Scientific Machine Learning Through Physics-Informed Neural Networks: Where we are and What’s Next

HIGHLIGHTS

• who: Salvatore Cuomo from the In this survey, we focus on how PINNs are used to address different scientific computing problems, the building blocks of PINNs, the aspects related to learning theory, what toolsets are already available, future directions and recent trends, and issues regarding accuracy and convergenceAccording to different problems, we show how PINNs solvers have been customized in literature by configuring the depth, layer size, activation functions and using transfer learning. This article can be considered as an extensive literature review on PINNs. It was mostly conducted by searching Scopus for the terms: ((physic* OR . . .
  

   

  

  If you want to have access to all the content you need to log in!

  Thanks :)

  Username or Email

  Password

  Remember me

  Lost your password?

  If you don't have an account, you can create one here.