arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2402.13562cs.CL
收藏

跨语言迁移中多源语言训练的分析

Analysis of Multi-Source Language Training in Cross-Lingual Transfer

Hanyang University(汉阳大学);Sony AI(索尼人工智能)

Seong Hoon Lim, Taejun Yun, Jinhyeon Kim, Jihun Choi, Taeuk Kim

AI总结:

本文研究跨语言迁移(XLT)的内部机制,提出多源语言训练(MSLT)技术,发现其能增加不同语言嵌入空间的混合以利用语言无关信息,并提出启发式方法识别有效语言组合以提升性能。

摘要:

多语言语言模型(LMs)对特定语言-任务对的成功适配,关键在于是否有为该条件量身定制的数据。尽管跨语言迁移(XLT)方法有助于缓解这一数据稀缺问题,但其有效性背后的机制仍存在持续的争议。在本研究中,我们聚焦于关于 XLT 内部机制的一个有前景的假设,即它促使多语言 LMs 更加重视语言无关或任务特定的特征。我们通过考察 XLT 的模式如何随过程中涉及源语言数量的变化而改变,来验证这一假设。我们的实验发现,在 XLT 中使用多种源语言——我们称之为 Multi-Source Language Training(MSLT)的技术——会导致不同语言的嵌入空间混合程度增加,这支持了 XLT 受益于利用语言无关信息的观点。另一方面,我们发现使用任意源语言组合并不总是保证更好的性能。我们提出了用于识别 MSLT 有效语言组合的简单启发式方法,并实证证明了其有效性。

英文摘要:

The successful adaptation of multilingual language models (LMs) to a specific language-task pair critically depends on the availability of data tailored for that condition. While cross-lingual transfer (XLT) methods have contributed to addressing this data scarcity problem, there still exists ongoing debate about the mechanisms behind their effectiveness. In this work, we focus on one of promising assumptions about inner workings of XLT, that it encourages multilingual LMs to place greater emphasis on language-agnostic or task-specific features. We test this hypothesis by examining how the patterns of XLT change with a varying number of source languages involved in the process. Our experimental findings show that the use of multiple source languages in XLT-a technique we term Multi-Source Language Training (MSLT)-leads to increased mingling of embedding spaces for different languages, supporting the claim that XLT benefits from making use of language-independent information. On the other hand, we discover that using an arbitrary combination of source languages does not always guarantee better performance. We suggest simple heuristics for identifying effective language combinations for MSLT and empirically prove its effectiveness.

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。

↑