Robots guarding nuclear site, learning like humans

Future-deniers, think again - if there was ever a headline that people in the 1950s would have expected to read in 2010, this is it. It seems that the U.S. military has developed and deployed robots - known as Mobile Detection Assessment Response Systems, or MDARS - to patrol a nuclear dumping ground in Nevada. From Wired:

The camera-equipped MDARS can scoot around pre-determined paths on its own, alerting flesh-and-blood guards when it encounters an intruder or a broken lock. In development by the Navy and General Dynamics since the early 1990s, the diesel-fueled sentry bot can operate for up to 16 hours, and reach a top speed of 20 mph. The U.S. military has experimented with using the MDARS machines to patrol some of its Hawthorne Army Depot in Nevada. The bots have even been tested with automatic weapons -- though I doubt that's the plan at the nuke site.

Check out this official video footage. Not quite as spine-chilling as Terminators, but getting there:



In the meantime, on the other side of the country, researchers at Carnegie Mellon are attempting to duplicate human learning processes in machines:

Since the start of the year, a team of researchers at Carnegie Mellon University -- supported by grants from the Defense Advanced Research Projects Agency and Google, and tapping into a research supercomputing cluster provided by Yahoo -- has been fine-tuning a computer system that is trying to master semantics by learning more like a human. Its beating hardware heart is a sleek, silver-gray computer -- calculating 24 hours a day, seven days a week -- that resides in a basement computer center at the university, in Pittsburgh. The computer was primed by the researchers with some basic knowledge in various categories and set loose on the Web with a mission to teach itself.

(Read more here.) Bring these two together, and what do you get? Well... we'll find out. (Along these lines - check out this NPR segment, "Can Unmanned Robots Follow the Laws of War?")