Graph Neural Networks (GNNs)

Our lab explores the application of Graph Neural Networks to tackle challenges in fault detection, diagnostics, and predictive maintenance across industrial and environmental domains. Our work focuses on advancing spatial-temporal modeling, improving generalization, and integrating domain knowledge for robust performance.

Advanced Methodologies in GNNs

Learning physics-consistent particle interactions

Filter-Informed Spectral Graph Wavelet Networks for Multiscale Feature Extraction and Intelligent Fault Diagnosis

Industrial Equipment and Predictive Maintenance

Graph Neural Networks for Dynamic Modeling of Roller Bearing

Early Fault Detection in IIoT Systems

Environmental Monitoring and Infrastructure Health

Leakage Detection in Water Distribution Networks

IoT Sensor Calibration for Air Pollution Monitoring