Skip to content
/ MT-LM Public

Domain-Specific Text Generation for Machine Translation (with LLMs) - scripts and config files for the paper

Notifications You must be signed in to change notification settings

ymoslem/MT-LM

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

19 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Domain-Specific Text Generation for Machine Translation

Scripts and config files for our paper, Domain-Specific Text Generation for Machine Translation

Summary

Using Large Language Models (LLMs) to generate synthetic data that simulate the domain and linguistic characteristics of the authentic data. You can check this article for an extended summary or the full paper.

Citation

@inproceedings{moslem-etal-2022-domain,
    title = "Domain-Specific Text Generation for Machine Translation",
    author = "Moslem, Yasmin  and
      Haque, Rejwanul  and
      Kelleher, John  and
      Way, Andy",
    booktitle = "Proceedings of the 15th biennial conference of the Association for Machine Translation in the Americas (Volume 1: Research Track)",
    month = sep,
    year = "2022",
    address = "Orlando, USA",
    publisher = "Association for Machine Translation in the Americas",
    url = "https://aclanthology.org/2022.amta-research.2",
    pages = "14--30",
    abstract = "Preservation of domain knowledge from the source to target is crucial in any translation workflow. It is common in the translation industry to receive highly-specialized projects, where there is hardly any parallel in-domain data. In such scenarios where there is insufficient in-domain data to fine-tune Machine Translation (MT) models, producing translations that are consistent with the relevant context is challenging. In this work, we propose leveraging state-of-the-art pretrained language models (LMs) for domain-specific data augmentation for MT, simulating the domain characteristics of either (a) a small bilingual dataset, or (b) the monolingual source text to be translated. Combining this idea with back-translation, we can generate huge amounts of synthetic bilingual in-domain data for both use cases. For our investigation, we used the state-of-the-art MT architecture, Transformer. We employed mixed fine-tuning to train models that significantly improve translation of in-domain texts. More specifically, our proposed methods achieved improvements of approximately 5-6 BLEU and 2-3 BLEU, respectively, on Arabic-to-English and English-to-Arabic language pairs. Furthermore, the outcome of human evaluation corroborates the automatic evaluation results.",
}

About

Domain-Specific Text Generation for Machine Translation (with LLMs) - scripts and config files for the paper

Topics

Resources

Stars

Watchers

Forks

Languages