Active learning

[marcobonici]

In this moment we are using a general purpose scheme to train our emulators, the LHS.
The goal of this issue is to test some active learning strategies.
This either requires something as ConformalPredictions.jl or the work done by Jaime Sullivan.