Skip to main content
  1. Home
  2. Social Media
  3. News

Facebook's text understanding engine is reading thousands of posts every second

Brad Jones
By

facebook messenger lite apple ios macbook iphone 4 5 6 app
Panithan Fakseemuang/123rf
Facebook has unveiled a new piece of technology that’s either impressive or invasive depending on how you look at it. DeepText is a “text understanding engine” that can apparently read and understand posts with near-human accuracy.

A blog post published on the Facebook Code subsite stresses the enormous importance of text-based communication to the social media giant. The company’s ability to recommend relevant content and filter out spam will reportedly be vastly improved thanks to the abilities of DeepText.

Recommended Videos

Using a complex system of neural networks and deep-learning techniques, DeepText is able to understand the textual content of posts written in more than 20 languages at a rate of several thousand posts per second.

Related

Poring over Facebook posts might not seem like the most challenging task, but difficulties arise for nonhuman readers when context comes into play. Previous attempts to create a system like DeepText were stymied by confusing elements like ambiguous terminology and slang.

Traditional methods of “teaching” a machine to read centered around the process of assigning each word to a numerical ID that’s easily recognized by a computer. However, this strategy requires words to be written exactly as in the teaching materials if the text is to be understood.

Conversely, DeepText uses a system that preserves semantic links between different terms. This means that words like ‘brother’ and ‘bro’ are connected, as well as different representations of the same term in different languages.

Facebook is an enormous service, and the company faces a big challenge in parsing all the text that makes its way onto the service and coming up with something that’s relevant. The DeepText system could be a huge help toward separating the wheat from the chaff.

In fact, DeepText is already being implemented in various Facebook services as a test of its capabilities. For instance, the system has been integrated into Messenger in such a way that it can determine when a user is looking to call a cab and offer up a means of doing so — but it will ignore any usage of words like “ride” that don’t refer to a taxi journey.

Development of DeepText is set to continue in collaboration with the Facebook AI Research group.

Editors' Recommendations

Topics
Brad Jones
Brad Jones
Former Digital Trends Contributor
Brad is an English-born writer currently splitting his time between Edinburgh and Pennsylvania. You can find him on Twitter…
Almost every state joins in on antitrust investigation into Facebook
facebook ftc fine not enough header

The attorneys general of nearly every state has joined New York in an antitrust investigation into Facebook. In total, 47 attorneys general are now looking into whether the social media giant “stifled competition” in a way that may have impacted user data and ad prices. 

New York Attorney General Letitia James announced the additional states’ participation in a press release on Tuesday, October 22. James initially began the investigation on September 6 on antitrust grounds. Attorneys general from both parties have now joined James to question Facebook’s business practices. 

Read more
Facebook acquires a startup that has built a mind-reading wristband
ctrl labs facebook acquisition crtl mind control device 2785 1500x1000

Facebook has acquired CTRL-Labs, a New York-based startup that has developed tech for controlling computers using decoded neural activity and the wave of a hand. Because, you know, who wants to bother using their mouse and keyboard to “like” posts or comment on someone’s holiday photos?

According to CNBC, the deal was worth between $500 million and $1 billion. A Facebook spokesperson told the publication that it spent under $1 billion. CTRL-Labs was founded in 2015 by Thomas Reardon, the creator of Microsoft Internet Explorer. It describes itself as an “applied research neuroscience company.”

Read more
Facebook has suspended ‘tens of thousands’ of apps over private data use
Facebook CEO Mark Zuckerberg in Washington, D.C.

Facebook says it has suspended “tens of thousands” of apps suspected of misusing private user data amid an investigation that it began last year following the Cambridge Analytica scandal.
The social media giant promised in March 2018 to take a look at all of the apps on Facebook that had access to large amounts of user information prior to when it changed its policy in 2014. The resulting suspensions were announced in a blog post on Friday. The investigation began after right-wing political analytics firm Cambridge Analytica improperly mined data from 87 million Facebook users. 
"Our App Developer Investigation is by no means finished. But there is meaningful progress to report so far," Ime Archibong, Facebook's vice president of product partnerships, wrote in the blog post. "To date, this investigation has addressed millions of apps. Of those, tens of thousands have been suspended for a variety of reasons while we continue to investigate."
Facebook said that it initially selected apps for investigation based on their user numbers, as well as how much data they could access from user accounts. It has since broadened that investigation to include apps that it thinks have the potential to abuse its policies, a process that involves both a background investigation of the developer, as well as an analysis of the app’s activity on the platform.
The investigation, Archibong wrote, involved “hundreds of people: Attorneys, external investigators, data scientists, engineers, policy specialists, platform partners, and other teams across the company.” 
While Facebook says that it removed tens of thousands of apps, it notes that the suspended apps, while plentiful, were only created by roughly 400 developers. An app being suspended doesn’t mean that it was necessarily posing a threat to people, and many were still in their testing phase when they were suspended, so they were never readily available to Facebook’s wider user base.
Facebook says it is not unusual for developers to have multiple test apps that never get rolled out. In some cases, apps were suspended because the developer did not respond to a request for information from Facebook, not due to any actual policy violations. One app, myPersonality, apparently shared information with researchers and companies but refused to take part in Facebook's audit -- so it was banned
The post goes on to says that Facebook hasn’t confirmed any other instances of misuse to date other than those it has already revealed to the public and that it has taken legal action against a handful of apps for a number of different reasons.
Beyond the investigation, Facebook says it has “made widespread improvements to how we evaluate and set policies for all developers that build on our platforms.” It also removed a number of APIs and has increased the number of people that work on its investigations team in order to better handle them in the future.

Read more