INSIDER THREAT DETECTION USING GRAPH NEURAL NETWORKS ON ENTERPRISE ACCESS LOGS
Subjects/Theme:
Insider Threat Detection, Graph Neural Networks, Enterprise Security, Access Logs, Anomaly Detection, Cybersecurity Analytics, Deep Learning, Behavioral ModelingDescription
Security and Privacy in AI Systems,
Edited By: Dr. Sunita Chaudhary, Dr. Joydeb Patra
ISBN (978-81-685212-9-2)
Insider threats represent a critical challenge in modern enterprise security, often resulting in significant financial and reputational damage. Traditional detection systems relying on rule-based or statistical methods struggle to capture complex relationships embedded within enterprise access logs. This paper proposes a novel framework leveraging Graph Neural Networks (GNNs) to model user-entity interactions and detect anomalous insider behavior. By transforming access logs into graph structures, where users, devices, and resources are represented as nodes and their interactions as edges, GNNs effectively learn relational patterns and detect deviations indicative of malicious intent. Experimental evaluation on benchmark datasets demonstrates improved detection accuracy, reduced false positives, and enhanced interpretability compared to traditional machine learning models. The study highlights the potential of GNN-based approaches in strengthening enterprise cybersecurity frameworks.
