Global-local integration for GNN-based anomalous device state detection in industrial control systems

作者：

Highlights：

• A GNN-based method, named the GLIN, is proposed to detect anomalous ICS devices.

• The GLIN features local-global semantic integration.

• The GLIN embeds global pooling and integration functionality.

• The GLIN surpasses existing learning baselines over the SWaT dataset.

摘要

•A GNN-based method, named the GLIN, is proposed to detect anomalous ICS devices.•The GLIN features local-global semantic integration.•The GLIN embeds global pooling and integration functionality.•The GLIN surpasses existing learning baselines over the SWaT dataset.

论文关键词：Global-local integration,Global pooling,Graph neural networks,Anomaly detection,Industrial control systems

论文评审过程：Received 27 April 2022, Revised 19 July 2022, Accepted 30 July 2022, Available online 4 August 2022, Version of Record 8 August 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2022.118345