A graph embedded in graph framework with dual-sequence input for efficient anomaly detection of complex equipment under insufficient samples

Haodong Yan, Fudong Li, Jinglong Chen, Zijun Liu, Jun Wang, Yong Feng, Xinwei Zhang

Abstract: Highlights•Propose GG-Nets to detect anomalies on the device side precisely and efficiently.•Learn the relationship between timestamps and sensors via the T-GAT and S-GAT.•Compress model size and improve inference speed through the duel input method.•Demonstrate the superiority via extensive experiments and discussions.