We develop an iterative (greedy) deep learning (DL) algorithm which is able
to transform an arbitrary probability distribution function (PDF) into the
target PDF. The model is based on iterative Optimal Transport of a series of 1D
slices, matching on each slice the marginal PDF to the