CharSpan：利用词汇相似性实现极低资源语言的零样本机器翻译
CharSpan: Utilizing Lexical Similarity to Enable Zero-Shot Machine Translation for Extremely Low-resource Languages
AI总结：
本文提出CharSpan方法，通过字符跨度噪声增强利用高资源语言与极低资源语言的词汇相似性，实现零样本机器翻译，在多个语系上达到最先进性能。
摘要：
我们通过利用来自“紧密相关”的高资源语言（HRL）的跨语言迁移，解决了从极低资源语言（ELRL）到英语的机器翻译（MT）任务。为ELRL开发MT系统具有挑战性，因为这些语言通常缺乏平行语料库和单语语料库，并且它们的表示在大型多语言语言模型中缺失。许多ELRL与某些HRL共享词汇相似性，这提供了一个新颖的建模机会。然而，现有的基于子词的神经MT模型并未显式利用这种词汇相似性，因为它们仅隐式地对齐HRL和ELRL的潜在嵌入空间。为了克服这一限制，我们提出了一种新颖的方法CharSpan，基于在HRL训练数据中引入“字符跨度噪声增强”。这作为一种正则化技术，使模型对HRL和ELRL之间的“词汇差异”更加鲁棒，从而促进有效的跨语言迁移。在来自三个不同语系的紧密相关HRL和ELRL对上的零样本设置中，我们的方法显著优于强基线，成为ELRL的最先进模型。
英文摘要：
We address the task of machine translation (MT) from extremely low-resource language (ELRL) to English by leveraging cross-lingual transfer from 'closely-related' high-resource language (HRL). The development of an MT system for ELRL is challenging because these languages typically lack parallel corpora and monolingual corpora, and their representations are absent from large multilingual language models. Many ELRLs share lexical similarities with some HRLs, which presents a novel modeling opportunity. However, existing subword-based neural MT models do not explicitly harness this lexical similarity, as they only implicitly align HRL and ELRL latent embedding space. To overcome this limitation, we propose a novel, CharSpan, approach based on 'character-span noise augmentation' into the training data of HRL. This serves as a regularization technique, making the model more robust to 'lexical divergences' between the HRL and ELRL, thus facilitating effective cross-lingual transfer. Our method significantly outperformed strong baselines in zero-shot settings on closely related HRL and ELRL pairs from three diverse language families, emerging as the state-of-the-art model for ELRLs.
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。
