The full article, “ Unsupervised Image-to-Image Translation Networks,” describes how this was achieved, primarily through a practice called unsupervised learning.