Log In Sign Up

End-to-End Neural Word Alignment Outperforms GIZA++

by   Thomas Zenkel, et al.

Word alignment was once a core unsupervised learning task in natural language processing because of its essential role in training statistical machine translation (MT) models. Although unnecessary for training neural MT models, word alignment still plays an important role in interactive applications of neural machine translation, such as annotation transfer and lexicon injection. While statistical MT methods have been replaced by neural approaches with superior performance, the twenty-year-old GIZA++ toolkit remains a key component of state-of-the-art word alignment systems. Prior work on neural word alignment has only been able to outperform GIZA++ by using its output during training. We present the first end-to-end neural word alignment method that consistently outperforms GIZA++ on three data sets. Our approach repurposes a Transformer model trained for supervised translation to also serve as an unsupervised word alignment model in a manner that is tightly integrated and does not affect translation quality.


page 1

page 2

page 3

page 4


Jointly Learning to Align and Translate with Transformer Models

The state of the art in machine translation (MT) is governed by neural a...

Leveraging Neural Machine Translation for Word Alignment

The most common tools for word-alignment rely on a large amount of paral...

Neural Baselines for Word Alignment

Word alignments identify translational correspondences between words in ...

Adding Interpretable Attention to Neural Translation Models Improves Word Alignment

Multi-layer models with multiple attention heads per layer provide super...

Generative latent neural models for automatic word alignment

Word alignments identify translational correspondences between words in ...

Smoothing parameter estimation framework for IBM word alignment models

IBM models are very important word alignment models in Machine Translati...

Word Alignment in the Era of Deep Learning: A Tutorial

The word alignment task, despite its prominence in the era of statistica...