• July 2012
    M T W T F S S
    « Jun   Aug »
  • Advertisements

How Many Computers to Identify a Cat? 16,000

New York Times (06/25/12) John Markoff

 Google researchers have created one of the largest neural networks for machine learning, consisting of 16,000 computer processors and more than one billion connections.  The researchers fed the system 10 million digital images from YouTube for processing, and it performed much better than previous efforts by doubling its accuracy in recognizing objects from a list of 20,000 distinct items.  The researchers also note the network taught itself to recognize cats.  The software-based neural network closely mirrors theories developed by biologists that suggest individual neurons are trained inside the brain to detect significant objects.  “The idea is that instead of having teams of researchers trying to find out how to find edges, you instead throw a ton of data at the algorithm and you let the data speak and have the software automatically learn from the data,” says Google and Stanford University researcher Andrew Y. Ng.  The system created a dream-like digital image of a cat by utilizing a hierarchy of memory locations to successively identify general features after being exposed to millions of images.  The researchers say the neural network provides new evidence that current machine-learning algorithms improve as the systems are given access to larger pools of data.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: