Long-distance reordering during search for hierarchical phrase-based SMT
Fabienne Braune, Anita Gojun and Alexander Fraser
In: EAMT 2012, 28-30 May 2012, Trento, Italy.
Long-distance reordering of syntactically divergent language pairs is
a critical problem. SMT has had limited success in handling these
reorderings during inference, and thus deterministic preprocessing
based on reordering parse trees is used. We consider German-to-English
translation using Hiero. We show how to effectively model
long-distance reorderings during search. Our work is novel in that we
look at reordering distances of up to 50 words, and conduct a detailed
manual analysis based on a new gold standard.