Embedding-Enhanced Giza++: Improving Alignment in Low- and High- Resource Scenarios Using Embedding Space Geometry
This addresses the challenge of improving word alignment accuracy for low- and high-resource language pairs, offering a novel approach that is not incremental but leverages embedding space geometry.
The paper tackles the problem of word alignment in natural language processing by introducing Embedding-Enhanced GIZA++, which outperforms the long-dominant GIZA++ method without relying on large models or supervision, achieving improvements of 8.5 to 12 AER in low-resource settings for three language pairs.
A popular natural language processing task decades ago, word alignment has been dominated until recently by GIZA++, a statistical method based on the 30-year-old IBM models. New methods that outperform GIZA++ primarily rely on large machine translation models, massively multilingual language models, or supervision from GIZA++ alignments itself. We introduce Embedding-Enhanced GIZA++, and outperform GIZA++ without any of the aforementioned factors. Taking advantage of monolingual embedding spaces of source and target language only, we exceed GIZA++'s performance in every tested scenario for three languages pairs. In the lowest-resource setting, we outperform GIZA++ by 8.5, 10.9, and 12 AER for Ro-En, De-En, and En-Fr, respectively. We release our code at https://github.com/kellymarchisio/ee-giza.