Filed under: Computers
Maru goes in the box. Maru goes out of the box. Maru goes back into the box, again. Not exactly genius (adorable, but not genius), though apparently scientists say there is something about the feline brain that may help advance the “thinking computer.” A machine that can replicate basic human cognition — abstract concepts, emotional data — is still pretty distant, but IBM researchers in Portland, Oregon have scaled it down a bit. Using a kitty cortex as a model, these neuro-computer scientists have claimed to have replicated a cat’s thinking pattern.
This doesn’t mean that Maru, or Mittens, or any other kitten, is dumb: the computer operates using 147,456 processors (while the average computer uses just a handful) and 144 terabytes of memory, and it is still 100 times slower than the way a cat thinks. The senior author of the findings, Dharmendra Modha, suggests that by creating a program based on the way a brain works, the emphasis is moved from “structured” data (say, adding numbers) and is placed on weighing different factors (like identifying an image, even if its blurry). So even though it is slow, the Portland supercomputer is apparently a breakthrough in cognitive computing.
Continue reading Kitty Brain Simulated By IBM, but Not Just for Adorable-ness
Kitty Brain Simulated By IBM, but Not Just for Adorable-ness originally appeared on Switched on Wed, 18 Nov 2009 12:50:00 EST. Please see our terms for use of feeds.
Read | Permalink | Email this | Comments
