Research map / Video / Video segmentation and tracking
MIDDAG: Where Does Our News Go? Investigating Information Diffusion via Community-Level Information Pathways
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Presents MIDDAG, an interactive system visualizing COVID-19 news information pathways on Twitter and Reddit at user and community levels, with pathway forecasting, susceptibility scoring, event extraction, and opinion summarization.
Problem
Understanding how COVID-19-related news spreads across social media, including forecasting propagation and characterizing user/community susceptibility and discussion content, which existing work does not visualize intuitively with comprehensive insights.
Approach
Combines community-level aggregation of user information pathways (via influence-passivity scoring), a GNN-based pathway link-prediction and autoregressive forecasting model, a RoBERTa-based contrastive susceptibility model, and T5-large instruction-tuned event extraction into one interactive visualization system.
Key results
- IP prediction GNN reaches 86.83% AUC on link prediction.
- Susceptibility model reaches 86.28% F1 for retweet behavior prediction.
- System covers 640 million tweets from 5.3 million users over May 15-30, 2020, plus 5,410 Reddit posts across 649 subreddits.
Limitations
- Analysis restricted to a 15-day subset of Twitter data due to dataset size.
- Susceptibility is indirectly validated through retweet prediction rather than direct belief measurement.
Experimental setup
Datasets: Twitter COVID-19 news dataset (640M tweets, 5.3M users) Reddit COVID-19 news posts (5,410 posts, 4,578 articles) ANTi-Vax CoAID 13 entity/relation/event/sentiment extraction datasets (UIE)
Metrics: AUC F1
Builds on: UIE Influence-Passivity algorithm
Closest papers by meaning
- Human Mobility Modeling during the COVID-19 Pandemic via Deep Graph Diffusion Infomax (AAAI 2023 · similarity 0.81)
- Make Information Diffusion Explainable: LLM-based Causal Framework for Diffusion Prediction (NeurIPS 2025 · similarity 0.80)
- VigDet: Knowledge Informed Neural Temporal Point Process for Coordination Detection on Social Media (NeurIPS 2021 · similarity 0.78)
- Characterizing Information Seeking Events in Health-Related Social Discourse (AAAI 2024 · similarity 0.78)
- Understanding Microtargeting Pattern on Social Media (AAAI 2025 · similarity 0.77)
- Promoting Fair Vaccination Strategies through Influence Maximization: A Case Study on COVID-19 Spread (AAAI 2024 · similarity 0.75)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.