Domain Adaptation of Machine Translation with Crowdworkers

TL;DR

This paper presents a framework that uses crowdworkers to efficiently collect domain-specific parallel data from the web, enabling quick adaptation of machine translation models to new domains with significant improvements in translation quality.

Contribution

It introduces a novel crowdworker-based data collection method for domain adaptation in machine translation, reducing time and cost compared to traditional approaches.

Findings

Collected domain-specific data in days at reasonable cost

Achieved an average BLEU score improvement of +7.8 points

Improved translation quality across five different domains

Abstract

Although a machine translation model trained with a large in-domain parallel corpus achieves remarkable results, it still works poorly when no in-domain data are available. This situation restricts the applicability of machine translation when the target domain's data are limited. However, there is great demand for high-quality domain-specific machine translation models for many domains. We propose a framework that efficiently and effectively collects parallel sentences in a target domain from the web with the help of crowdworkers. With the collected parallel data, we can quickly adapt a machine translation model to the target domain. Our experiments show that the proposed method can collect target-domain parallel data over a few days at a reasonable cost. We tested it with five domains, and the domain-adapted model improved the BLEU scores to +19.7 by an average of +7.8 points compared to a general-purpose translation model.