Facebook auto-translates 200 million lines of text every day
Facebook reaches more than 1.6 billion users in every corner of the Earth. Making sure everybody understands one another, at least at the linguistic level, has fallen to the company’s machine learning algorithms. According to Alan Packer, Facebook’s Director of Engineering for language technology, the digital babelfish is doing a pretty great job of it.
Speaking at MIT’s Emtech Digital conference in San Francisco on Monday, Packer also noted that Facebook’s translation service understands more than 40 languages and dialects. What’s more, the service understands both the words themselves and their larger social context. If, for example, the system sees that the post it’s translating is asking for hotel recommendations in Paris, it could promote that post to the user’s French friends, suggest a specific Parisian pal for insights or search other public posts for those recommendations.
While Packer remained mum on whether a universally-translated Facebook experience could garner users more international connections, he did tell TechCrunch, “The mission of the translation team is removing language as a barrier to making the world more open and connected.” Of course, with the recent advent of 24-hour Facebook Live kitten cams, who needs words?
(39)
