Twitter acquires ML startup Fabula AI to stem the tide of disinformation

The acquisition will help Twitter build out its machine learning team

It didn't take long after election night 2016 for people to start questioning the role that disinformation on social media had on the outcome, nor for those companies to admit that it was a problem. Yet, now it's two and a half years later, and these companies still haven't figured out how to effectively combat the spread of fake news.

While social media companies have taken steps such as making it easier for users to report fake stories, as well giving users more warnings about which information might not be true, the sheer amount of disinformation coming from foreign actors is so overwhelming that the only truly effective way to combat it is with technology that can detect it early, and then remove it before it can take hold.

That's why Twitter announced on Monday that it acquired Fabula AI, a London-based company that incorporates what it calls "Geometric Deep Learning," basically using AI technology to learn from social networks. No financial terms of the deal were disclosed.

This acquisition will help Twitter continue to build out its machine learning team; five of the members from the Fabula team have joined the Cortex team at Twitter, which is "the central ML org at Twitter with the goal to advance machine learning inside & outside Twitter," CTO Parag Agrawal wrote. Fabula founder Michael Bronstein will be leading graph deep learning research at Twitter, while also maintaining his position as the Chair in Machine Learning & Pattern Recognition at Imperial College.

Meanwhile, Twitter will also be using Fabula's technology to help combat the spread of disinformation on its platform. 

"Graph deep learning is a novel method for applying powerful ML techniques to network-structured data," Agrawal wrote. "The result is the ability to analyze very large and complex datasets describing relations and interactions, and to extract signals in ways that traditional ML techniques are not capable of doing."

Specifically, Twitter plans to use this technology to "help people feel safe on Twitter and help them see relevant information."

"Specifically, by studying and understanding the Twitter graph, comprised of the millions of Tweets, Retweets and Likes shared on Twitter every day, we will be able to improve the health of the conversation, as well as products including the timeline, recommendations, the explore tab and the onboarding experience," Agrawal said. 

Combating fake news is the first only the first part of Twitter's plan for Fabula'a technology; eventually, the company also plans to use it to help fight spam as well as abuse, another problem that Twitter has been fighting for years. 

"We’ll integrate Fabula’s technology where it makes sense to strengthen our systems and operations in the coming months.  It will likely take us some time to be able to integrate their graph deep learning algorithms into our ML platform," a Twitter spokesperson told VatorNews. 

This is far from Twitter's first machine learning acquisition: the company bought deep learning startup Madbits, which uses machine learning for image search, in July 2014. In June of 2015 it purchased Whetlab, another London-based machine learning startup focused on making machine learning easier to implement. In 2016 the company bought Magic Pony, a company that uses machine learning to enhance visual processing. 

The acquisition of Fabula, however, is more important, as it gets to a deeper question lurking within Twitter, and social media in general; one that had Facebook CEO Mark Zuckerberg hauled in front of Congress, and the American public. It's the question of whether or not these platforms can be trusted, and if they are harmful to society at large, especially if they are having an impact on democracy and elections. 

Disinformation is by no means only a problem facing Twitter, but all social media platforms; Facebook made its own acquisition akin to Fabula last year, buying Bloomsbury AI to combat fake news on its platform as well. Both companies are taking the issue seriously, as well they should; not until these questions have been sufficiently answered will social media regain the public trust, something that it has lost in the last few years. 

Founded in 2018, Fabula had raised two rounds of undisclosed seed funding from investors that included Elio Leoni Sceti, according to Crunchbase.

(Image source: fabula.ai)