|
|
|
|
|
|
|
Researchers at the Max Planck Institute for Mathematics in the Sciences have demonstrated software for robots that allows them to "learn" to move through trial and error, using an artificial neural network.
|
|
|
|
|
Contributed:
|
|
Added by:
|
|
Viewed:
|
|
|
On august 12 2008
|
|
Andres
|
|
775 times
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Member comment:
Have something to say about this video? Signup and
login to share your thoughts with the community.
|
|
If they seem awfully clumsy it is because these artificial neural networks are still learning to move to a target location. I wonder how this would look several generations down the road. Would it improve drastically thousands of generation afterwards by sticking with the same algorithm or would it not really improve much at all?
|
|
|
Add Videos
