An Open Source Artificial Life Project Called OpenWorm

OpenWorm is a very cool project that also scares me a little bit: a collaborative, open source attempt to construct an artificial life form -- a simple worm, computationally created from the cellular level to a point where it's sophisticated enough to solve, as the site explains, "basic problems such as feeding, mate-finding and predator avoidance". This would be the first digital life form of its kind, but if the project is successful, more sophisticated species are sure to follow.

I first heard about this open source project because OpenSim pioneer John Hurliman recently joined OpenWorm's development team, helping with improving the code's deployment processes. "In the future I'd like to help with the physical and neural simulation aspects," he tells me. "It strongly overlaps with the simulation research I did at Intel, even though we were virtual world/avatar focused, for the implementation the theory is equally applicable to a project like this."

How's progress on the worm itself going?

"Not a lot from the 'download and run it' perspective, it's a pretty massive undertaking," John tells me. "I was just in a Google Hangout meeting with the team this morning showing the latest demo, which showed a particle simulation of five connected muscle segments moving together through a body of water. The neuron simulation is also working and there have been some early tests of sending the outputs of that simulation (signals to the muscles to contract or relax) to the physics simulation. It will probably be 3-5 months before both simulations have been debugged and integrated to the point where the average curious developer could build and run the project and see the muscular shell of a worm swimming around."

That's right -- since it's open source, other teams can work on evolving worms and other organisms of their own. More news when the worm turns. (Yes, I went there.)

Please share this post with people you like:

Tweet

Comments

[Kayla]

This has me excited =D

[Arcadia Codesmith]

I've wondered if perhaps one of the factors retarding progress in artificial intellegence has been the fact that you can't divorce cognative development from sensory and physical inputs.

Creating artificial sentience may be light-years removed from building a humble virtual worm, but I see it as a productive avenue for exploration. As a gamer, I'm thinking about autonomous agents as NPCs, but that's just a very small subset of the potential.

[Giulio Prisco]

So who is going to suggest to build a Second Life interface to OpenWorm and put artificial life worms in the virtual wild?