As more and more content is created on Facebook, the platform needs to become smarter in order to filter out spam or recommend the right content. This is where Facebook’s reliance on AI and machine learning comes into play of course. In order to know the relevance of any text, it needs to able to understand it. So, Facebook built a “text understanding engine” called DeepText that can read and understand everything you post.
DeepText is a “text understanding engine” meaning that it can read and understand text almost as well as humans can. It is in fact extremely accurate in doing so by using a system of deep-learning and and neural networks to understand text in over 20 different languages. Every second, it can read and understand thousands of posts.
As is the case with any machine that reads and analyses text, there is alway the issue of context – machines find it very hard to understand context in human language due to the use of ambiguity and of course slang. This is made worse by both ambiguous words and slang WITHIN any language and not just between multiple languages.
If you like our stories, there is an easy way to stay updated:
DeepText somehow surmounts this problem by keeping the semantic links between words. In essence, it remembers the different instances of specific words being used in order to keep their relevance. Again, this changes across different languages as well. DeepText is not only planned to be used to offer relevant information to users. In the near future, it will be used in Messenger, which is bound to become a much more intelligent system with which users can interact in different ways.
If you are more interested in this, have a look here.
In any case, the important thing here is that Facebook is listening… or actually… reading… everything you write.
More from Facebook
Facebook announced new AI-powered detection technology that will help find and remove non-consensual intimate images (aka "revenge porn") on its …
Facebook announced that it's making several updates to its ad metrics - one of these is the replacement of relevance …