Inconvergent: The Algorithmic Art of Anders Hoff

by admin, October 2, 2016
  •  
  •  
  •  
  •  
  •  

dg_d4e_01_andershoff_600x400-002

Near a window, steam rises in spirals from a cup of tea. A column of sunlight intervenes, producing emergent ribbons of heat and ephemeral color. It’s an ordinary cup of tea, and steam is an ordinary product of nature’s processes. But to say that there is no art in nature is only a question of agency and scale.

This is a question Anders Hoff might ask. He is a search engine consultant and a mathematician in Oslo. He is also an artist, but a kind of artist that’s new on the scene, one with a different sort of medium–the kind you’ll start to see more of.

Anders is an artist of networks.

When he’s not spending six to nine hours a day solving complicated problems in search, he’s most likely listening to metal and using math to create generative artwork like this:

pic1inconvergent

It takes a few days to produce something like this. He writes a type of mathematical function, a generative algorithm, that defines the behavior of an agent-based data set. An agent is just something that acts on a network of things. When you’re creating an artificial network on a computer, you can define what you want the agents to be and do–to a point.

The agents are linked to other agents, and you can give them simple rules, but–like the simple rules in nature that create incredibly complex patterns–the agents are part of a network, and together they evolve the overall network in ways you can’t anticipate.

pic2inconvergent-300x300

In the case of this piece, which Anders describes as something akin to flower petals, or cabbage, or the inside of your intestines, the agents are the vertices of the mesh. You start with a simple triangular mesh and the vertices interact with each other in some way. Every vertex is connected to five or six other vertices. This creates the whole mesh–the network. These vertices act on simple rules that attract them to their neighbors and cause them to avoid the unconnected vertices of their mesh. In the end (if there is an end), it’s the complete network of all the vertices interacting that makes the mesh move or behave and evolve as it does.

He creates with a beginner’s mind.

“I like how you can get interesting structures arising from these simple rules,” Anders says, and it’s that simple curiosity that drives him to create algorithms that evolve into living artifacts that resemble familiar patterns in nature, analogs to biological mechanisms, like this:

pic3inconvergent

He doesn’t have an end result in mind. “I’ve tried to make it as naive as possible and still see if I can get that behavior. Nature doesn’t solve differential equations, but nature does evolve – so I want to make something as naive as possible.”

kelp

Sometimes he sets out to make a thing, and he makes exactly that one thing. But most times he starts with nothing other than a vague idea, perhaps an interesting mesh that someone else has made, and plays with it until he’s tired of it. That’s how this all started. When he was 16, Anders stumbled upon a site called Complexification, which was created by Jared Tarbell, one of Etsy’s cofounders. The site featured animation created by agent-based systems. Anders started playing with Tarbell’s algorithms, copying them at first and then making his own original designs.

Years later while studying physics and working on his master’s in numerical mathematics, Anders stumbled upon Complexification when he was supposed to be cramming for his finals, and out of curiosity and a bit of procrastination a whole portfolio of generative art emerged.

differential_9k

Back to our agency question. Where does the work of one network artist end and another’s begin? How does the artist know when the art is complete?

With art that grows, it’s difficult to say. When it feels complete, when it feels original, when it’s time to do something new.

For Anders, it’s a curiosity-driven process, the reward of creating simple behavior and producing something interesting. He puts his algorithms on github for other coders to play with and hopes someone will make something entirely different. Evolution is the propagation of novelty in nature, and even in art—where nature and the subconscious are exposed in complicity.

In an emerging medium like generative algorithmic art, you could think of the network itself as the artist. A network of artificial agents designed by a network of generative artists, evolving through each node, shown to other networks as art.

“The waking have one world in common, whereas each sleeper turns away to a private world of his own.” – Heraclitus

To learn how Dialog can help your business, contact us at 512.697.9425 or LetsChat@DialogGroup.com.

This article was originally published in design4emergence, a network science community sponsored and nurtured by Dialog and Panarchy, the world’s first network design firm.

Comments are closed.