面向未见低资源语言的对比对齐指令调优大语言模型进行机器翻译
Tuning LLMs with Contrastive Alignment Instructions for Machine Translation in Unseen, Low-resource Languages
AI总结：
本文提出对比对齐指令AlignInstruct，通过跨语言判别器增强跨语言监督，微调BLOOMZ模型，在24种未见语言和48个翻译方向上显著提升机器翻译质量，并改善零样本性能。
摘要：
本文引入了对比对齐指令（AlignInstruct）来解决大语言模型（LLMs）在机器翻译（MT）中的两个挑战。其一是将支持的语言扩展到以前未见过的语言。其二是低资源语言数据的缺乏。通过机器翻译指令（MTInstruct）进行模型微调是应对第一个挑战的直接方法。然而，MTInstruct受到第二个挑战中固有的弱跨语言信号的限制。AlignInstruct通过利用统计词对齐构建的跨语言判别器来强调跨语言监督。我们基于在多达24种未见语言上微调BLOOMZ模型（1b1、3b和7b1）的结果表明：（1）LLMs可以使用MTInstruct有效地翻译未见语言；（2）AlignInstruct在涉及英语的48个翻译方向上带来了翻译质量的一致提升；（3）基于判别器的指令作为跨语言指令优于其生成式对应物；（4）AlignInstruct在30个零样本方向上提高了性能。
英文摘要：
This article introduces contrastive alignment instructions (AlignInstruct) to address two challenges in machine translation (MT) on large language models (LLMs). One is the expansion of supported languages to previously unseen ones. The second relates to the lack of data in low-resource languages. Model fine-tuning through MT instructions (MTInstruct) is a straightforward approach to the first challenge. However, MTInstruct is limited by weak cross-lingual signals inherent in the second challenge. AlignInstruct emphasizes cross-lingual supervision via a cross-lingual discriminator built using statistical word alignments. Our results based on fine-tuning the BLOOMZ models (1b1, 3b, and 7b1) in up to 24 unseen languages showed that: (1) LLMs can effectively translate unseen languages using MTInstruct; (2) AlignInstruct led to consistent improvements in translation quality across 48 translation directions involving English; (3) Discriminator-based instructions outperformed their generative counterparts as cross-lingual instructions; (4) AlignInstruct improved performance in 30 zero-shot directions.
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。
