Terms of the deal weren’t disclosed; however the acquisition will underpin a research group at Twitter led by Sandeep Pandey that will work toward finding new ways to leverage machine learning across natural language processing (NLP), recommendations systems, reinforcement learning, and graph deep learning. The group will also address ml ethics.
Founded in 2018, Fabula has developed a patented AI system it calls “geometric deep learning” — effectively algorithms that learn from massive and complex information sets gleaned from social networks.
“Fake news” has become an umbrella buzzword to describe the deliberate spread of misinformation; however Fabula AI is basically about helping identify the authenticity of any data that circulates on social media — despite intent. Studies have shown that false news spreads quicker than real news online, a pattern that can be used to help spot misinformation. This is what Fabula focuses on: detecting differences in how content is spreading on social media and allocating an authenticity score.
“As this technology detects the spread pattern, it is language and locale independent; actually, it can be used even once the content is encrypted,” the corporate says on its homepage. “We additionally believe that such an approach, given it’s based on the propagation pattern through large social networks, is much more resilient to adversarial attacks.”
It seems Fabula is more of an acqui-hire than anything else, with the team joining Twitter’s Cortex unit, which is the social network giant’s in-house group of researchers and engineers working on machine learning tech for Twitter. Fabula cofounder and chief scientist Michael Bronstein will now lead graph deep learning at Twitter whereas also retaining a separate position as chair of machine learning and pattern recognition at Imperial College.
As with most of the main social media platforms, Twitter has faced its share of criticism for the way it’s used to spread misinformation. This latest move is intended to “improve the health of the conversation” on Twitter, according to CTO Parag Agrawal, and will expand over time to assist stop other kinds of spam and platform abuse.
“By studying and understanding the Twitter graph, comprised of the millions of tweets, retweets, and likes shared on Twitter daily, 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.
Twitter has made a number of machine learning acquisitions through the years, including visual intelligence startup Madbits back in 2014 and Whetlab and Magic Pony in the following years.
But Fabula is a significantly notable acquisition, as the underlying technology is squarely centered on fighting the spread of misinformation online. This could prove to be something of a trend — last year, Facebook snapped up Bloomsbury AI, a startup building NLP smarts that could even be used to help combat fake news.
With the 2020 U.S. presidential election on the horizon, social media companies will be under intense scrutiny for their handling of fake news — which is partly why Twitter is trying to invest in automation to get rid of the bad eggs.
“Machine learning plays a key role in powering Twitter and our purpose of serving the public conversation,” Agrawal said.