Artificial Intelligence and Physics-Informed Machine Learning for Sustainability Challenges

The artificial intelligence and physics-informed machine learning for sustainability challenges thematic area leverages cutting-edge artificial intelligence and machine learning techniques to address pressing global sustainability issues. By integrating data-driven models with physics-based principles, researchers create robust tools for solving complex problems of the built and natural environments. 

Example research topics include: 

• Machine Learning
  Example projects: Computing fragility of infrastructure; Detecting erosion; Predictive models for structural deterioration; Predictive models for earthquake ground motions;  Improving flow measurement in rivers and streams; Better and faster flood and erosion modeling; Forecasting carbon cycling, river water quantity and quality in a rapidly changing world (e.g., intensified climate change and human activities); Predicting traveler’s behavior and planning electric vehicle charging facilities; Predicting electric vehicle performance with different weather patterns and payloads 
• Artificial Intelligence
  Example projects: Multi-agent deep reinforcement learning for infrastructure management; Decision-making under uncertainty; building AI models for transportation infrastructure deterioration prediction; Predicting how automated vehicles may impact pavement performance
  
• Computational Solid Mechanics
  Example projects: Uncertainty quantification of structural performance; System identification based on monitoring data
  
• Computational Fluid Mechanics
  Example projects: Improving passage of aquatic organisms; Roadway hydraulic structures; Improving pavement drainage; Nature-based solutions
  
• Numerical Methods in Geotechnical Engineering 
  
• Example projects: Material point method for penetration problems; Assessing levee stability and failure mechanisms using coupled material point method
  
• Physics-based Modeling for Understanding Processes, Drivers, and Patterns of Changing Hydrological Systems
  Example projects: Understanding global low flow dynamics under climate change with next-generation, differentiable global hydrologic models; Chesapeake Bay HydroML: Advanced Streamflow and Water Quality Predictions for the Chesapeake Bay Watershed; Ecosystem dynamics models; Deep learning techniques to understand nitrate concentration dynamics