Zhiwei Wang is a civil engineer whose research focuses on enhancing the resilience of infrastructure systems against natural hazards by utilizing generative artificial intelligence (AI). His work aims to address critical challenges in assessing and managing risks to bridges, buildings, and transportation networks, particularly in the face of increasingly severe events such as earthquakes, hurricanes, and climate-driven extreme weather conditions. His training at MIDAS will focus on advanced machine learning, uncertainty quantification, and interdisciplinary collaboration, to translate cutting-edge AI methods into practical tools for engineers and policymakers.
Traditional engineering models for risk assessment often face limitations due to high uncertainty, limited data, and the difficulty of simulating rare but catastrophic events. Zhiwei’s research aims to overcome these challenges by leveraging advanced generative AI techniques—such as normalizing flows, diffusion models, and generative adversarial networks (GANs)—to generate realistic synthetic data and simulate a broad range of extreme scenarios that are otherwise difficult to capture. He also intends to use generative models for simulation model updating, integrating observed monitoring data to calibrate and correct predictive models, thereby reducing model bias and improving reliability. These efforts aim to enable more robust, data-driven approaches to understanding infrastructure vulnerability and to support improved strategies for design, maintenance, and risk mitigation.