L3S Best Paper of the quarter (Q2/2024)
Category: NLP
A Trustworthy Approach to Classify and Analyze Epidemic-Related Information From Microblogs
Authors: Thi Huyen Nguyen, Marco Fisichella, Koustav Rudra
Published in IEEE Transactions on Computational Social Systems
https://ieeexplore.ieee.org/abstract/document/10530086
The paper in a nutshell:
We introduce a trustworthy approach for the classification of tweets during disease outbreaks. Our method can efficiently identify tweets from different crucial classes such as signs and symptoms, transmission, prevention measures, etc., and extract explanations for the output decisions simultaneously. Besides, we propose a simple graph-based ranking method to generate short summaries of tweets.
What is the potential impact of your findings?
Time-critical analysis of crisis-related information helps humanitarian organizations and governmental bodies gain actionable information and plan for situational response. Our work aims to support affected communities and stakeholders obtain rapid and crucial updates without being overwhelmed by massive content posted on social media during epidemics. By integrating interpretability, our study can assist end-users to understand and enhance their trust in the decisions of ML models for real-life applications.
What is new about your research?
Unlike most of the previous works on epidemic-related tweet classification, which merely focus on model performance but not model transparency, our model is interpretable by design. The proposed classifier improves both the accuracy and explainability of the classification. Besides, our graph-based model is simple, yet efficient for generating concise situational updates of tweets from different classes during epidemic situations.
