Sliced Iterative Normalizing Flows

                         Biwei Dai 1 Uro Seljak 1 2

             Abstract                models usually parameterize the mapping using neural net-
    We develop an iterative (greedy) deep learning      works, and the training of these models typically consists
    (DL) algorithm which is able to transform an arbi-    of minimizing a dissimilarity measure between the model
    trary probability distribution function (PDF) into    distribution and the target  ...