Following the Russian Federation's full-scale invasion of Ukraine in February 2022, a multitude of information narratives emerged within both pro-Russian and pro-Ukrainian communities online. As the conflict progresses, so too do the information narratives, constantly adapting and influencing local and global community perceptions and attitudes. This dynamic nature of the evolving information environment (IE) underscores a critical need to fully discern how narratives evolve and affect online communities. Existing research, however, often fails to capture information narrative evolution, overlooking both the fluid nature of narratives and the internal mechanisms that drive their evolution. Recognizing this, we introduce a novel approach designed to both model narrative evolution and uncover the underlying mechanisms driving them. In this work we perform a comparative discourse analysis across communities on Telegram covering the initial three months following the invasion. First, we uncover substantial disparities in narratives and perceptions between pro-Russian and pro-Ukrainian communities. Then, we probe deeper into prevalent narratives of each group, identifying key themes and examining the underlying mechanisms fueling their evolution. Finally, we explore influences and factors that may shape the development and spread of narratives.

Othering—the process of portraying an outgroup as fundamentally different and inferior—often escalates into framing the outgroup as an existential threat, thereby legitimizing exclusion and violence. Throughout history, othering has played a central role in conflicts, from genocides in Nazi Germany and Rwanda to contemporary hostility toward migrants in the US and Europe. Traditional computational methods, such as those used for hate speech detection, frequently overlook the subtle, context-dependent nature of othering language, limiting their effectiveness in real-time detection and analysis. Our work addresses these limitations through three key contributions: (1) a computational framework that combines sociological theory with large language models (LLMs) to identify and analyze othering language, (2) an in-depth examination of othering discourse dynamics, focusing on attention patterns and its interplay with moral framing, and (3) a rapid domain adaptation enabling robust analysis across different platforms and contexts. We apply our framework to a large corpus of Telegram messages from Russo-Ukrainian war bloggers and political discourse on Gab, revealing several previously unquantified patterns: othering rhetoric surges during crises, often intertwines with moralized language, and escalates during critical periods. Our findings demonstrate that this approach not only surpasses existing hate and fear speech detection methods but also offers actionable insights for anticipating and mitigating threats to social cohesion in conflict-prone environments.

Formally announced to the public following former President Donald Trump’s bans and suspensions from mainstream social networks in early 2022 following his role in the January 6 Capitol Riots, Truth Social was launched as an ``alternative'' social media platform that claims to be a refuge for free speech, offering a platform for those disaffected by the content moderation policies of then existing, mainstream social networks. The subsequent rise of Truth Social has been driven largely by hard-line supporters of the former president as well as those affected by the content moderation of other social networks. These distinct qualities combined with the its status as the main mouthpiece of the former president positions Truth Social as a particularly influential social media platform and give rise to several research questions. However, outside of a handful of news reports, little is known about the new social media platform partially due to a lack of well-curated data. In the current work, we describe a dataset of over 823,000 posts to Truth Social and and social network with over 454,000 distinct users. In addition to the dataset itself, we also present some basic analysis of its content, certain temporal features, and its network.

The spread of COVID-19 throughout the world has led to cataclysmic consequences on the global community, which poses an urgent need to accurately understand and predict the trajectories of the pandemic. Existing research has relied on graph-structured human mobility data for the task of pandemic forecasting. To perform pandemic forecasting of COVID-19 in the United States, we curate Large-MG, a large-scale mobility dataset that contains 66 dynamic mobility graphs, with each graph having over 3k nodes and an average of 540k edges. One drawback with existing Graph Neural Networks (GNNs) for pandemic forecasting is that they generally perform information propagation in a flat way and thus ignore the inherent community structure in a mobility graph. To bridge this gap, we propose a Hierarchical Spatio-Temporal Graph Neural Network (HiSTGNN) to perform pandemic forecasting, which learns both spatial and temporal information from a sequence of dynamic mobility graphs. HiSTGNN consists of two network architectures. One is a hierarchical graph neural network (HiGNN) that constructs a two-level neural architecture: county-level and region-level, and performs information propagation in a hierarchical way. The other network architecture is a Transformer-based model that captures the temporal dynamics among the sequence of learned node representations from HiGNN. Additionally, we introduce a joint learning objective to further optimize HiSTGNN. Extensive experiments have demonstrated HiSTGNN's superior predictive power of COVID-19 new case/death counts compared with state-of-the-art baselines.