Multilingual Spoken Language Corpus Development for Communication Research

Takezawa, Toshiyuki

doi:10.1007/11939993_78

Cited by 26 publications

(34 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Dima et al (2018) presented a small dependency treebank of travel domain sentences in Modern Standard Arabic. The corpus is created by translating the selected 2,000 sentences from the Basic Travelling Expression Corpus (BTEC) presented by Takezawa in Takezawa (2006). Different parallel corpora were proposed for MSA such as those presented by: 1) Ziemski et al (2016) dealing, in addition to MSA, with five other languages: Chinese, English, French, Russian, Spanish.…”

Section: Building Resourcesmentioning

confidence: 99%

Arabic natural language processing: An overview

Guellil

Saadane²,

Azouaou

et al. 2021

Journal of King Saud University - Computer and Information Scie

100

View full text Add to dashboard Cite

a b s t r a c tArabic is recognised as the 4th most used language of the Internet. Arabic has three main varieties: (1) classical Arabic (CA), (2) Modern Standard Arabic (MSA), (3) Arabic Dialect (AD). MSA and AD could be written either in Arabic or in Roman script (Arabizi), which corresponds to Arabic written with Latin letters, numerals and punctuation. Due to the complexity of this language and the number of corresponding challenges for NLP, many surveys have been conducted, in order to synthesise the work done on Arabic. However these surveys principally focus on two varieties of Arabic (MSA and AD, written in Arabic letters only), they are slightly old (no such survey since 2015) and therefore do not cover recent resources and tools. To bridge the gap, we propose a survey focusing on 90 recent research papers (74% of which were published after 2015). Our study presents and classifies the work done on the three varieties of Arabic, by concentrating on both Arabic and Arabizi, and associates each work to its publicly available resources whenever available.

show abstract

Section: Building Resourcesmentioning

confidence: 99%

Arabic natural language processing: An overview

Guellil

Saadane²,

Azouaou

et al. 2021

Journal of King Saud University - Computer and Information Scie

100

View full text Add to dashboard Cite

show abstract

“…In previous studies, BTEC has been evaluated as the basic dataset for S2ST purpose [1,4]. Similarly, Bengali translated BTEC also needs to be analyzed in view of the following important points:…”

Section: O Utline Of a Nalysismentioning

confidence: 99%

Bengali Basic Travel Expression Corpus: A statistical analysis

Khan

Basu

et al. 2014

2014 17th Oriental Chapter of the International Committee for the Co-Ordination and Standardization of Speech Databases and Ass

View full text Add to dashboard Cite

The Japanese-English aligned Basic Travel Expression Corpus (BTEC) has been used as a basic dataset for development of real-world Speech-to-Speech Translation (S2ST) systems in related prior studies. This paper presents a detailed statistical analysis on the Bengali translated BTEC text and its phonetic transcriptions for development of English-Bengali speech translation applications in travel domain. In different level of analysis hierarchy, the study focuses on the lexical and phonetical status of the analyzed corpus based on frequency spectrums, estimated population size, coverage ratio, goodness of fit of Large Number of Rare Events (LNRE) model and transition patterns . The experimental observations provide necessary insights on sufficiency of the analyzed corpus with respect to the travel domain as well as for building basic components of English-Bengali S2ST system.The Universal Speech Translation Advanced Research (U ST AR) Consortium is an international research consortium comprised of 29 institutes from 23 countries (as of June, 2014) across the world. It provides a uniform framework for pursuing collaborative research; sharing resources and developing infrastructures and resources for building web based speech-to speech translation (S2ST) applications in English to native languages of the respective participating institutes. Under the same framework, since 2014, Centre for Development of Advanced Computing, Kolkata (in India) is developing resources and components required for English to Bengali speech translation focusing on the travel domain. As the prime common dataset to build the language specific components of cross language S2ST applications under U-ST AR framework, the Basic Travel Expression Corpus (BTEC) has been used and translated into several native languages by the respective member institutes.As reported in earlier studies, the BTEC, developed by Advanced Telecommunication Research (ATR) laboratory, Japan, is an English-Japanese aligned corpus [1,2]. It is qualitatively well-examined, consistent and having wide coverage of travel and tourism related basic expressions. These expressions are highly oriented towards real world conversations on various topics like transportation, travel activities, sight-seeing, shopping, dining, staying, asking help in trouble, basic queries, greetings etc. much likely to be spoken by tourists travelling to other countries. In prior works, the corpus has been quantitatively evaluated and statistically analyzed in languages like Japanese [2] and Hindi [3]. The statistical analysis presented in this paper, is carried out on the Bengali translated version of the BTEC and is being reported for the first time. To introduce Bengali language, it is good to mention here that Bengali belongs to the eastern Indo-Aryan branch of the Indo-European family of languages. It is the official state language of East Indian state West Bengal and the national language of Bangladesh. As per Wiki, it is one of the most spoken languages, ranked seventh in world. II. PREPARING B EN...

show abstract

“…One of the main contributions of this work is using the neural MT approach for the Chinese-Spanish language pair. In the last years, there has appeared more and more resources for this language pair available in [Ziemski et al, 2016a] or from TAUS [Takezawa, 2006]. The TAUS corpus is around 2,890,000 sentences, the Bible corpus about 30,000 sentences and the BTEC corpus about 20,000 sentences.…”

Section: Data and Preprocessingmentioning

confidence: 99%

Chinese–Spanish neural machine translation enhanced with character and word bitmap fonts

Costa-jussà

Aldón

Fonollosa

2017

Machine Translation

View full text Add to dashboard Cite

Recently, machine translation systems based on neural networks have reached state-of-the-art results for some pairs of languages (e.g., German–English). In this paper, we are investigating the performance of neural machine translation in Chinese–Spanish, which is a challenging language pair. Given that the meaning of a Chinese word can be related to its graphical representation, this work aims to enhance neural machine translation by using as input a combination of: words or characters and their corresponding bitmap fonts. The fact of performing the interpretation of every word or character as a bitmap font generates more informed vectorial representations. Best results are obtained when using words plus their bitmap fonts obtaining an improvement (over a competitive neural MT baseline system) of almost six BLEU, five METEOR points and ranked coherently better in the human evaluation.Peer ReviewedPostprint (published version

show abstract

Multilingual Spoken Language Corpus Development for Communication Research

Cited by 26 publications

References 7 publications

Arabic natural language processing: An overview

Arabic natural language processing: An overview

Bengali Basic Travel Expression Corpus: A statistical analysis

Chinese–Spanish neural machine translation enhanced with character and word bitmap fonts

Contact Info

Product

Resources

About