NeuralPDE.jl is a Julia package that employs neural networks to obtain a solution of high dimensional differential equations and hence overcoming the curse of dimensionality.One of the techniques that uses Scientific Machine Learning to solve PDEs is Physics Informed Neural Networks (PINNs). The project aims at extending the application of the PINN algorithm to integro- differential equations. The project also propose to add support for constrained equations and different geometry for PINNs.

Organization

Student

Ashutosh Bharambe

Mentors

  • Chris Rackauckas
  • Kirill Zubov
close

2021