Robot Self-Evolution Talk @ TED
Writing by Evan Ackerman on Monday, 15 of October , 2007 at 2:26 am
Natasha from the TED (Technology, Entertainment and Design) Conference has just let us know that a new talk by Hod Lipson on self aware robots is now available for viewing on their website. You can check out the video here (and it is worth watching), but basically Lipson discusses two robot self-modeling projects of his. One of them is the Starfish quadruped robot that’s able to teach itself to walk (we wrote about it a few months ago), and the other project is the Golem Project, which some of you may remember from back in 2000. Read more after the jump.
The Golem Project was an attempt to evolve viable (i.e. moving) mechanical life forms through mutations and natural selection, with idle computers combining basic mechanical bits and pieces and preserving the best results. One of the early stabs at distributed computing, you could download the Golem@Home program to your computer (which I did, incidentally), where it would chug away, creating life while you were on the couch watching TV. The virtual bots created by your computer were uploaded to the Golem server, where they were “raced” against other virtual bots, until a few designs came out on top. The winners were constructed; you can see them in action in this video:
Here’s what the Golem Project concluded from the whole experiment: “After accumulating several Million CPU hours on this project and reviewing many evolved creatures we have concluded that merely more CPU is not sufficient to evolve complexity: The evolutionary process appears to be hitting a complexity barrier that is not traversable using direct mutation-selection processes, due to the exponential nature of the problem. We are now developing new theories about additional mechanisms that are necessary for the synthetic evolution of complex life forms. Some of these new mechanisms are based on ideas of modularity, regularity, hierarchy, symbiosis and co-evolution. These ideas are resulting in a new generation of artificial-life systems.” Hmm. I might have put it a bit more positively, but at least they’re still working on the concepts. Here’s my idea, though… Why not just design a program that evolves a better evolutionary program, huh?
[ TED Talks ]
Leave a comment
Category: Artificial Intelligence, Research
- Add this post to
- Del.icio.us -
- Digg
No comments yet.