Neural Network-based Word Alignment through Score Aggregation
Computation and Language
2016-07-01 v1
Abstract
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.
Cite
@article{arxiv.1606.09560,
title = {Neural Network-based Word Alignment through Score Aggregation},
author = {Joel Legrand and Michael Auli and Ronan Collobert},
journal= {arXiv preprint arXiv:1606.09560},
year = {2016}
}