more. But as a collective, the ant colony is capable of finding creative solutions to complex problems: from finding the shortest path to food to organising a full-scale army, maintaining burial grounds and herding caterpillars. If you observe ant colonies in bird view, they almost look like a computer network. Hofstadter has often used this ant example, talking about Aunt Hillary (the ant hill) as a sentient being.
Artificial creativity works in the same way as an ant colony: a number of simple algorithms (or agents) work together, share and distribute information in the hope of developing emergent behaviour to solve new and unexpected problems.
Choices and decisions
Art and graphic design are visual things, they speak and communicate visually.
And the tricky problem here is that a computer has no equivalent for eyes. This makes life extremely hard for an agent trying to be artificially creative. How can it assess/evaluate what it created? How can it see its own work? Did it create something innovative? Did it live up to the expectations? How can it experience the feeling of getting a composition right, nudging the elements a bit more to the left to get the visual harmony exactly working? Or experience the feeling that things looked better when they were inside its head... and look kinda boring once visualised?
A simple answer would be: there is no need for a computer algorithm to be able to see its own work: it takes its pleasure in other factors, like the pure mathematical harmony of numbers. However, what would the practical use of such a program be, to us as people? We would naturally like the agent to design stuff humans like - we don't see the magical harmony of equations, or the tally of tail recursions. We like colors that fit well together, interesting contrasts, and intense composition in art, things we can see, and ponder on. But let's say we are broad-minded, and accept the fact that artificially creative programs will design things all by themself that look strange to us, things that take getting used to. We are broad-minded. We want to expand our creative horizons. We don't want art to die in a circle of repetition.
So the computer is free to do what it likes. And we are talking about art, not commercial design. Now there is an even bigger problem. When is an agent satisfied? One the common pitfalls in art and graphic design is the decision-making process. When is a project finished? Do you add a small black line to the layout, or not? If blue and red are both good options, are we going to use blue, or red? Detailed choices like these can drive an artist nuts: he or she might start out in blue, revert to red, back to blue, start working with lines, reverts to circles, to end in green (and find out the audience doesn't really has an appetite for green).
To a computer, choices and decisions are a complex problem as well. It can either act totally random, or totally determined (actually this is not entirely true but I'm simplifying). Total randomness is not creativity however. The really neat agent, the one that is determined and knows what its doing, needs expertise, it needs a base from which to draw conclusions. This is the same with any human artist: after some practice he or she will tackle choices more easily, based on past experiences on what works and what not.
Relations
The agent needs to be able to draw determined conclusions from a base of expertise and knowledge. The agent needs data it can interrelate, combine, juxtapose in an infinitely creative number of ways to reach a determined conclusion. It needs other agents that act in a similar way with which it can talk, share opinions. This is where the real creativity takes place: in the conceptual phase of creating, in the linking of facts to other facts, the joining of fact and fantasy, the translation of words to new words, the association of images to concepts. The visualisation is merely a last step that fits the conceptual model as-best-as-can. Form follows function, as they say in graphic design.
An agent asked about cheese that creates a design with pictures of cheese ripped of the internet is much less interesting than an agent that knows cheese is connotated with mice, sheep, Switzerland, the moon and uses those elements as imagery. This agent may not necessarily make something more beautiful, but definitely something more creative. If we can get an algorithm to think like a human, to make equally crazy associations, to devise equally heartquenching combinations, then we surely have something creative.
The problem of handing the Frankenstein-monster a brain, of handing an artificially creative algorithm knowledge in the field of art, is not an easy problem, but it is solveable however to my accord. In the last two years I have made a slow evolution from generative art to artificially creative art, grinding my teeth on the problem of knowledge along the way. Here are some examples, you can read more about them by following the links:
 | The internet is freely accessible to any computer. The internet is a large pool (or puddle if you wish) of human knowledge, creativity and insanity. Prism, a color-finding agent we (Frederik De Bleser, Lucas Nijs, Tom De Smedt, Experimental Media group, St. Lucas School of Arts, Antwerp, Belgium) created for NodeBox, uses the internet as its base knowledge to match the right colors to the right concept. |
 | Flocking and swarming algorithms, like boids have been around for years, and as a system are excellent to describe natural composition, instead of using stale grids. Using boids we made something called Blines, a proof-of-concept of an agent that finds a harmonious composition all by itself, based on its knowledge of organisation in nature: bird flocks, fish schools, spiralling formations, the golden ratio. |
 | Semantic databases like WordNet and ConceptNet have also been around for a while now, and are excellent for a creative agent to use and misuse the relations between words and commonsense data. Flowerewolf is a poetry agent we created for NodeBox that writes poetry on any given subject, by moving from synonym to synonym in WordNet, fetching out consonating nouns and adjectives, creating alliterations, and so on. Photobjects is a database of thousands of images loosely linked to ConceptNet, which should create strange compositions matching the feel of any given concept. I'll tell you about the should in a minute. |
Social relevance
There is a nice paper by Rob Saunders and John S. Gero (Key Centre of Design Computing and Cognition, University of Sydney) online, entitled Emergent Notions of Creativity in Artificial Societies of Curious Agents.
Agent-based algorithms derive their strength from their social environment. Likewise, interesting art and graphic design have a social relevance. Either it is telling us something about the state and fate of the world, or it is providing a glimpse into the mind of the artist who sees the world in a different way, or we are confronted with a series of war photogrpahs, or harassed with commercials - in all cases the work of art is playing a social role between the artist, the audience, and the world.
Art and design want to uncover things, reveal and reinvent the world.
The first experiments with Photobjects didn't really work out, the computer generated some funny image compositions, but they had no real social meaning. One of my current experiments involves coupling the Photobjects agent to an online agent here on Replica called Hyperpolator. Hyperpolator was a joke; it tries to summarise what is going on in the world in key concepts. Admittedly, it has a totally pessimistic and twisted view on the world (usually involving the concepts terrorism, president, war and sex), but as a proof-of-concept it is OK enough to feed Photobjects with socially relevant information.
A third agent (PhotoBot) attempts to process the compositions Photobjects and Hyperpolator create to make them more elegant and designish. All combined, the computer created artwork for an abstract assignment like: make something that visualises the world today. The result is something a second-year graphic design student would create.
What our little setup needs now is the ability to learn.
Computer-generated composition involving the concept "terrorism".
Computer-generated composition involving the concepts "religion" and "war".
Tom De Smedt, PhD. on Artificial Creativity, St. Lucas School of Arts, Antwerp, Belgium, 2005
http://research.nodebox.net