Research Experience for Undergraduates (REU), NHERI 2022: Implementing Physics Constraints into Graph Network-based Simulator for Natural Hazard Predictions

This project focuses on the process of further physically constraining a graph-based machine learning model, Graph Network-based Simulator (GNS), to limit the room for inaccuracy in the model’s predictions. GNS can be used to simulate particle dynamics which has the potential to predict many physically complex conditions such as natural hazards. With the advantage of natural hazard predictions, individuals may be warned of possible catastrophic damage caused by a natural hazard in advance which could greatly reduce loss through damages, property, and life. Professionals can use the data in this project to further improve the GNS model's predictions. This project is intended for anyone with an interest in Machine Learning and a desire to lessen the damages caused by natural hazards.