Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Modeling Information Narrative Detection and Evolution on Telegram during the Russia-Ukraine War

Patrick Gerard, Svitlana Volkova, Louis Penafiel, Kristina Lerman, Tim Weninger

TL;DR

The paper addresses how information narratives evolve in online communities during the Russia–Ukraine war by analyzing Telegram data from Russian- and Ukrainian-leaning channels. It introduces a two-tier methodology that combines online hierarchical clustering (OnlineAgglomerative HAC) with a MacroNarrative layer (and HITL refinement) to detect, track, and synthesize evolving narratives over time. Key findings include rapid, event-aligned reactions in both communities, distinct narrative priorities (e.g., humanitarian or geopolitical framing), and a Bucha Massacre case study illustrating cross-community divergence. The approach offers a dynamic, scalable tool for Operations in the Information Environment (OIE) to monitor information flows, adapt strategies, and enhance situational awareness in volatile information landscapes.

Abstract

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.

Modeling Information Narrative Detection and Evolution on Telegram during the Russia-Ukraine War

TL;DR

The paper addresses how information narratives evolve in online communities during the Russia–Ukraine war by analyzing Telegram data from Russian- and Ukrainian-leaning channels. It introduces a two-tier methodology that combines online hierarchical clustering (OnlineAgglomerative HAC) with a MacroNarrative layer (and HITL refinement) to detect, track, and synthesize evolving narratives over time. Key findings include rapid, event-aligned reactions in both communities, distinct narrative priorities (e.g., humanitarian or geopolitical framing), and a Bucha Massacre case study illustrating cross-community divergence. The approach offers a dynamic, scalable tool for Operations in the Information Environment (OIE) to monitor information flows, adapt strategies, and enhance situational awareness in volatile information landscapes.

Abstract

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.
Paper Structure (50 sections, 4 figures, 5 tables)

This paper contains 50 sections, 4 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Related Work and Background
  3. Narrative Detection
  4. Information Spread and Narrative Evolution
  5. Narrative Dynamics during RU-UA War
  6. Clustering and Topic Modeling Limitations
  7. Methodology
  8. Our Model -- Clustering Streaming Text Data
  9. Hierarchical Agglomerative Clustering (HAC) for Dynamic Narrative Analysis
  10. Adapting HAC to Evolving Data Streams
  11. Exploiting Online Data Integration
  12. Advancing Our Model -- Macro-Narratives
  13. Macro-Narrative Detection
  14. Step 1: Discovery and Refinement of Narrative Seeds
  15. Step 2: Human-in-the-Loop (HITL) Refinement
  16. ...and 35 more sections

Figures (4)

  • Figure 1: Contrasting information narratives surrounding the Bucha Massacre across Russian and Ukranian communities.
  • Figure 2: Themes discovered in narratives about Denazifying Ukraine.
  • Figure 3: Information narrative evolution shown using max-normalized frequency of posts in key narratives within Ukrainian and Russian communities over time (summed over 3-day time-periods for clarity), where max-normalization adjusts frequency counts relative to each narrative's peak activity for comparative clarity.
  • Figure 4: Information narrative evolution of dominant theme distribution of story clusters focused on Russian atrocities. To make this we look at the top 2 dominant themes in top 5 most populated story clusters; then, we aggregate this at a monthly timeframe (thus, having 4 points for potential pivoting) to understand general flow of these themes.