Technology from the London-based start-up is to be used to track down misinformation and spam and ‘improve the health of the conversation’
Twitter said it has acquired Fabula AI, a startup based in London whose machine learning technology can be used to spot misinformation as it spreads online.
Staff acquired with Fabula is to form the backbone of a new research group at Twitter, led by Twitter senior director of engineering Sandeep Pandey, that is to look for new ways of applying machine learning across Twitter’s AI systems.
Fabula co-founder and chief scientist Michael Bronstein is to lead research into graph deep learning at Twitter, as well as retaining his position as chair of machine learning and pattern recognition at Imperial College London.
The start-up, founded only last year, develops an AI system it calls “geometric deep learning” that is tuned to finding patterns in large data sets from social networks.
The company focuses on tracking the way information spreads on social media and allocating it an authenticity score based on a range of characteristics.
The technology is language- and locale-independent and can function even when content is encrypted, the company says.
Twitter said the acquisition is intended to “improve the health of the conversation” on the network and is planned to expand to include detecting spam and other abusive practices.
“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,” said Twitter chief technology officer Parag Agrawal in a statement.
The acquisition is comparable to Facebook’s purchase of Bloomsbury AI last year, whose AI technology can also be used to help combat misinformation.