Search:

teak teak furniture dansk teak teak lumber teak wood teak patio furniture teak coffee table teak garden furniture teak outdoor furniture teak oil teak table danish teak teak chairs teak molding teak meubelen teak bench teak credenza teak patio teak chair teak flooring teak dresser

Autonomous Helicopters Teach Themselves to Fly Stunts

Teak Furniture Center
Teak Furniture Center Teak Furniture Center
Teak Furniture Center

Stanford computer scientists have developed an artificial intelligence system that enables robotic helicopters to teach themselves to fly difficult stunts by watching other helicopters perform the same maneuvers. The technique is known as "apprenticeship learning." The result is an autonomous helicopter than can fly dazzling stunts on its own. Stanford University: http://www.stanford.edu/ Stanford News Service: http://news-service.stanford.e... Stanford on YouTube: http://www.youtube.com/stanfor...

Channel: Education
Uploaded: December 31, 1969 at 6:59 pm
Author: stanforduniversity

Length: 03:25
Rating: 4.74
Views: 27225

Tags: apprenticeship  artificial  autonomous  computer  Helicopter  intelligence  robotics  robots  science  technology  

Video Url:


Embed Code:

Video Comments

krbosak (December 31, 1969 at 6:59 pm)
chemical, this is a general discussion on what AI and neurals can do. My point and probably your is that they can interpolate smoothly nonlinear multidimensional data with best quality. But most ppl claim this can provide reliable extrapolations towards new maneuvers. I agree, but those extrapolations are in rpactice so close to tested soutions that it will be hard to call those as new maneuvers. Certainly their helico will make really beautiful rolls.
robertlam18 (December 31, 1969 at 6:59 pm)
So was the red helicopter flew by human or their program?
drummerboy827 (December 31, 1969 at 6:59 pm)
This is completely different. In this case, they have written a program to mimic maneuvers done by a heli flown by a pro. Not too difficult to do given they know the inputs the pilot gave and how the heli reacted. It's just a matter of converging on a solution that produces the same results.The DARPA challenge requires the cars to make decisions based on their surroundings and a destination. My favorite is they have to know the 'cue' of cars at a 4-way stop sign, and react accordingly.
osgeek (December 31, 1969 at 6:59 pm)
It's more than just playing back a recording. It's imitating the moves and even improving on them. Big difference. This kind of thing is a big step forward.
chromophore (December 31, 1969 at 6:59 pm)
Andrew Ng is an extremely cool prof!
tronjon1551 (December 31, 1969 at 6:59 pm)
they will take over the world!!!!!!!!!!!!
UnconcernedCitizen (December 31, 1969 at 6:59 pm)
I am bored and this video did not help.
caracalla212 (December 31, 1969 at 6:59 pm)
Félicitation Monsieurs, j'ai découvert votre performance sur un site français dédié aux sciences. C'est du très bon travail. Encore une fois félicitation. Great job from Paris France.
monaghandavid (December 31, 1969 at 6:59 pm)
This type of learning should be used for the vehicles in the DARPA challenge or later autonomous vehicles.
XxnesiexX (December 31, 1969 at 6:59 pm)
i'm going to work fot standford if thats what you do all day

Teak Furniture Center



Teak Furniture Center © 2007 All Rights Reserved.