arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2108.03533cs.AIcs.CL
收藏

利用迁移学习提高相似语言翻译质量

Improving Similar Language Translation With Transfer Learning

Natural Language Processing Lab(自然语言处理实验室);The University of British Columbia(英属哥伦比亚大学)

Ife Adebara, Muhammad Abdul-Mageed

摘要:

我们研究了基于预训练神经机器翻译模型的迁移学习,以在(低资源)相似语言之间进行翻译。这项工作是我们对WMT 2021相似语言翻译共享任务的贡献的一部分,我们在该任务中提交了不同语言对的模型,包括法语-班巴拉语,西班牙语-加泰罗尼亚语和西班牙语-葡萄牙语。我们的加泰罗尼亚语-西班牙语($82.79$ BLEU)和葡萄牙语-西班牙语($87.11$ BLEU)模型在官方共享任务评估中排名前1,我们是唯一提交法语-班巴拉配对模型的团队。

英文摘要:

We investigate transfer learning based on pre-trained neural machine translation models to translate between (low-resource) similar languages. This work is part of our contribution to the WMT 2021 Similar Languages Translation Shared Task where we submitted models for different language pairs, including French-Bambara, Spanish-Catalan, and Spanish-Portuguese in both directions. Our models for Catalan-Spanish ($82.79$ BLEU) and Portuguese-Spanish ($87.11$ BLEU) rank top 1 in the official shared task evaluation, and we are the only team to submit models for the French-Bambara pairs.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑