Robots can do amazing things these days — just look at the recent winner of the DARPA Challenge. Tanya Lewis of Live Science wrote about how "spectators fell in love with them, cheering when the robots succeeded and feeling sorry for them when they failed."
Perhaps they might be less in love if they read "The Rise of the Robots: Technology and the Threat of a Jobless Future." In the book, Martin Ford mentions a study from the University of Oxford that studied more than 700 job types in America and determined that almost 50 percent of them "will ultimately be susceptible to full machine automation." He suggests that most people believe that "occupations that require physical manipulation of the environment will always be safe" — and they are wrong. It’s too bad that he couldn’t embed the video above into his book; it shows that robots can do very sophisticated physical manipulation indeed.
The Yaskawa Bushido Project is a demonstration of a robot, specifically a MOTOMAN MH24, that has been taught Samurai sword techniques copied from Isao Machii, holder of a number of Guinness records, including the "Fastest tennis ball (708 km/h) cut by sword." The robot masters recorded the sword master’s moves and programmed the robotic arm to make the same moves at the same speed. It’s both impressive (that arm moves FAST!) and a demonstration of the robot’s limitations. It takes a lot of careful analysis and programming to teach that arm what to do.
At UC Berkeley, it's a slightly different story. BRETT (Berkeley Robot for the Elimination of Tedious Tasks) doesn’t need pre-programming; it’s designed to learn from its mistakes. Unlike MOTOMAN with the sword, this robot adapts to its environment. In a news release, Berkeley faculty member Trevor Darrell explains how BRETT is different:
Most robotic applications are in controlled environments where objects are in predictable positions.The challenge of putting robots into real-life settings, like homes or offices, is that those environments are constantly changing. The robot must be able to perceive and adapt to its surroundings.
With more data, you can start learning more complex things. We still have a long way to go before our robots can learn to clean a house or sort laundry, but our initial results indicate that these kinds of deep learning techniques can have a transformative effect in terms of enabling robots to learn complex tasks entirely from scratch. In the next five to 10 years, we may see significant advances in robot learning capabilities through this line of work.
More in MNN and TreeHugger: