Languages Through the Looking Glass of BPE Compression

Author:

Gutierrez-Vasques Ximena1,Bentz Christian2,Samardžić Tanja3

Affiliation:

1. University of Zürich, URPP Language and Space. ximena.gutierrezvasques@uzh.ch

2. University of Tübingen, Department of General Linguistics. chris@christianbentz.de

3. University of Zürich, URPP Language and Space, tanja.samardzic@uzh.ch

Abstract

Abstract Byte-pair encoding (BPE) is widely used in NLP for performing subword tokenization. It uncovers redundant patterns for compressing the data, and hence alleviates the sparsity problem in downstream applications. Subwords discovered during the first merge operations tend to have the most substantial impact on the compression of texts. However, the structural underpinnings of this effect have not been analyzed cross-linguistically. We conduct in-depth analyses across 47 typologically diverse languages and three parallel corpora, and thereby show that the types of recurrent patterns that have the strongest impact on compression are an indicator of morphological typology. For languages with richer inflectional morphology there is a preference for highly productive subwords on the early merges, while for languages with less inflectional morphology, idiosyncratic subwords are more prominent. Both types of patterns contribute to efficient compression. Counter to the common perception that BPE subwords are not linguistically relevant, we find patterns across languages that resemble those described in traditional typology. We thus propose a novel way to characterize languages according to their BPE subword properties, inspired by the notion of morphological productivity in linguistics. This allows us to have language vectors that encode typological knowledge induced from raw text. Our approach is easily applicable to a wider range of languages and texts, as it does not require annotated data or any external linguistic knowledge. We discuss its potential contributions to quantitative typology and multilingual NLP.

Publisher

MIT Press

Subject

Artificial Intelligence,Computer Science Applications,Linguistics and Language,Language and Linguistics

Reference119 articles.

1. Quechua;Adelaar,2004

2. JW300: A wide-coverage parallel corpus for low-resource languages;Agić,2019

3. Typological distinctions in word-formation;Aikhenvald,2007

4. Polysynthetic structures of Lowland Amazonia;Aikhenvald,2017

5. Al-Rfou, Rami . 2015. Polyglot: A Massive Multilingual Natural Language Processing Pipeline. Ph.D. thesis, State University of New York at Stony Brook.

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3