Researchers have found a way to allow robots to learn how to cook by watching videos on YouTube.
The researchers from NICTA, Australia's Information and Communications Technology Research Centre of Excellence, and the University of Maryland developed a way for teaching robots how to properly use tools by having them watch videos on YouTube.
The work was made possible using a kind of artificial intelligence called deep learning.
Deep learning uses training systems that are called artificial neural networks for acquiring massive amounts of information gathered from different kinds of sources, including pictures and sounds, 'Tech Times' reported.
The information is presented to the artificial neural networks, which then send inferences in response.
For the study, the researchers used convolutional neural networks for the identification of activities such as a hand's grasp on an object, and for recognising specific objects.
The system is then capable of predicting the action that involves both the hand's grasping action and the object.
Researchers collected data from a total of 88 videos in YouTube of different people cooking. The researchers then generated commands that robots could execute, resulting in the robots learning how to cook.
"We believe this preliminary integrated system raises hope towards a fully intelligent robot for manipulation tasks that can automatically enrich its own knowledge resource by 'watching' recordings from the World Wide Web," researchers said.
