Collective narratives

collaboration work with USC team (check pdf 😉)

[abstract]: Narrative is a foundation of human cognition and decision making. As such narratives play a crucial role in political campaigns and spread of conspiracy theories and misinfor- mation. With the pervasive use of social media, the competition and spreading of narratives on social media can have profound impact on societal issues. Yet, computational understanding narratives circulated online is faced with many challenges such as the scale of the problem and the inherent difficulty of reliably identifying and extracting narratives from texts. Furthermore, many existing methods do not con- sider the temporality of the collective narratives—how nar- ratives emerge, spread, and die. In this study, we propose a comprehensive narrative discovery framework that fill this gap by combining change point detection, semantic role la- beling (SRL), and automatic aggregation of narrative frag- ments into narrative networks. The model was tested on two Twitter data, one related to COVID-19, and the other is about 2017 French Election. Our results demonstrate that our ap- proach can recover major narrative shifts that correspond to the major events.