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 2203.01976cs.CL
收藏

基于重叠的词汇生成促进相关语言间的跨语言迁移

Overlap-based Vocabulary Generation Improves Cross-lingual Transfer Among Related Languages

Indian Institute of Technology Bombay(印度理工学院孟买分校);Google Research(谷歌研究院)

Vaidehi Patil, Partha Talukdar, Sunita Sarawagi

摘要:

预先训练的多语言模型,如mBERT和XLM-R,已经证明了zero-shot跨语言迁移到低网络资源语言(LRL)的巨大潜力。然而,由于有限的模型容量,高网络资源语言(HRL)和LRL之间的可用单语语料库的大小的巨大差异并没有提供足够的范围的共同嵌入LRL与HRL,从而影响下游的LRL的任务性能。在本文中,我们认为,沿着词汇重叠的维度在一个语系中的语言之间的相关性可能会被利用来克服一些语料库的限制LRL。我们提出了重叠BPE(OBPE),一个简单而有效的修改BPE词汇生成算法,提高了相关语言的重叠。通过对多个NLP任务和数据集的广泛实验,我们观察到OBPE生成了一个词汇表,通过与HRL共享的标记来增加LRL的表示。这导致从相关HRL到LRL的改进的zero-shot转移,而不降低HRL表示和准确性。与以前的研究,驳回了标记重叠的重要性,我们表明,在低资源相关的语言设置,标记重叠的事项。综合地将重叠减少到零可能导致zero-shot传输精度下降多达四倍。

英文摘要:

Pre-trained multilingual language models such as mBERT and XLM-R have demonstrated great potential for zero-shot cross-lingual transfer to low web-resource languages (LRL). However, due to limited model capacity, the large difference in the sizes of available monolingual corpora between high web-resource languages (HRL) and LRLs does not provide enough scope of co-embedding the LRL with the HRL, thereby affecting downstream task performance of LRLs. In this paper, we argue that relatedness among languages in a language family along the dimension of lexical overlap may be leveraged to overcome some of the corpora limitations of LRLs. We propose Overlap BPE (OBPE), a simple yet effective modification to the BPE vocabulary generation algorithm which enhances overlap across related languages. Through extensive experiments on multiple NLP tasks and datasets, we observe that OBPE generates a vocabulary that increases the representation of LRLs via tokens shared with HRLs. This results in improved zero-shot transfer from related HRLs to LRLs without reducing HRL representation and accuracy. Unlike previous studies that dismissed the importance of token-overlap, we show that in the low-resource related language setting, token overlap matters. Synthetically reducing the overlap to zero can cause as much as a four-fold drop in zero-shot transfer accuracy.

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。

↑