Hi everybody,
I wanted to add a new tutorial to the documentation. My idea is to compare a physics informed neural network (PINN) to traditional numerical methods (like finite element or finite difference) for solving common partial differential equations (PDEs).
I would cover few examples like Burgers Equation or the Heat Equation, and document the accuracy, convergence rate , training and solving time etc.
Hopefully this would be very valuable for people who have fair idea of traditional methods and want to understand some practical implications of using PINNs.
Is this a contribution worth pursuing? Looking forward to feedback.