Vassar college Professor John H. Long is a marine biologist, by training, and, now, a roboticist by trade. Essentially, he builds robot populations closely modeled on extinct (and living) fish, and then subjects them to simulated evolutionary pressure—to hype it up a bit, he “pits them against each other”—to learn things about why historical animals evolved as they did.
I was skeptical, at first. Seems like a cool idea, but can you really build a robot that’s enough like a real fish to draw reliable conclusions? Then I listened to this video interview, embedded above, from Vassar’s newspaper The Miscellany News.
And I’m still skeptical. But less so. John is a passionate and engaging talker, and has had a bunch of papers in Zoology and other prestigious journals with titles like “The Importance of Body Stiffness in Undulatory Propulsion,” “Swimming fundamentals: turning performance of leopard sharks (Triakis semifasciata) is predicted by body shape and postural reconfiguration,” and “Fish out of water: terrestrial jumping by fully aquatic fishes.”
As these titles hint, Dr. Long’s focus is really on biomechanics, e.g. how one body shape might confer behavioral advantages over another. Which makes it easier for me, at least, to understand how his approach could actually work:
We were interested in the evolution of structures like the backbone, for example, but we weren’t interested in the evolution of the brain. So we didn’t touch the brain. So all we did was evolve the body. So here’s a big surprise: You evolve the body, and you get smarter robots. You don’t need to touch the brain to become smarter. You can just have a “smarter body,” if you will. And the reason that’s a surprise is because, as humans, we’re so focused on our heads, and the giant size of our brain—right?—that we think this is the ultimate answer to any question about animal intelligence or human intelligence. Sure: Brains are important. But they’re not all that’s important.
Dr. Long has a book out, and will be speaking at World Maker Faire New York this Saturday at 4:30PM. Still skeptical? Come on out and hear what he has to say.
4 thoughts on “Maker Faire New York: Studying Evolution with Robots”
It’s really neat but I’m not sure I’d call it evolution. There’s an intelligence that is doing the “pitting” and deciding what constitutes a “win”. Just like pretty much any evolution model I’ve seen, there’s an intelligence involved.
I wish I knew how to build robots. That was a childhood dream of mine… inspired by R2D2, of course.
Comments are closed.