The nerve!

In this work we demonstrate how modern-day statistical MT system can be constrained to search for translations obeying particular length, meter, and rhyming constraints, whether a single constraint, or any one of a set. We further demonstrate that the hypothesis space is often rich enough that these constraints can be satisfied. The impact on translation quality, however, is quite profound, as is to be expected. It seems that at the present state of machine translation, one does indeed have to choose betweengetting either the form or the meaning right. In the present form, however, we can already find good translations, as a sort of found poetry, by translating a large quantity of text, whether poetic or not.