TED-MWE: a bilingual parallel corpus with MWE annotation Towards a methodology for annotating MWEs in parallel multilingual corpora

نویسندگان

  • Johanna Monti
  • Federico Sangati
  • Mihael Arcan
چکیده

English. The translation of Multiword expressions (MWE) by Machine Translation (MT) represents a big challenge, and although MT has considerably improved in recent years, MWE mistranslations still occur very frequently. There is the need to develop large data sets, mainly parallel corpora, annotated with MWEs, since they are useful both for SMT training purposes and MWE translation quality evaluation. This paper describes a methodology to annotate a parallel spoken corpus with MWEs. The dataset used for this experiment is an English-Italian corpus extracted from the TED spoken corpus and complemented by an SMT output. Italiano. La traduzione delle polirematiche da parte dei sistemi di Traduzione Automatica (TA) rappresenta un sfida irrisolta e benché i sistemi abbiano compiuto notevoli progressi, traduzioni errate di polirematiche occorrono ancora molto di frequente. È necessario sviluppare ampie collezioni di dati principalmente corpora paralleli annotati con polirematiche che siano utili sia per l’addestramento della TA di tipo statistico sia per la valutazione della qualità della traduzione delle polirematiche. Questo contributo descrive una metodologia per annotare un corpus parallelo del parlato con le polirematiche e il corpus stesso. La collezione di dati usata per questo esperimento è un corpus inglese-italiano estratto dal TED, corpus del parlato, integrato dalla traduzione di un sistema statistico di TA. ∗Johanna Monti is author of sections 2 and 3.2, Federico Sangati is author of sections 4 and 5, Mihael Arcan is author of sections 3.1 and 4.

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Annotation of Multi-Word Expressions in Czech Texts

Multi-word expressions (MWEs) are difficult to define and also difficult to annotate. Some of them cause serious errors in the traditional annotation pipeline tokenization – morphological analysis – morphological disambiguation. Many cases of incorrect annotation in Czech corpora are known. To narrow the research topic, we focus only in fixed MWEs – those with fixed word order and no ellidable ...

متن کامل

Discovering Light Verb Constructions and their Translations from Parallel Corpora without Word Alignment

We propose a method for joint unsupervised discovery of multiword expressions (MWEs) and their translations from parallel corpora. First, we apply independent monolingual MWE extraction in source and target languages simultaneously. Then, we calculate translation probability, association score and distributional similarity of co-occurring pairs. Finally, we rank all translations of a given MWE ...

متن کامل

Querying Multi-word Expressions Annotation with CQL

This paper demonstrates a solution for querying corpora with multi-word expression (MWE) annotation using a concordance system. Namely, the PARSEME multilingual corpora, which contain manually annotated verbal multi-word expression (VMWE) in 18 languages, are converted to a suitable vertical format so that they can be explored using the Corpus Query Language (CQL). VMWEs encompass a range of ca...

متن کامل

Paying Attention to Multi-Word Expressions in Neural Machine Translation

Processing of multi-word expressions (MWEs) is a known problem for any natural language processing task. Even neural machine translation (NMT) struggles to overcome it. This paper presents results of experiments on investigating NMT attention allocation to the MWEs and improving automated translation of sentences that contain MWEs in English→Latvian and English→Czech NMT systems. Two improvemen...

متن کامل

Statistically-Driven Alignment-Based Multiword Expression Identification for Technical Domains

Multiword Expressions (MWEs) are one of the stumbling blocks for more precise Natural Language Processing (NLP) systems. Particularly, the lack of coverage of MWEs in resources can impact negatively on the performance of tasks and applications, and can lead to loss of information or communication errors. This is especially problematic in technical domains, where a significant portion of the voc...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2015