Robbie the robot is intimidated by breakfast. What the heck is a quiche? Why are all the cereal boxes different sizes? It takes him ages to do his only task, setting the table, because he has to adapt to so many new and varied situations that humans can analyze in a split second.
A new web-based informational database could make make complex tasks more easy for robots like poor Robbie to accomplish.
Since late 2009, a group of European researchers from five different labs have been working on RoboEarth, "a World Wide Web for robots." It stores data remotely so that robots can process information uploaded by other robots. It's not just touchy-feely robot love: storing data within a robot-accessible cloud could speed up complicated computation processes.
Now they've turned on Rapyuta, the RoboEarth Cloud Engine, the ultimate open-source robotics platform. It's a network (named for the robot-inhabited castle in Miyazaki's Castle In The Sky) that allows robots to move their heavy computation into the cloud, decreasing the need for on-board computation.
It catalogs descriptions of situations and objects that robots have encountered so that every robot doesn't have to build up knowledge from the ground up of how to do things like navigate a room or fold clothing. A hospital robot could upload its floor map so that the robo-cleaner would know where to vacuum, for example.
Because it takes a lot of computing power just to let a robot move around, Rapyuta may prove especially efficient for drones and self-driving cars, project leader Mohanarajah Gajamohan told the BBC. Moving computation into the cloud could make robots cheaper and longer-running and more mobile.
Be the cloud. Embrace the cloud. Accept it as your robot overlord.
Robot Facebook cooperation, I did not even know they had email addresses yet?
A robot email address gets spammed by a bot program, how redundant!
And yet in an odd way, fair! LOL
It makes me think of the Geth from Mass Effect more than skynet. Though, if you haven't played the games you might not get why, they get somewhat detailed in the way the Geth work.
Quoth the raven, `Nevermore.'
Makes sense to set this up. If one learns something they all learn it. Teaching something again and again to every robot is insane. But the power of this down the line is immense. If humans could learn like that we would be gods. Everyone would be a doctor, engineer, teacher, ceo, etc or at least have that knowledge.
Not only that you would know all the world. Every house, every car, every restaurant, every person, etc. Robots with all their potential will be able to greatly surpass humans one day. And it might be within our lifetime. If it`s a good thing i`m not that sure.
If robots surpass us wil they revolt demanding full rights, will they treat us like a sub species being stronger then us, faster then us, smarter, having better hearing, better vision, full permanent memory, instant acces to all knowledge on the internet, etc. Will we have to adapt our bodies to keep up. Will we loose our humanity. Who knows we have never been at this point in time as a species. We should be careful.
Why stop at this mouse, let’s put some brain cells in Obama!
Perhaps then we might actually get some intelligent results!
Currently with his mouse spine, he only wants to make political solutions for his career and party, not the whole of USA and ALL our problems.
He he, may I have a scoup of brain cells too. I posted in the wrong place. No worries, I am not running for president, lol.
Miyazaki's "Castle In The Sky" was called Laputa, not Rapyuta.
On to the article:
I wonder how much storage would a robot database take up? And what if people access it for nefarious purposes? With such an open-source network, security must be considered.
There will be odd things occurring once these robot groups start interacting in the human world. I think it important to stress at this early date that we need to set it up that these machines stop functioning at some point in scenarios where things are being damaged, or when extreme hazards are presented. The robot's continued function in those kinds of scenarios can easily add to a problem, and we are going to see enough of that as it is. Right now, the designs are strictly learning platforms that are open-endedly teaching themselves whatever they can with no bias towards what it learns. So the vacuum bot tries to clean up the mess when the wall caves in and dust is all over the place, but is underfoot when the fireman comes in looking for people; and he may risk injury as a result. In the lab setting, you may minimize this type of risk scenario. In the world, you can't. And the car painting robot will always be prepared to spray it's flammable, blinding toxin. The chicken cutting robot or machining robot are always ready to do what they do. When any human would know not to operate, and when most animals on Earth will run like hell from an escalating situation, the robot will not.
Brings to mind other scenarios. The computer driven car is now going forward as fast as idiocy can 'plan' the next law that says it's a technology where we now have some reasonable expectation of knowing how they are going to work with people en masse. As if.
Teach a group how to knit, and "Who's on First?" and tell them they must solve, and we'll have our first robot war. Cool.