Google and Human Brain

The Internet has created the most efficient generations of multitasking so far. Many people open at least four tabs at a time (Google, Facebook, Youtube, Pandora, Wikipedia, Gmail, etc.). People always jump from one page to another and click the link to open a new tab and browser. The way to master knowledge has changed from deep reading to rapid skimming. Information is displayed every time a Web page is opened, and ads and other information are displayed on most pages.

Recently, scientists of artificial intelligence are beginning to seek inspiration from the brain. Demis Hassibis, co-founder of Google Deep Mind Group, says: "The only evidence we have is the human brain, so learn how it accomplishes these functions." I agree. I have been studying the brain for over 30 years. In 2004, I wrote a book called "Intelligence", which proposed a way artificial intelligence needs brain theory. In 2005, I co-founded Numenta. It is a company specializing in reverse engineering of the new cortex, the most important part of the human brain, and the most important intelligence part. By understanding how cells in the brain cooperate to create perception and behavior, I understand what intelligence is. There is a similarity between the workings of the brain and the work of the current AI technology, which indicates that AI is progressing in the right direction.

You are a member of the Google Brain team and want to know the content of the YouTube video. This is a true story of the investigation by the YouTube Cat Detector, evoking general enthusiasm for artificial intelligence. In this article, the Google Brain team and the researchers at Stanford University Quoc Le and Andrew Ng describe algorithms to categorize YouTube videos into categories containing cats and categories. They are not planning to look for cats, but this algorithm automatically combines videos containing cats (and thousands of other objects in 22,000 object categories defined in ImageNet) without explicit training data.