Hi, nice article!
But I still have some confusion between the concept of PINN and Neural Operator. Please correct me if I’m wrong:

  • PINN: A method to SOLVE PDE using NN by feeding the NN with some data on boundary conditions and other random data (if any), and by imposing the PDE as an additional loss function of the NN.

  • Neural Operator: A new type/architecture/variants(?) of NN that learns from data from PDE simulation. So, it can map the input and the output of the PDE better than a conventional NN.

Can you help me? Would love to have some discussions with you. Cheers!

Best regards,
Adam

GitHub-flavored Markdown & a sane subset of HTML is supported.