Wired reports on cat recognition. Two wins here: cats are the best; and evolution beats ID.
Google’s mysterious X lab built a neural network of 16,000 computer processors with one billion connections and let it browse YouTube, it did what many web users might do — it began to look for cats.
The “brain” simulation was exposed to 10 million randomly selected YouTube video thumbnails over the course of three days and, after being presented with a list of 20,000 different items, it began to recognize pictures of cats using a “deep learning” algorithm.
Take that ID suckers! If a few thousand processors can do this, then a few billion years for evolution to result in systems that recognise and operate in their environment (i.e. life) is a snip. The BBC reports:
The work of the team stands at odds with many image-recognition techniques, which depend on telling a computer to look for specific features of a target object before any are presented to it.
Damn! I’ve been using the Godly method of divinely commanding my software to work, when all the time I should have used evolutionary techniques. Note to self on next sales pitch:
Here’s a computer. Here’s some random code I threw together. Give it a try, and let me know how it goes. It should figure itself out eventually. Disclaimer: being evolutionary, when it does eventually work there’s no telling what it will work at.
On second thoughts, that does sound a little like how I work.