Neural Baselines for Word Alignment
This work addresses word alignment for machine translation, but it is incremental as it applies existing neural approaches to established models.
The paper tackled the problem of unsupervised word alignment for machine translation by evaluating neural versions of IBM-1 and hidden Markov models, showing that these neural models vastly outperform their discrete counterparts in most settings across four language pairs.
Word alignments identify translational correspondences between words in a parallel sentence pair and is used, for instance, to learn bilingual dictionaries, to train statistical machine translation systems , or to perform quality estimation. In most areas of natural language processing, neural network models nowadays constitute the preferred approach, a situation that might also apply to word alignment models. In this work, we study and comprehensively evaluate neural models for unsupervised word alignment for four language pairs, contrasting several variants of neural models. We show that in most settings, neural versions of the IBM-1 and hidden Markov models vastly outperform their discrete counterparts. We also analyze typical alignment errors of the baselines that our models overcome to illustrate the benefits-and the limitations-of these new models for morphologically rich languages.