JuliaML backend for parameter estimation

[ChrisRackauckas]

Following up on #47, it would be nice to add JuliaML support to the parameter estimation. This would look a lot like the Optim.jl support shown here in the docs DifferentialEquations.jl (specifically, through DiffEqParamEstim.jl) spits out an objective function for Learn.jl or something like that to solve.

The code to get this done will likely be very simple. For example, this is the entire code to get it for Optim.jl: https://github.com/JuliaDiffEq/DiffEqParamEstim.jl/blob/master/src/DiffEqParamEstim.jl#L21 . But we will need some JuliaML support since I'm not familiar with how to use their tools (and some of it still needs to be developed).

@ahwillia

[ChrisRackauckas]

Migrated to SciML/DiffEqParamEstim.jl#2