MultiMWE: Building a Multi-lingual Multi-Word Expression (MWE) Parallel Corpora

Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020), pages 2970–2979 Marseille, 11–16 May 2020 c⃝ European Language Resources Association (ELRA), licensed under CC-BY-NC 2970 MultiMWE: Building a Multi-lingual Multi-Word Expression (MWE) Parallel Corpora Lifeng Han1, Gareth J.F. Jones1 and Alan F. Smeaton2 1 ADAPT Research Centre 2 Insight Centre for Data Analytics School of Computing, Dublin City University, Glasnevin, Dublin 9, Ireland lifeng.han@adaptcentre.ie; {gareth.jones, alan.smeaton}@dcu.ie Abstract Multi-word expressions (MWEs) are a hot topic in research in natural language processing (NLP), including topics such as MWE detection, MWE decomposition, and research investigating the exploitation of MWEs in other NLP fields such as Machine Translation. However, the availability of bilingual or multi-lingual MWE corpora is very limited. The only bilingual MWE corpora that we are aware of is from the PARSEME (PARSing and Multi-word Expressions) EU Project. This is a small collection of only 871 pairs of English-German MWEs. In this paper, we present multi-lingual and bilingual MWE corpora that we have extracted from root parallel corpora. Our collections are 3,159,226 and 143,042 bilingual MWE pairs for German-English and Chinese-English respectively after filtering. We examine the quality of these extracted bilingual MWEs in MT experiments. Our initial experiments applying MWEs in MT show improved translation performances on MWE terms in qualitative analysis and better general evaluation scores in quantitative analysis, on both German-English and Chinese-English language pairs. We follow a standard experimental pipeline to create our MultiMWE corpora which are available online. Researchers can use this free corpus for their own models or use them in a knowledge base as model features. Keywords: Multi-lingual Corpus, Multi-word Expression, Machine Translation, Language Resource, Evaluation 1. Introduction The use of multi-word expressions (MWEs) has become a hot topic in research in the field of natural language pro- cessing (NLP). Topics of interests in MWEs include is- sues such as MWE detection (Maldonado et al., 2017), MWE decomposition, and the integration of MWEs into other NLP applications such as Machine Translation (MT). However, to support research into the multilingual use of MWEs, the availability of bilingual or multi-lingual MWE corpora is very limited. The only freely available bilingual MWE corpora that we are aware of, at the submission time, is from the PARSEME (PARSing and Multi-word Expres- sions) EU Project1. This corpus is quite small, contain- ing only 871 pairs of English-German MWEs. In this pa- per we present details of multi-lingual and bilingual MWE corpora that we have extracted from parallel corpora. We examine the quality of these extracted bilingual MWEs in MT experiments. Results for our initial experiments of ap- plying MWEs into the MT process show improved trans- lation evaluation scores on German-English and Chinese- English language pairs. These initial results justify further development of MWEs and their use in MT and potentially other NLP applications. We follow a standard experimental pipeline (Rikters and Bojar, 2017) to extract our bilingual MWEs. Our MultiMWE corpora are freely available on- line2. This paper is organized as follows: Section 2 provides some …

Content truncated to 500 words