Discriminative word alignment with conditional random fields

TLDR

The paper proposes a novel method for inducing word alignments from sentence‑aligned data. The method employs a Conditional Random Field conditioned on both source and target texts, enabling arbitrary overlapping features and efficient globally optimal training and decoding. With only a few hundred training sentences, the model outperforms the state of the art, achieving alignment error rates of 5.29 % for French‑English and 25.8 % for Romanian‑English.

Abstract

In this paper we present a novel approach for inducing word alignments from sentence aligned data. We use a Conditional Random Field (CRF), a discriminative model, which is estimated on a small supervised training set. The CRF is conditioned on both the source and target texts, and thus allows for the use of arbitrary and overlapping features over these data. Moreover, the CRF has efficient training and decoding processes which both find globally optimal solutions.We apply this alignment model to both French-English and Romanian-English language pairs. We show how a large number of highly predictive features can be easily incorporated into the CRF, and demonstrate that even with only a few hundred word-aligned training sentences, our model improves over the current state-of-the-art with alignment error rates of 5.29 and 25.8 for the two tasks respectively.