Abstract
Dynamic social networks, characterized by continuously evolving node and edge structures, demand advanced analytical models capable of both representation learning and efficient inference. Graph Neural Networks (GNNs) have emerged as powerful tools for learning on such structured data. This paper explores the integration of GNN architectures with dynamic social network data for real-time community detection. It compares traditional methods with GNN-based frameworks, outlines key architectural components, and provides empirical insights into performance benchmarks. Real-world applications in misinformation detection, trend prediction, and social recommendation systems are discussed.
View more »