Epidemiological Modeling

Agent-based models simulate heterogeneous individuals and interactions to capture emergent epidemic dynamics. Emerging epidemic estimation is affected by systematic biases that need remediation. Bayesian methods are widely used to estimate parameters and quantify uncertainty in outbreak settings. Transfer learning helps forecast new diseases in low-data settings by using knowledge from related diseases. Deep learning has been applied to phylogenetic inference, infectious disease detection from records, and epidemic forecasting. Differentiable agent-based models support gradient-based calibration and sensitivity analysis through automatic differentiation.