Deeptext – Facebook’s text understanding engine

Facebook recently announced a new way in which its Facebook platform can now understand posts and messages with “near-human accuracy.”

Text is the most common form of communication on Facebook, whether it’s post updates or using Facebook Messenger with your friends or family. Facebook wants to understand this better in order to continue to fulfil its objective of serving you the most relevant content or filtering out content that you do not wish to see, including spam.

With this in mind, Facebook’s DeepText technology uses machine learning to try and understand what it is you would like to see, or the help you may need. DeepText has the potential to further improve Facebook experiences by understanding posts better to extract intent, sentiment, and entities (e.g. people, places, events), using mixed content signals like text and images, and automating the removal of objectionable content like spam. Many celebrities and public figures use Facebook to start conversations with the public. These conversations often draw hundreds or even thousands of comments. Finding the most relevant comments in multiple languages while maintaining comment quality is currently a challenge. One additional challenge that DeepText may be able to address is surfacing the most relevant or high-quality comments. The technology analyses several thousand posts per second, spanning more than 20 languages.


As a consumer, this artificial intelligence technology means that it should improve your Facebook experience by serving you relevant content, beyond this however, Facebook also plans to make this technology helpful

DeepText is already being tested on some Facebook experiences such as Messenger, where DeepText is being used by the AML Conversation Understanding team to get a better understanding of when someone might want to go somewhere. It’s used for intent detection. For example, if you were to type in “I need a taxi” in Facebook Messenger, DeepText could interpret what you need and present you with the relevant options to order a taxi.

Beyond this, DeepText could also help people find the right Facebook tools in order to help them. For example, if someone did a Facebook post update saying ‘Bike for sale, $200’, Deeptext could detect what the post was about, extract the meaningful information such as the object being sold and the price and offer the user the option to place the object for sale on Facebook’s Marketplace.


Whilst Facebook has not specifically said they would open this technology up to advertisers (yet!), the opportunity for you as a potential advertiser is huge. Imagine a post where someone says ‘I’m hungry’ or ‘Anyone for pizza’ and you as a Pizza restaurant or Pizza delivery shop, within a 3-mile radius of the person posting, are able to present your advert for Pizza together, perhaps, with a voucher or free delivery.

Of course, the privacy implications of this technology may make you nervous, after all what Facebook is now highlighting is that it scans the content of all of your messages and this has potential privacy issues, but at the same time DeepText could help Facebook reduce harassment and abuse and that has to be a good thing.

With DeepText and its other image recognition technology, Facebook could make huge strides in how it serves content to it’s users and providing any advertising is relevant and helpful, the opportunities could be endless, so watch this space!