Ricerc@Sapienza

In recent years, we have seen a drastic disruption in the way news, and information in general, are consumed.
In fact, while just twenty years ago the only way we had to get in contact with the world was through what we now call "traditional media", like televisions and newspapers, nowadays the use of the internet, in particular social media, is paramount.
Unfortunately, social media for news consumption is a double-edged sword. On the one hand, its low cost, easy access, and rapid dissemination of information leads people to seek out and consume news from social media. On the other hand, it enables the wide spread of "fake news", (i.e, low quality news with intentionally false information). The extensive spread of fake news has the potential for extremely negative impacts on individuals and society. Therefore, fake news detection on social media has become an emerging research area that is attracting tremendous attention. Up until recently, fake news detection methods have largely borrowed techniques from the broad data mining field, focusing mostly on single aspects of the issue at a time, mostly content-based. A key weakness of these kind of approaches is that often the interpretation of the news requires the knowledge of political or social context or "common sense", which current natural language processing algorithms are still missing. Recent studies though, have empirically shown that fake and real news spread differently on social media, forming propagation patterns that could be harnessed for the automatic fake news detection.
In this context, geometrical deep learning emerges as a suitable candidate to pick up such useful features. The underlying core algorithms are a generalization of classical convolutional neural networks to graphs, allowing the fusion of heterogeneous data such as content, user profile and activity, social graph, and news propagation.
This new framework, has brought a significant advancement with respect to results in the literature.