The concept of robots that learn may seem relatively simply. We see robots and machines that are able to perform complex tasks flawlessly and many believe that this type of technology means that artificially intelligent robots are right around the corner (or here already). The biggest challenge has been in programming robots with the ability to adjust to changing environments and circumstances, applying what they know and extrapolating what they do not know to make a proper choice in new situations.
Sounds easy, but it’s not.
Not yet, at least.
Researchers in Japan have developed a “Self-Replicating Neural Network” that allows robots to think much in the same way that humans do – assess a problem and solve it.
“So far, robots, including industrial robots, have been able to do specific tasks quickly and accurately,” said Osamu Hasegawa, Associate Professor of Imaging Science and Engineering Lab at the Tokyo Institute of Technology. “But if their environment changes slightly, robots like that can’t respond.”
This new technology apparently changes things.
“This robot remembers only basic knowledge, and and it can apply that knowledge to its immediate situation.”
The potential is not necessarily in the technology individually but as a network. The theory is that eventually all of these robots being used for human tasks will be connected. If one robot learns something new, all robots learn something new.
Have you ever seen iRobot? I hope this doesn’t head down that road.