The research, funded by DARPA's Mathematics of Sensing, Exploitation and Execution (MSEE) program, is in the early stages, but is showing promising signs. The team have currently developed a system that allowed their test robots to watch a series of cooking tutorial videos and perform the tasks shown with zero human input.
On top of this, the program enables the robots to store whatever knowledge they've learned and build on it to become better at what they do. This means they can learn a lot faster, at a fraction of the cost to roboticists; plus this can expand to robots teaching other robots.
“This system allows robots to continuously build on previous learning—such as types of objects and grasps associated with them—which could have a huge impact on teaching and training,” Ghanadan said.
“Instead of the long and expensive process of programming code to teach robots to do tasks, this research opens the potential for robots to learn much faster, at much lower cost and, to the extent they are authorized to do so, share that knowledge with other robots. This learning-based approach is a significant step towards developing technologies that could have benefits in areas such as military repair and logistics.”
While this sounds amazing, we hope this isn't a terrifying step towards Skynet! Feel free to take a look at University of Maryland's paper for more info.