Kyoto University & JST Trilingual Technical Term Dictionary (TriTechDict) †

Description †

TriTechDict is a large-scale Japanese-English-Chinese technical term dictionary, containing about 3.3M entries. The original entries are from a Japanese-English dictionary provided by the Japan Science and Technology Agency (JST) (550k entries), and an English-Chinese dictionary provided by the Institute of Science and Technology information of China (ISTIC) (3.4M entries). We automatically translated the Japanese and Chinese entries in the original dictionaries into Chinese and Japanese, respectively, using a pivot based statistical machine translation system. Manual evaluation indicates that the accuracy of the constructed dictionary is 90%. The size of the constructed trilingual dictionary is about 3.3M after merging overlapped term pairs in the two translated dictionaries.

We release TriTechDict as a search interface, where users can judge the correctness of the entries, and enter more appropriate translations. We hope that the quality and coverage of the dictionary could be further improved by collective intelligence.

This work is supported by the JST MT project "Project on Practical Implementation of Japanese to Chinese-Chinese to Japanese Machine Translation."

Search Interface †

TriTechDict

Reference †

• Raj Dabre, Chenhui Chu, Fabien Cromieres, Toshiaki Nakazawa and Sadao Kurohashi.
  Large-scale Dictionary Construction via Pivot-based Statistical Machine Translation with Significance Pruning and Neural Network Features,
  Proceedings of the 29th Pacific Asia Conference on Language, Information and Computing (PACLIC2015), pp.293-301, Shanghai, China, (2015.10)

Contact and Bug Report †

• MAIL: nl-resource at nlp.ist.i.kyoto-u.ac.jp