Semisupervised Machine Fault Diagnosis Fusing Unsupervised Graph Contrastive Learning

By learning effective information from unlabeled nodes, node-level graph data-driven diagnosis methods perform better than graph-level methods. However, features of unlabeled nodes, indirectly involved in graph feature learning, are not fully utilized. To overcome aforementioned limitations, a semisupervised machine fault diagnosis fusing unsupervised graph contrastive learning (GCL) is proposed. A new GCL framework, where positive and negative graphs are generated by calculating Pearson correlation coefficient, is fused into the graph transformer network (GTN). Furthermore, a new combined loss, including a supervised cross-entropy loss and a new unsupervised GCL loss, is designed for GTN training. Contrastive learning of positive and negative graphs is guided by the unsupervised GCL loss. While the semisupervised graph feature learning for original graphs is mainly driven by the supervised cross-entropy loss, where the GTN for graph feature learning shares parameters. Experimental results on public and real datasets show the proposed method achieves a competitive performance.

Automated summary

Node-level graph data-driven diagnosis methods outperform graph-level methods by effectively learning information from unlabeled nodes. However, the features of these nodes, indirectly involved in graph feature learning, are not fully utilized. To address this, a semisupervised machine fault diagnosis method is proposed, which combines unsupervised graph contrastive learning (GCL). A new GCL framework is fused into the graph transformer network (GTN) to generate positive and negative graphs based on Pearson correlation coefficient. The GTN training includes a supervised cross-entropy loss and a new unsupervised GCL loss, guiding the contrastive learning of positive and negative graphs. This method achieves competitive performance according to experimental results on public and real datasets.