I <3 CE: Osman Lab at i3CE 2026

Shuojia did an oral presentation at the i3CE meeting in Korea, where she shared her recent work on using graph neural networks (GNNs) to improve sparse water quality prediction in water distribution systems. Her presentation focused on a practical challenge in water infrastructure: utilities often have limited sensor coverage, making it difficult to monitor water quality across an entire network in real time. Her work compared different GNN designs to understand when edge information and temporal history can help models make better predictions under sparse data conditions.
The meeting was a meaningful experience both academically and personally. It gave Shuojia the chance to present her research to a broader civil engineering and infrastructure community, receive feedback, and learn how other researchers are applying data-driven methods to real-world infrastructure problems. More importantly, the conference helped her think about how to communicate technical machine learning work in a way that connects to practical water management needs, while gaining new perspectives on how AI can support more resilient infrastructure systems.