The purpose of this project is to develop a flexible library of NN solvers (NeuralNetDiffEq.jl) which are able to utilize parallelization, and plugs into the DifferentialEquations.jl ecosystem giving user an easy option to switch to these methods. The problem to be solved and advantages of the proposed approach are discussed in the Introduction of the proposal. The library will implement as much of DE solver interface features we have in julia along with some integration with other related packages. The goals and timeline for the project are also discussed in the proposal. I have a single ODE solver as a prototype and code can be checked here.



Akshay Sharma


  • Lyndon White`
  • Chris Rackauckas