A Discriminative Neural Model for Cross-Lingual Word Alignment
This work addresses the problem of improving word alignment accuracy for machine translation and downstream tasks like named entity recognition projection, though it is incremental as it builds on existing Transformer models.
The paper tackles cross-lingual word alignment by introducing a discriminative neural model integrated into a Transformer-based machine translation system, achieving major improvements of 11-27 F1 over unsupervised baselines on English-Chinese and English-Arabic alignment with small labeled datasets (~1.7K-5K sentences).
We introduce a novel discriminative word alignment model, which we integrate into a Transformer-based machine translation model. In experiments based on a small number of labeled examples (~1.7K-5K sentences) we evaluate its performance intrinsically on both English-Chinese and English-Arabic alignment, where we achieve major improvements over unsupervised baselines (11-27 F1). We evaluate the model extrinsically on data projection for Chinese NER, showing that our alignments lead to higher performance when used to project NER tags from English to Chinese. Finally, we perform an ablation analysis and an annotation experiment that jointly support the utility and feasibility of future manual alignment elicitation.