Semi Supervised Graph Anomaly Detection

Semi-supervised graph anomaly detection aims to identify unusual patterns in graph data using a limited number of labeled nodes, leveraging the structure and features of the graph to improve detection accuracy. Current research focuses on developing novel generative models, incorporating multi-task learning and active learning strategies, and improving model generalizability through data augmentation techniques, often employing graph neural networks. These advancements are crucial for enhancing the robustness and reliability of various applications, including web security, microservice system monitoring, and other domains where graph-structured data is prevalent.