Facebook is using an artificial intelligence method called deep learning to analyze data collected from its 1 billion users in a new way. The goal of the project — which is run by an eight person group known as the "AI team" — is to understand the meaning and importance behind the things users share on Facebook, according to a report from MIT Technology Review. If the team can do that, the social network might be able to predict our actions online, show us content that is more relevant to our interests, and better target advertisements, the report says.
Facebook isn't the first company to bring deep learning into a consumer context — Google and noted futurist Ray Kurzweil are applying the technique to web search and natural language processing. Last year, Google also used deep learning to identify cats in YouTube videos without its software ever being told that cats exist. IBM's Watson supercomputer used deep learning to defeat humans on the game show Jeopardy and now it's tackling health care issues. Deep learning uses a
"The data set is increasing in size." multi-layered approach to data, parsing information to build up a body of knowledge that can then be used to figure out concepts, or understand what objects sound and look like. Artificial intelligence researchers argue that deep learning will one day be able to understand information the way a human brain does.
Currently, an average users could potentially see upwards of 1,500 posts in the News Feed, but Facebook filters it down to between 30 and 60. The posts that make it to News Feed are what the company considers the items a user will be most interested in. But, the company tells Technology Review that it can do far better. "The data set is increasing in size, people are getting more friends, and with the advent of mobile, people are online more frequently," Mike Schroepfer, Facebook's chief technology officer, says in the report. "It's not that I look at my news feed once at the end of the day; I constantly pull out my phone while I'm waiting for my friend, or I'm at the coffee shop. We have five minutes to really delight you." Shroepfer also says that deep learning could eventually help users better organize photos and even decide what to share on Facebook.