Our team will develop fundamental and generalizable computational advances.
- Develop a computational theory of local-to-global dynamics over multi-scale multi-layer (MSML) networks.
- Develop computational foundations for forecasting, control, and optimization problems in epidemiology.
- Discover fundamental limits to forecasting and inference.
- Develop new statistical and machine learning techniques for ensemble modeling, spatial detection of weak signals, and change detection.
- Explore the use of crowdsourced and active learning methods for inferring individual and community level awareness and behavioral changes.
- Characterize the joint effects of network structure and local interactions on spreading processes.
- Develop HPC-enabled rigorous solutions for scalable validation, calibration, sensitivity analysis, and uncertainty quantification for spreading processes on MSML networks.
Expected outcomes and computational advances:
- Foundations of computational network science, especially with respect to:
- an algorithmic local-to-global theory that seeks to understand the composed dynamics as a function of network structure and local dynamics, and
- algorithmic foundations of spreading processes over MSML networks.
- Rigorous algorithms and practical scalable methods for inferring properties of MSML networks.
- New theory-guided deep learning methods that work on MSML networks and operate on sparse observational data.
- New methods for transfer learning that can be used to develop innovative interventions based on a particular disease in a specific region to another disease in a related region.