Physics-Informed Neural Networks

Neural networks have been used extensively in many fields with impressive results.
Most applications use data as the sole learning source. Recently, researchers have proposed
to use additional information about the latent data that gives birth to information-informed
machine learning. One of the recent applications has been in the field of differential
equations where we want the model to verify some equations, boundary conditions,
and possibly some data. In this manuscript, we will dive into the family of Physics-
Informed Neural Networks, exploring mathematical properties relating to its convergence
and error.