Tech
BRETT: The Robot that Learns to Complete Tasks Like a Human (VIDEO)
Catherine Griffin
First Posted: May 27, 2015 08:33 AM EDT
Imagine a robot that learns by trial and error--just like a human. UC Berkeley researchers have now developed algorithms that enable robots to learn motor tasks in this way, marking a milestone in artificial intelligence.
Seeing is believing, though, and that's why the researchers demonstrated the new algorithm through a robot. They had the machine complete various tasks, such as putting a clothes hanger on a rack, assembling a toy plane, screwing a cap on a water bottle and more. What made this extraordinary was that the robot didn't have any pre-programmed details about its surroundings.
"What we're reporting one here is a new approach to empowering a robot to learn," said Pieter Abbeel of UC Berkeley's Department of Engineering and Computer Sciences, in a news release. "The key is that when a robot is faced with something new, we won't have to reprogram it. The exact same software, which encodes how the robot can learn, was used to allow the robot to learn all the different tasks we gave it."
Currently, most robots are in controlled environments where objects are in predictable positions. In this case, though, the researchers were able to give the robot a way to learn about its environment and adapt.
"For all our versatility, humans are not born with a repertoire of behaviors that can be deployed like a Swiss army knife, and we do not need to be programmed," said Sergey Levine, one of the researchers. "Instead, we learn new skills over the course of our life from experience and from other humans. This learning process is so deeply rooted in our nervous system, that we cannot even communicate to another person precisely how the resulting skill should be executed. We can at best hope to offer pointers and guidance as they learn it on their own."
The new robot, nicknamed BRETT for Berkeley Robot for the Elimination of Tedious Tasks, can master a typical assignment in about 10 minutes when given the relevant coordinates for the beginning and end of the task. When not given the location for objects in the scene, the learning process takes about three hours.
Either way, the robot is a huge leap forward for artificial intelligence. As data processing improves, it's likely that this system will be able to learn more complex things, as well.
Want to see BRETT in action? Check out the video below, courtesy of YouTube.
Related Stories
Drone Flies Autonomously and Learns New Routes Using Navigation System
Control Swarms of Robots with a Simple Swipe of Your Finger on Your Tablet
For more great science stories and general news, please visit our sister site, Headlines and Global News (HNGN).
See Now:
NASA's Juno Spacecraft's Rendezvous With Jupiter's Mammoth Cyclone
©2024 ScienceWorldReport.com All rights reserved. Do not reproduce without permission. The window to the world of science news.
More on SCIENCEwr
First Posted: May 27, 2015 08:33 AM EDT
Imagine a robot that learns by trial and error--just like a human. UC Berkeley researchers have now developed algorithms that enable robots to learn motor tasks in this way, marking a milestone in artificial intelligence.
Seeing is believing, though, and that's why the researchers demonstrated the new algorithm through a robot. They had the machine complete various tasks, such as putting a clothes hanger on a rack, assembling a toy plane, screwing a cap on a water bottle and more. What made this extraordinary was that the robot didn't have any pre-programmed details about its surroundings.
"What we're reporting one here is a new approach to empowering a robot to learn," said Pieter Abbeel of UC Berkeley's Department of Engineering and Computer Sciences, in a news release. "The key is that when a robot is faced with something new, we won't have to reprogram it. The exact same software, which encodes how the robot can learn, was used to allow the robot to learn all the different tasks we gave it."
Currently, most robots are in controlled environments where objects are in predictable positions. In this case, though, the researchers were able to give the robot a way to learn about its environment and adapt.
"For all our versatility, humans are not born with a repertoire of behaviors that can be deployed like a Swiss army knife, and we do not need to be programmed," said Sergey Levine, one of the researchers. "Instead, we learn new skills over the course of our life from experience and from other humans. This learning process is so deeply rooted in our nervous system, that we cannot even communicate to another person precisely how the resulting skill should be executed. We can at best hope to offer pointers and guidance as they learn it on their own."
The new robot, nicknamed BRETT for Berkeley Robot for the Elimination of Tedious Tasks, can master a typical assignment in about 10 minutes when given the relevant coordinates for the beginning and end of the task. When not given the location for objects in the scene, the learning process takes about three hours.
Either way, the robot is a huge leap forward for artificial intelligence. As data processing improves, it's likely that this system will be able to learn more complex things, as well.
Want to see BRETT in action? Check out the video below, courtesy of YouTube.
Related Stories
Drone Flies Autonomously and Learns New Routes Using Navigation System
Control Swarms of Robots with a Simple Swipe of Your Finger on Your Tablet
For more great science stories and general news, please visit our sister site, Headlines and Global News (HNGN).
See Now: NASA's Juno Spacecraft's Rendezvous With Jupiter's Mammoth Cyclone