On Wed, Sep 12, 2012 at 9:03 AM, Mike Tintner <[email protected]> wrote: > What we have now - the pure fantasy that machines can leap to human-level > intelligence (incl. maths and aesthetic appreciation) - in a couple of > bounds - that there is no natural order of evolution of capabilities - is, > to use Sergio's word, absolutely "preposterous". A Don Quixote joke of pure > dreamers.
Some people believe that. The way Google will probably solve the "funny video" problem is to take millions of videos that have already been rated by humans, and divide them into a training set and a test set. Then they will experiment with various machine learning algorithms on the training set (running on a million cores) and measure their accuracy on the test set. They won't know in advance which algorithms will work, if any, because the algorithm for humor is poorly understood. They will probably have some guesses, however. It requires understanding of high level features (e.g. cat faces. That part is done). There also has to be an element of surprise, which is why a joke isn't funny the second time you hear it. The result could very well be beyond human comprehension. It works, but we don't understand why it works. -- Matt Mahoney, [email protected] ------------------------------------------- AGI Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-c97d2393 Modify Your Subscription: https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-2484a968 Powered by Listbox: http://www.listbox.com
