Darwinbots Forum
General => Off Topic => Topic started by: deoxymoron on April 25, 2010, 07:32:06 PM
-
Very interesting site about actual robots learning their environment. food for thought- relatable to DBs, kind of...
site:
http://ccsl.mae.cornell.edu (http://ccsl.mae.cornell.edu)
robot learning:
http://ccsl.mae.cornell.edu/emergent_self_models (http://ccsl.mae.cornell.edu/emergent_self_models)
-
"Robot generates several self-models that match sensor data collected while performing previous actions. It does not know which model is correct."
Thats nice and all, but there could be thousands of possible connections, many of them aren't correct, and many will appear to be correct with the data it has. Also, it leaves the proccess of actually picking "self-models" pretty vague. Does it pick them out of a hat? Thats not usually a good way to solve complex problems. Still, I like the concept, I just wish it would be a little less vague.
-
Interesting. I'd be interested in knowing how long it takes an actual physical robot to figure out locomotion. And yeah, as Househalter said, how is it picking self models?
-
Interesting. I'd be interested in knowing how long it takes an actual physical robot to figure out locomotion. And yeah, as Househalter said, how is it picking self models?
actualy quite fast see the movie http://ccsl.mae.cornell.edu/research/selfm...ent_428x240.wmv (http://ccsl.mae.cornell.edu/research/selfmodels/videos/resilient_428x240.wmv)
-
VLC won't play it
-
It's a neat video except the gait it evolves is really really suboptimal. Might just be that the robot's actual design sucks, though (hardware issue instead of software issue).
-
If I see the video it suggests it evolves a model of its self, not an optimal walking stragegy