A Different Path to AI
Google has extended the same “statistical machine learning” it so successfully uses for spelling — the did you mean… functionality — to language translation. With that comes a different path to AI:
Creating a translation machine has long been seen as one of the toughest challenges in artificial intelligence. For decades, computer scientists tried using a rules-based approach — teaching the computer the linguistic rules of two languages and giving it the necessary dictionaries.
But in the mid-1990s, researchers began favoring a so-called statistical approach. They found that if they fed the computer thousands or millions of passages and their human-generated translations, it could learn to make accurate guesses about how to translate new texts.
Extensions of the functionality include a free directory assistance service, 800-GOOG-411, a voice search system and Google Goggles, a search based on matching photos taken by a cellphone. Imagined future functionality includes translating a cellphone photo of, for example, a German menu into English.
RELATED: Yesterday, Google launched its Public Data Explorer. Specifically designed for avid data crunchers, it’s pretty cool:
One Response to “A Different Path to AI”
I find translating to be an annoying task at best. I've been using Google's translator to translate an entire piece of text and then go back and make minor corrections. I've seen a vast improvement in the translators ability over the years….it is quite impressive.