Accurate Online Posterior Alignments for Principled Lexically-Constrained Decoding

TL;DR

This paper introduces a novel online posterior alignment method for machine translation that improves alignment accuracy and enhances lexically constrained translation quality across multiple language pairs.

Contribution

The paper presents a new online posterior alignment technique that integrates seamlessly with constrained decoding and outperforms existing methods in alignment error rates.

Findings

Consistent reduction in alignment error rates across five language pairs.

Significant BLEU score improvements at constrained positions.

Effective integration with existing decoding algorithms.

Abstract

Online alignment in machine translation refers to the task of aligning a target word to a source word when the target sequence has only been partially decoded. Good online alignments facilitate important applications such as lexically constrained translation where user-defined dictionaries are used to inject lexical constraints into the translation model. We propose a novel posterior alignment technique that is truly online in its execution and superior in terms of alignment error rates compared to existing methods. Our proposed inference technique jointly considers alignment and token probabilities in a principled manner and can be seamlessly integrated within existing constrained beam-search decoding algorithms. On five language pairs, including two distant language pairs, we achieve consistent drop in alignment error rates. When deployed on seven lexically constrained translation tasks, we achieve significant improvements in BLEU specifically around the constrained positions.