Are Unaligned Words Important for Machine Translation
Yuqi Zhang, Evgeny Matusov, Hermann Ney · RWTH Publications (RWTH Aachen) · 2009
In this paper, we deal with the problem of a large number of unaligned words in automatically learned word alignments for machine translation (MT). These unaligned words are the reason for ambiguous phrase pairs extracted by a statistical phrase-based MT system. In translation, this phrase ambiguity causes deletion and insertion errors. We present hard and optional deletion approaches to remove the unaligned words in the source language sentences. Improvements in translation quality are achieved both on large and small vocabulary tasks with the presented methods. 1