Enhancing regional scale natural hazard simulations with artificial intelligence

Partner: Matt Schoettler, NHERI SimCenter, Government

Overview

Project Description

The NHERI Computational Modeling and Simulation Center (SimCenter) provides next-generation computational modeling and simulation software tools, user support, and educational materials to the natural hazards engineering research community with the goal of advancing the nation’s capability to simulate the impact of natural hazards on structures, lifelines, and communities. In addition, the Center will enable leaders to make more informed decisions about the need for and effectiveness of potential mitigation strategies.

The SimCenter is currently undertaking activities that focus on the identification of building vulnerabilities to multiple natural hazards, such as earthquakes, tsunamis, and hurricanes. One of the prospective methods being employed is artificial intelligence (AI) to support or enhance regional scale simulations. Specifically, deep neural networks are being trained to extract building information from street view or satellite images, which improves modeling assumptions by reducing uncertainties. Two important aspects in this work are data collection and annotation.

Student interns will support the data collection effort by using Python to download data from online databases or program efficient web scraping methods to gather building inventory information. Student interns will support the data annotation effort by using CVAT to annotate images for improving training algorithm accuracy. Annotation involves identifying building components in the images and assigning labels to images. Rigorous quality control is essential for delivering trustworthy products with application to real engineering problems and intended for use by a broad community. Students will meet regularly a postdoctoral scholar supervisor on the SimCenter’s software development team.

Expected Deliverable

Data collection, annotation, and application of deep learning neural networks.

What would a successful semester look like to you?

Student interns would work with postdoctoral scholars for data collection, annotation, and applying deep learning neural networks to supplement missing information in building inventories to reduce uncertainties in natural hazard simulations.

Data

Models

Conclusion