Neural Network-based Word Alignment through Score Aggregation
This addresses the problem of improving word alignment accuracy for machine translation systems, but it is incremental as it builds on existing neural methods.
The paper tackles word alignment for machine translation by proposing a neural network with score aggregation for unsupervised training, resulting in improved alignment accuracy over Fast Align by 7 AER on English-Czech, 6 AER on Romanian-English, and 1.7 AER on English-French.
We present a simple neural network for word alignment that builds source and target word window representations to compute alignment scores for sentence pairs. To enable unsupervised training, we use an aggregation operation that summarizes the alignment scores for a given target word. A soft-margin objective increases scores for true target words while decreasing scores for target words that are not present. Compared to the popular Fast Align model, our approach improves alignment accuracy by 7 AER on English-Czech, by 6 AER on Romanian-English and by 1.7 AER on English-French alignment.