The impact of a PEER funded research project "Machine Learning for Analysis and Risk Management of Complex Infrastructure Systems" is highlighted below. The project Principal Investigator (PI) is Jack W. Baker, Professor of Civil & Environmental Engineering, Stanford University. The Research Team includes Rodrigo Silva Lopez, PhD Student, Stanford University.
A calibrated neural network model could significantly impact decision-making for agencies such as Caltrans. It would enable quick and convenient evaluation of the benefits of upgrading particular bridges and road links. Rather than requiring a detailed traffic assignment and routing calculation for each scenario, a fast and simple calculation could be made. This could enable network risk metrics to be considered in standardized evaluations of capital projects such as bridge retrofits or network upgrades. If interpretable algorithms such as LIME result in informative predictions of key components, that result could be of broad importance in showing the value of using machine learning algorithms to understand performance complex infrastructure networks.