Large-Scale English-Japanese Simultaneous Interpretation Corpus: Construction and Analyses with Sentence-Aligned Data

The evaluation campaign of the International Conference on Spoken Language Translation (IWSLT 2021) featured this year four shared tasks: (i) Simultaneous speech translation, (ii) Offline speech translation, (iii) Multilingual speech translation, (iv) Low-resource speech translation. A total of 22 teams participated in at least one of the tasks. This paper describes each shared task, data and evaluation metrics, and reports results of the received submissions.

show abstract

Findings of the Iwslt 2021 Evaluation Campaign

Anastasopoulos

Bojar

Bremerman³

et al. 2021

Proceedings of the 18th International Conference on Spoken Language Translation (IWSLT 2021)

Self Cite

View full text Add to dashboard Cite

show abstract

WCC-JC 2.0: A Web-Crawled and Manually Aligned Parallel Corpus for Japanese-Chinese Neural Machine Translation

Tian¹,

Mao

Han

et al. 2023

Electronics

View full text Add to dashboard Cite

Movie and TV subtitles are frequently employed in natural language processing (NLP)applications, but there are limited Japanese-Chinese bilingual corpora accessible as a dataset to trainneural machine translation (NMT) models. In our previous study, we effectively constructed a corpusof a considerable size containing bilingual text data in both Japanese and Chinese by collectingsubtitle text data from websites that host movies and television series. The unsatisfactory translationperformance of the initial corpus, Web-Crawled Corpus of Japanese and Chinese (WCC-JC 1.0), waspredominantly caused by the limited number of sentence pairs. To address this shortcoming, wethoroughly analyzed the issues associated with the construction of WCC-JC 1.0 and constructed theWCC-JC 2.0 corpus by first collecting subtitle data from movie and TV series websites. Then, wemanually aligned a large number of high-quality sentence pairs. Our efforts resulted in a new corpusthat includes about 1.4 million sentence pairs, an 87% increase compared with WCC-JC 1.0. As aresult, WCC-JC 2.0 is now among the largest publicly available Japanese-Chinese bilingual corporain the world. To assess the performance of WCC-JC 2.0, we calculated the BLEU scores relative toother comparative corpora and performed manual evaluations of the translation results generated bytranslation models trained on WCC-JC 2.0. We provide WCC-JC 2.0 as a free download for researchpurposes only.

show abstract

Large-Scale English-Japanese Simultaneous Interpretation Corpus: Construction and Analyses with Sentence-Aligned Data

Cited by 2 publications

References 24 publications

Findings of the Iwslt 2021 Evaluation Campaign

Findings of the Iwslt 2021 Evaluation Campaign

WCC-JC 2.0: A Web-Crawled and Manually Aligned Parallel Corpus for Japanese-Chinese Neural Machine Translation

Contact Info

Product

Resources

About