BinaryAlign: Word Alignment as Binary Sequence Labeling

TL;DR

BinaryAlign introduces a unified binary sequence labeling approach for word alignment that outperforms existing methods across high and low resource language pairs, with extensive analysis on multilingual models and error types.

Contribution

It presents a novel, unified word alignment method based on binary sequence labeling that surpasses prior approaches in diverse resource scenarios.

Findings

Outperforms existing word alignment models in various resource settings

Effective across multiple language pairs, including non-English languages

Provides detailed error analysis and model comparisons

Abstract

Real world deployments of word alignment are almost certain to cover both high and low resource languages. However, the state-of-the-art for this task recommends a different model class depending on the availability of gold alignment training data for a particular language pair. We propose BinaryAlign, a novel word alignment technique based on binary sequence labeling that outperforms existing approaches in both scenarios, offering a unifying approach to the task. Additionally, we vary the specific choice of multilingual foundation model, perform stratified error analysis over alignment error type, and explore the performance of BinaryAlign on non-English language pairs. We make our source code publicly available.