HARD TECHNOLOGY GROUP HTG NEWS TIPS

Every day, Facebook performs some 4.5 billion automatic translations — and as of yesterday, they’re all processed using neural networks. Previously, the social networking site used simpler phrase-based machine translation models, but it’s now switched to the more advanced method. “Creating seamless, highly accurate translation experiences for the 2 billion people who use Facebook is difficult,” explained the company in a blog post. “We need to account for context, slang, typos, abbreviations, and intent simultaneously.” The big difference between the old system and the new one is the attention span. While the phrase-based system translated sentences word by word, or by looking at short phrases, the neural networks consider whole sentences at a time. They do this using a particular sort of machine learning component known as an LSTM or long short-term memory network. The benefits are pretty clear. Compare these two examples from Facebook of a Turkish-to-English translation. Th...

Facebook's latest artificial intelligence system is known as "DeepText," and it has some pretty serious implications as far as privacy goes. The social network is interested in implementing the system that can help analyze thousands of posts per second across 20 different languages, all with near-human accuracy. What does that mean for you as a Facebook user? Soon, the company will be able to identify what you're saying in posts, messages and other content available via your profile and other pages to better offer suggestions and other helpful options. For example, if you happen to post a status about needing a ride or converse with a friend about when you might need to be picked up, Facebook might suggest you hit up Uber or Lyft services via Messenger transportation integration. The key is distinguishing between slang, jargon and other intricacies of human language that can sometimes make this sort of thing difficult for artificial intelligence to understand. ...

Earlier this month, Facebook announced that it had begun using a language AI called DeepText on its platform. Language is complex, and in order to better understand intent (an important part of flagging hate speech, for example), any computer program needs to figure out how humans use language. Now, Instagram, which is owned by Facebook, has announced that it's begun using DeepText to eliminate comments that violate Instagram's Community Guidelines. DeepText is currently in limited deployment on Facebook, but immediately after learning about the AI, Instagram's top people wanted to test it out on their own platform. They first focused on spam, rather than mean or spiteful comments, asking human workers to wade through a giant set of comments and flag spam by hand. They then fed most of this data into DeepText, which created algorithms based on what it found in the spam comments. The team then turned around and tested the algorithm on the portion of the human-filtered ...