Identifying Social Media Bots and Inauthentic Users: A Lightweight Graph-Based Enhancement Framework
Identifying Social Media Bots and Inauthentic Users: A Lightweight Graph-Based Enhancement Framework
Shay Matok,M.Sc student at the department of Industrial Engineering
Advisor: Prof. Irad Ben-Gal
Abstract:
Social media platforms have revolutionized communication and information access, but they are also increasingly exploited by bots and inauthentic users. These malicious accounts can distort online discourse, amplify false narratives, and coordinate hidden networks to manipulate public opinion. While existing detection methods often rely on either simplistic models using partial information or complex deep learning models requiring extensive labeled datasets and significant computational power, they fall short in balancing accuracy and efficiency. This presentation introduces a lightweight, two-stage framework that integrates a weaker classifier—such as Random Forest or XGBoost—with a Relational Graph Attention Network (RGAT) to refine predictions based on social connections. The proposed approach enhances the detection of inauthentic activity in low-resource settings by leveraging relational signals and coordination patterns often overlooked by other methods, while maintaining modularity and scalability.
Bio:
Shay Matok is an MSc student in the faculty of industrial engineering. His current research focuses on Identifying inauthentic users in social media while using graph based methods.
