Advancing agent-based simulation scalability
High-fidelity simulation models provide the most accurate representation of the spread of COVID and related diseases but are subject to computational limitations. Agent-based modelling, which treats each individual as a unique agent with objectives/properties/etc., is an example of such a high-fidelity model. The project’s objective is to realize a vastly improved scalability of this type of analysis, using modern techniques in the field of deep learning and AI.
- Specifically in representation learning of dynamical processes.
- Co-Project Investigators: Christian Muise (Queen’s University), Gias Uddin (University of Calgary), Manos Papagelis (York University) and Morgan Craig (Université de Montréal)