Research Projects
The REMISS project develops a new methodology that allows a trust and credibility score to be awarded both to social network accounts and to messages that may be false. For this reason, it combines various techniques that, until now, have been developed in independent silos and have given powerful results in terms of detection, although they are not very interpretable. The project proposes a methodology based on propagation models, multimodal analysis and text characterization derived from metrics obtained in the laboratory. This analysis will allow to understand the motivations behind the exchange of information (and disinformation) and the types of characteristics that make a text credible and susceptible to being propagated. Disinformation on social networks is part of a widespread phenomenon that affects the exchange and dissemination of content. The problems it generates are well known, from the amplification of general distrust on the part of society, to the exacerbation of the mechanisms of social polarization or stigmatization of vulnerable groups. This situation has gone beyond the merely professional sphere, both for researchers and journalists, and the threat of the generation and dissemination of false content and news is already internalized by the majority of society. Understanding how fake news is generated and spread is critical to developing evidence-based policies and actions to detect and mitigate its spread. Eurecat coordinates the project through its Big Data & Data Science Unit, which participates in research in computational social sciences and cognitive analysis. The REMISS consortium is made up of five other partners: CCMA and Verificat, which provide verified and unverified data; ESADE, with extensive experience in computational social sciences and behavioral analysis; and CVC and UVEG, leading experts in the areas of image analysis, text analysis, and natural language processing techniques.