arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2401.05072cs.CL
收藏

在大语言模型中对齐翻译特定理解与通用理解

Aligning Translation-Specific Understanding to General Understanding in Large Language Models

Harbin Institute of Technology(哈尔滨工业大学);Peng Cheng Laboratory(鹏城实验室)

Yichong Huang, Baohang Li, Xiaocheng Feng, Chengpeng Fu, Wenshuai Huo, Ting Liu, Bing Qin

AI总结:

本文提出DUAT翻译过程,通过融入通用理解并利用跨语言解释与外部工具改进难词检测和解释生成,在Challenge-WMT基准上显著提升翻译质量并降低直译性。

摘要:

大语言模型(LLMs)在理解复杂文本方面展现出卓越能力,为实现接近人类的翻译表现提供了一条有前景的路径。然而,本研究揭示了LLMs内部翻译特定理解与通用理解之间的错位。这种理解错位导致LLMs在翻译某些复杂概念时出现误译或逐字直译,尽管它们在通用场景(如问答)中能够准确理解这些概念。为了使翻译特定理解与通用理解对齐,我们提出了一种新颖的翻译过程DUAT(Difficult words Understanding Aligned Translation,难词理解对齐翻译),显式地融入对引发理解不一致的复杂内容的通用理解来指导翻译。具体而言,DUAT对难翻译的词进行跨语言解释,并利用生成的解释来增强翻译。此外,我们重构了外部工具,以改进DUAT在检测难词和生成有用解释方面的能力。我们在自建的基准Challenge-WMT上进行了实验,该基准由容易发生误译的样本组成。在高资源与低资源语言对上的人工评估结果表明,DUAT显著促进了理解对齐,提升了翻译质量(COMET最高提升+3.85),并将翻译的直译性降低了25%至51%。

英文摘要:

Large Language models (LLMs) have exhibited remarkable abilities in understanding complex texts, offering a promising path towards human-like translation performance. However, this study reveals the misalignment between the translation-specific understanding and the general understanding inside LLMs. This understanding misalignment leads to LLMs mistakenly or literally translating some complicated concepts that they accurately comprehend in the general scenarios (e.g., QA). To align the translation-specific understanding to the general one, we propose a novel translation process, DUAT (Difficult words Understanding Aligned Translation), explicitly incorporating the general understanding on the complicated content incurring inconsistent understanding to guide the translation. Specifically, DUAT performs cross-lingual interpretation for the difficult-to-translate words and enhances the translation with the generated interpretations. Furthermore, we reframe the external tools to improve DUAT in detecting difficult words and generating helpful interpretations. We conduct experiments on the self-constructed benchmark Challenge-WMT, consisting of samples that are prone to mistranslation. Human evaluation results on high-resource and low-resource language pairs indicate that DUAT significantly facilitates the understanding alignment, which improves the translation quality (up to +3.85 COMET) and reduces the literality of the translation by -25% to -51%.

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。

↑