A Kernel of Truth: Determining Rumor Veracity on Twitter by Diffusion Pattern Alone
This addresses the challenge of misinformation detection on social media by providing a robust method that avoids manipulation-prone signals like text and user identities.
The paper tackled the problem of detecting rumor veracity on Twitter by analyzing only the diffusion patterns of information, without using text or user identities, and demonstrated that graph kernels on cascade structures can accurately predict truth or falsehood, even early in propagation.
Recent work in the domain of misinformation detection has leveraged rich signals in the text and user identities associated with content on social media. But text can be strategically manipulated and accounts reopened under different aliases, suggesting that these approaches are inherently brittle. In this work, we investigate an alternative modality that is naturally robust: the pattern in which information propagates. Can the veracity of an unverified rumor spreading online be discerned solely on the basis of its pattern of diffusion through the social network? Using graph kernels to extract complex topological information from Twitter cascade structures, we train accurate predictive models that are blind to language, user identities, and time, demonstrating for the first time that such "sanitized" diffusion patterns are highly informative of veracity. Our results indicate that, with proper aggregation, the collective sharing pattern of the crowd may reveal powerful signals of rumor truth or falsehood, even in the early stages of propagation.