Tech
Robots Learn to Cook from YouTube Videos (VIDEO)
Catherine Griffin
First Posted: Jan 13, 2015 10:46 AM EST
You can learn all sorts of things off of YouTube: how to play guitar, how to sew a shirt, how to bake a cake. Now, it seems as if robots can do the same.
Scientists have long wanted to create robots that can teach themselves. Now, they've succeeded. The researchers have developed robotic systems that can learn intricate grasping and manipulation movements required for cooking by watching online cooking videos. The key is that these robots can "think" for themselves and determine the best combination of observed motions that will allow them to efficiently accomplish a given task.
"We chose cooking videos become everyone has done it and understands it," said Yiannis Aloimonos, one of the researchers, in a news release. "But cooking is complex in terms of manipulation, the steps involved and the tools you use. If you want to cut a cucumber, for example, you need to grab the knife, move it into place, make the cut and observe the results to make sure you did them properly."
One of the key challenges was devising a way for the robots to parse individual steps appropriately. The robots had to be able to recognize each distinct step, assign it to a "rule," and then string together these behaviors.
"We are trying to create a technology so that robots eventually can interact with humans," said Cornelia Fermuller, one of the researchers. "So they need to understand what humans are doing. For that, we need tools so that the robots can pick up a human's actions and track them in real time. We are interested in understanding all of these components. How is an action performed by humans? How is it perceived by humans? What are the cognitive processes behind it?"
The robots rely on specialized software architecture known as deep-learning neural networks. Interestingly, though, these robots can be flexible and learn "on the job." This could be huge in terms of the next phase of automation. In fact, the researchers predict that these robots could herald the "next industrial revolution."
The findings can be found online here.
Want to see for yourself? Check out the video below, courtesy of YouTube.
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
TagsRobot ©2024 ScienceWorldReport.com All rights reserved. Do not reproduce without permission. The window to the world of science news.
More on SCIENCEwr
First Posted: Jan 13, 2015 10:46 AM EST
You can learn all sorts of things off of YouTube: how to play guitar, how to sew a shirt, how to bake a cake. Now, it seems as if robots can do the same.
Scientists have long wanted to create robots that can teach themselves. Now, they've succeeded. The researchers have developed robotic systems that can learn intricate grasping and manipulation movements required for cooking by watching online cooking videos. The key is that these robots can "think" for themselves and determine the best combination of observed motions that will allow them to efficiently accomplish a given task.
"We chose cooking videos become everyone has done it and understands it," said Yiannis Aloimonos, one of the researchers, in a news release. "But cooking is complex in terms of manipulation, the steps involved and the tools you use. If you want to cut a cucumber, for example, you need to grab the knife, move it into place, make the cut and observe the results to make sure you did them properly."
One of the key challenges was devising a way for the robots to parse individual steps appropriately. The robots had to be able to recognize each distinct step, assign it to a "rule," and then string together these behaviors.
"We are trying to create a technology so that robots eventually can interact with humans," said Cornelia Fermuller, one of the researchers. "So they need to understand what humans are doing. For that, we need tools so that the robots can pick up a human's actions and track them in real time. We are interested in understanding all of these components. How is an action performed by humans? How is it perceived by humans? What are the cognitive processes behind it?"
The robots rely on specialized software architecture known as deep-learning neural networks. Interestingly, though, these robots can be flexible and learn "on the job." This could be huge in terms of the next phase of automation. In fact, the researchers predict that these robots could herald the "next industrial revolution."
The findings can be found online here.
Want to see for yourself? Check out the video below, courtesy of YouTube.
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