Shuming Liang

Dr. Shuming Liang serves as a Postdoctoral Research Fellow at The Data Science Institute within the Faculty of Engineering and Information Technology at University of Technology Sydney (UTS). His research bridges machine learning theory with practical applications in critical infrastructure management, particularly focusing on water distribution systems and transportation networks.

His educational background includes a Ph.D. in Data Science from UTS (2020-2024), a Master of Science in Astrophysics from the Chinese Academy of Sciences (2010-2014), and a Bachelor of Science in Physics from the University of Jinan (2006-2010).

Dr. Liang's research interests center on Machine Learning with specialization in Graph Neural Networks, where he develops novel frameworks for anomaly detection, link prediction, and infrastructure failure forecasting. His work uniquely integrates time-series data, geospatial data, and satellite imagery to solve real-world problems in water resource management and transportation optimization. Recent publications demonstrate his expertise in creating multi-task learning approaches that enhance the interpretability and scalability of GNNs for industrial applications.

His publication trend shows increasing focus on practical implementations of machine learning in civil infrastructure, with recent work emphasizing anomaly detection in sensor networks (2025), generalization techniques for GNNs (2025), and applications in water pipe network management (2021-2024). His research consistently addresses the challenge of applying theoretical machine learning advances to solve tangible problems in critical infrastructure systems.

Dr. Liang actively collaborates with industry partners through funded research projects, particularly with Sydney Water Corporation, where his work contributes to proactive maintenance strategies for water distribution networks. His research has practical implications for reducing water loss, improving infrastructure reliability, and optimizing resource allocation in urban environments.

His laboratory work focuses on developing data-driven frameworks that integrate multiple data sources (pipe features, network structure, geographical information, and temporal failure patterns) to create comprehensive predictive models for infrastructure management. Current projects involve implementing these models in real-world water utility operations to enable targeted pipe inspection and renewal planning.