/ News
Deep Learning Robot Takes 10 Days to Teach Itself to Grasp
Leave a human baby with some toys and it’ll quickly learn to pick them up. Now a robot with deep learning capabilities has done the same thing.
One of the goals of general purpose robots is to interact in an intelligent way with everyday objects. But robotic grasping skills are embarrassingly poor. Ask a robot to pick up a TV remote or a bottle of water or a toy gun and it will endlessly fumble with it—unless specifically programmed to pick up that object in a specially controlled environment.
That’s in stark contrast to human grasping capabilities. A human baby quickly learns to grasp such objects, even in the most cluttered and unstructured environments.
And therein lies an important clue. Could robots learnt to grasp like babies, by repeated trial and error?
Today, Lerrel Pinto and Abhinav Gupta at Carnegie Mellon University in Pittsburgh show how this is possible. These guys have equipped a robot called Baxter with powerful deep learning capabilities, placed a table full of ordinary objects in front of it and then left it to learn, like a baby playing in a high chair.
Baxter is a modern two-armed industrial robot designed to perform repeatable tasks in environments such as factory floors. Each arm has a standard two-fingered parallel gripper and a high resolution camera to allow the robot to see what it is grasping close up. It also has a Microsoft Kinect sensor to provide an overview of the table in front of it.
Pinto and Gupta programmed Baxter to grasp an object by isolating it from its neighbors, then to pick a random point in that area, rotate the grippers to a certain angle and then move the grippers down vertically to attempt a grasp. The robot then lifts its arm and determines whether the grasp has been successful using force sensors. For each point, it repeats the grasping process 188 times, each time after rotating the gripping angle by 10 degrees.
To allow the robot to learn, Pinto and Gupta placed a variety of objects on the table in front of Baxter, and simply left it for up to 10 hours a day, without human intervention. If the robot dropped an object on the floor, there were plenty of others it could continue with.
Baxter’s deep learning approach was fairly standard too. The researchers gave Baxter a conventional neural network that had been pretrained in object recognition, thereby giving Baxter some essential skills to start with. However, two layers of this network were devoted to learning to grasp in these random experiments.
The team then used a second stage of learning to improve Baxter’s skills. Having picked up some basics, the team used these on a new set of objects, some of which Baxter had already seen and others that were entirely novel.
Over 700 hours, Baxter performed some 50,000 grasps on 150 different objects, each time learning whether the approach was successful or not. These objects included a TV remote, a variety of plastic toys such as guns, cars, and so on, and other similarly sized objects. That produced a significant body of learning which allowed Baxter to predict whether or not a grasp would be successful almost 80 percent of the time.
To find out just how good it was, Pinto and Gupta compared Baxter to a number of other approaches. For example, one of these was a common sense approach in which the robot was programmed to pick up an object near its center, at its narrowest point while avoiding areas that were too narrow to grip (since the grippers could not close completely).
This lead to a grasping accuracy of only 62 percent, while other approaches were only a little better. None matched Baxter’s performance.
That’s interesting work that could have an important impact on the way robots like Baxter interact with the world. A key part of this approach is that it occurred in a cluttered, relatively unstructured environment like the ones humans cope with easily. What’s more, the grasping skills were more or less entirely self-taught.
Of course, Baxter and its convolutional neural network have a long way to go before they can match the grasping skills of a young child. An important skill to learn next will be how hard to grip, something that is important when handling delicate objects without damaging them.
Perhaps the ultimate test for Baxter will be the toothpaste challenge—the successful placement of a pea-sized blob of toothpaste on a toothbrush. That’s something humans learn at a young age (while mysteriously forgetting it again during their teenage years).
But at the rate robots like Baxter are learning, it won’t be long before humans lose their mastery of toothpaste world, just as they lost their mastery of the chess world almost 20 years ago.
Ref: arxiv.org/abs/1509.06825 : Supersizing Self-Supervision: Learning to Grasp from 50K Tries and 700 Robot Hours
Source: http://www.technologyreview.com/view/542076/deep-learning-robot-takes-10-days-to-teach-itself-to-grasp/
/ About us
Founded by Russian entrepreneur Dmitry Itskov in February 2011 with the participation of leading Russian specialists in the field of neural interfaces, robotics, artificial organs and systems.
The main goals of the 2045 Initiative: the creation and realization of a new strategy for the development of humanity which meets global civilization challenges; the creation of optimale conditions promoting the spiritual enlightenment of humanity; and the realization of a new futuristic reality based on 5 principles: high spirituality, high culture, high ethics, high science and high technologies.
The main science mega-project of the 2045 Initiative aims to create technologies enabling the transfer of a individual’s personality to a more advanced non-biological carrier, and extending life, including to the point of immortality. We devote particular attention to enabling the fullest possible dialogue between the world’s major spiritual traditions, science and society.
A large-scale transformation of humanity, comparable to some of the major spiritual and sci-tech revolutions in history, will require a new strategy. We believe this to be necessary to overcome existing crises, which threaten our planetary habitat and the continued existence of humanity as a species. With the 2045 Initiative, we hope to realize a new strategy for humanity's development, and in so doing, create a more productive, fulfilling, and satisfying future.
The "2045" team is working towards creating an international research center where leading scientists will be engaged in research and development in the fields of anthropomorphic robotics, living systems modeling and brain and consciousness modeling with the goal of transferring one’s individual consciousness to an artificial carrier and achieving cybernetic immortality.
An annual congress "The Global Future 2045" is organized by the Initiative to give platform for discussing mankind's evolutionary strategy based on technologies of cybernetic immortality as well as the possible impact of such technologies on global society, politics and economies of the future.
Future prospects of "2045" Initiative for society
2015-2020
The emergence and widespread use of affordable android "avatars" controlled by a "brain-computer" interface. Coupled with related technologies “avatars’ will give people a number of new features: ability to work in dangerous environments, perform rescue operations, travel in extreme situations etc.
Avatar components will be used in medicine for the rehabilitation of fully or partially disabled patients giving them prosthetic limbs or recover lost senses.
2020-2025
Creation of an autonomous life-support system for the human brain linked to a robot, ‘avatar’, will save people whose body is completely worn out or irreversibly damaged. Any patient with an intact brain will be able to return to a fully functioning bodily life. Such technologies will greatly enlarge the possibility of hybrid bio-electronic devices, thus creating a new IT revolution and will make all kinds of superimpositions of electronic and biological systems possible.
2030-2035
Creation of a computer model of the brain and human consciousness with the subsequent development of means to transfer individual consciousness onto an artificial carrier. This development will profoundly change the world, it will not only give everyone the possibility of cybernetic immortality but will also create a friendly artificial intelligence, expand human capabilities and provide opportunities for ordinary people to restore or modify their own brain multiple times. The final result at this stage can be a real revolution in the understanding of human nature that will completely change the human and technical prospects for humanity.
2045
This is the time when substance-independent minds will receive new bodies with capacities far exceeding those of ordinary humans. A new era for humanity will arrive! Changes will occur in all spheres of human activity – energy generation, transportation, politics, medicine, psychology, sciences, and so on.
Today it is hard to imagine a future when bodies consisting of nanorobots will become affordable and capable of taking any form. It is also hard to imagine body holograms featuring controlled matter. One thing is clear however: humanity, for the first time in its history, will make a fully managed evolutionary transition and eventually become a new species. Moreover, prerequisites for a large-scale expansion into outer space will be created as well.
Key elements of the project in the future
• International social movement
• social network immortal.me
• charitable foundation "Global Future 2045" (Foundation 2045)
• scientific research centre "Immortality"
• business incubator
• University of "Immortality"
• annual award for contribution to the realization of the project of "Immortality”.
/ Follow us at:
/ Contact:
For additional information, partnerships and sponsorship, contact:
info@2045.com
Media requesting information and interviews, contact PR@2045.com
LinkedIn
LiveJournal
Google
Twitter
Facebook
Я.ру
ВКонтакте
Mail.ru