AINeutralarXiv – CS AI · Jun 96/10
🧠
Knowledge-Inclusive Adaptive Physics-Informed Neural Network for Microbial Interaction Modelling
Researchers propose a Physics-Informed Neural Network (PINN) framework that incorporates multiple knowledge sources—including peer-reviewed literature and network structures—to improve microbial community modeling beyond traditional equation-based approaches. The framework, applied to generalized Lotka-Volterra modeling, demonstrates significant performance improvements of up to 53% over existing methods, with additional gains of up to 23-47% when knowledge is integrated.