The science behind the comic
Fake news detection in social networks using geometric deep learning (GoodNews)
Social media has become one of our main sources of news. However, as the past elections in the US and
the UK have shown, it has also become a platform to share intentionally false information. But how to
detect, what content is true or false? Michael Bronstein approaches the topic from a new angle. Instead
of analysing the content, his team examines the news spreading patterns in social networks.
With the ERC Consolidator Grant, Prof. Bronstein developed a novel class of geometric deep learning algorithms. His new research project GoodNews, funded with the Proof of Concept grant, is based on these algorithms and aims to convert the geometric deep learning technology into a commercial application that could be used for detecting fake news. Bronstein's team has established a company Fabula AI and will use the grant to develop a prototype system with real data from social media and solidify the intellectual property portfolio.
Read the abstract here.
Researcher: Michael Bronstein
Host Institution: Universita Della Svizzera Italiana, Switzerland