重温图对比学习在异常检测中的应用（英文版）

英文标题：Revisiting Graph Contrastive Learning for Anomaly Detection中文摘要：本文提出了基于对比学习的图神经网络异常检测方法 MAG，强调了多 GNN 模块与图扩展模块在提高检测性能中的作用，并通过两个变种 L-MAG 和 M-MAG 在 Cora 和 Pubmed 数据集上证明了其有效性。英文摘要：Combining Graph neural networks (GNNs) with contrastive learning for anomalydetection has drawn rising attention recently. Existing graph contrastiveanomaly detection (GCAD) methods have primarily focused on improving detectioncapability through graph augmentation and multi-scale contrast modules.However, the underlyi