A Koopman-PINN Framework for Epidemic Models: Parameter Inference and Forecasting
Researchers have developed a new framework called Koopman-PINN that combines Koopman operator theory with physics-informed neural networks for improved epidemic modeling. This approach maps epidemic states into a latent space where dynamics are more linear, enhancing interpretability and long-term forecasting stability. The framework was tested on synthetic monkeypox data and real-world SARS-CoV-2 data from Germany, Morocco, and Sweden, demonstrating superior performance in parameter estimation and trajectory reconstruction compared to existing methods. AI
IMPACT This framework offers a more accurate and stable approach to modeling and forecasting epidemics, potentially improving public health responses.