GRINDA - Graph-based Intelligence for Network Disruption and Analysis

This work aims to develop a robust, data-driven framework for identifying key agents within criminal networks. By leveraging innovative Graph Neural Networks\u2014specifically Graph Attention Networks and Graph Transformers\u2014and advanced outlier detection techniques, we are creating a powerful tool for law enforcement. By transforming real-world crime data into graph structures, the approach aims to capture both structural and contextual features of criminal interactions, enabling the detection of strategically important individuals who are not necessarily central in traditional metrics. The ultimate objective is to enhance law enforcement capabilities in mapping, disrupting, and monitoring criminal networks through scalable, intelligent analysis tools, thereby contributing to a safer society.