Unravelling the Performance of Physics-informed Graph Neural Networks for Dynamical Systems

We briefly explain the theoretical formulation highlighting the similarities and differences in the inductive biases and graph architecture of these systems.