Facebook Using DeepText Artificial Intelligence To Make Its Computers Understand Everything You Post
The social network giant, Facebook on Wednesday announced a new, artificially intelligent language processing engine, called DeepText, a system developed by Facebook’s AI (artificial intelligence) labs that scans 10,000 posts every second in 20 languages. It applies a computer-science technique called deep learning to the old problem of teaching computers to understand human language.
“Understanding the various ways text is used on Facebook can help us improve people’s experiences with our products, whether we’re surfacing more of the content that people want to see or filtering out undesirable content like spam,” the company said in a blog post.
“Every minute, Facebook users make 4,000 new posts,” with 180 million comments made on public Facebook posts per day, says Hussein Mehanna, an engineering director with Facebook’s Applied Machine Learning team. “This is a step toward building machines that can communicate with humans in a smart way,” he added.
As Facebook puts it, DeepText is “deep learning-based text understanding engine.” DeepText has already been tested in Facebook Messenger, wherein the AI learned how to know when someone might need a taxi. Facebook writes, “DeepText is used for intent detection and entity extraction to help realize that a person is not looking for a taxi when he or she says something like, ‘I just came out of the taxi,’ as opposed to, ‘I need a ride.’” This is the major difference, says Facebook, from DeepText and other related AI; DeepText is a more thorough reader.
Also, if it detects that you are trying to sell something in a status update, AI can automatically format post with the price and item details. And if Facebook AI can determine what kinds of comments on celebrities’ posts are interesting, it can show the ones that you will actually want to read. Facebook is also building new deep learning architectures that can understand text and visual content together.
Facebook explains that “to get closer to how humans understand text, we need to teach the computer to understand things like slang and word-sense disambiguation. As an example, if someone says, “I like blackberry,” does that mean the fruit or the device?”
DeepText could help Facebook weed out harassment and abuse. If it can automatically clamp down hate speech from status updates that just has controversial terms, it could prevent a human from ever suffering the harassment and having to report it manually. Facebook already says that its AI systems report more offensive photos than humans on its social network.
However, such a tool might also seem like an invasion of privacy, having your private messages scanned by Facebook. The company has had lawsuits filed against it in the past on more than one occasion for scanning the contents of private messages.
With DeepText and its other image recognition AIs, Facebook could make huge enhancements in how it directs content from authors to viewers. Well, whether it really happens. Only time will tell.