Elamur Research⌕
☰
home / papers / 2008 / Automatic Construction of a Japanese-Chinese Dictionary via English

Automatic Construction of a Japanese-Chinese Dictionary via English

Summary

The authors seek a way to obtain a dictionary for a language pair that lacks a complete direct bilingual dictionary. To this end, they combine two dictionaries that share a pivot language, in this case English. Simply composing a Japanese-English and a Chinese-English dictionary produces many false equivalents. This is caused by the polysemy of English words: for example, "plant" means both "factory" and "plant (vegetation)." To filter out such equivalents, the method uses non-parallel monolingual corpora of both languages on a shared topic. Word associations are extracted from the corpora using pointwise mutual information. Japanese associations are then aligned with Chinese ones through the "noisy" dictionary. For each input word, a correlation matrix between the associated words and the translation candidates is computed iteratively. The authors had previously developed this algorithm for unsupervised word sense discrimination. A candidate is retained if it is supported by a sufficient proportion of associated words. The result is a dictionary adapted to the topic of the corpora. Each translation in it is accompanied by context words, so a translation can be chosen depending on the context. Preliminary experiments on nouns showed that the approach works, but its parameters still need to be optimized. The method is useful for low-resource language pairs.

Key points

  • Resources: the LDC Chinese-English Translation Lexicon together with the EDR Japanese-English Bilingual Dictionary. Corpora: Mainichi (2000–2005, 632 MB) and Xinhua News Agency (1999–2004, 473 MB). For segmentation, JUMAN and the Chinese morphological analyzer of Nakagawa and Uchimoto (2007) were used.
  • Settings: Japanese nouns with frequency ≥100, Chinese nouns with frequency ≥30. The co-occurrence window is 25 content words on each side, and the MI threshold is 1.0. At most 700 associated words were used per matrix.
  • Coverage: at least two translation candidates were available for 8,284 of 10,003 Japanese nouns and for 8,426 of 9,288 Chinese nouns.
  • Selection by "support": the support of a candidate is the proportion of associated words for which that candidate has the highest correlation. In the example with the word 工場, a threshold of 0.1 removed the false 植株 and 作品, but also the correct 厂 along with them.
  • Bidirectional selection: the intersection of the J→C and C→J results is taken at a minimum support of 0.025. This raised precision by about 4%, while recall remained almost unchanged. Quality is limited by the coverage of the source dictionaries: after the candidates were manually supplemented, the results for the examples improved.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

Fields: Linguistics & languages

Cite

Cite as Hiroyuki Kaji, Shin'ichi Tamamura and Dashtseren Erdenebat. 2008. Automatic Construction of a Japanese-Chinese Dictionary via English. Proceedings of the Sixth International Conference on Language Resources and Evaluation (LREC'08).

The paper itself, not this page. BibTeX from the library's record of the work.

@inproceedings{kaji2008automatic,
  title = {{Automatic Construction of a Japanese-Chinese Dictionary via English}},
  author = {Hiroyuki Kaji, Shin'ichi Tamamura and Dashtseren Erdenebat},
  booktitle = {Proceedings of the Sixth International Conference on Language Resources and Evaluation (LREC'08)},
  year = {2008},
  url = {https://aclanthology.org/L08-1573/}
}

References

14 references, 5 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Bond, Francis, Ruhaida Binti Sulong, Takefumi Yamazaki, and Kentaro Ogura. (2001). Design and construction of a machine-tractable Japanese-Malay dictionary. In Proceedings of Machine Translation Summit VIII, pp. 53-58. find
  2. Church, Kenneth and Patrick Hanks. (1990). Word association norms, mutual information, and lexicography. Computational Linguistics, 16(1): 22-29. our card
  3. Fung, Pascale and Lo Yuen Yee. (1998). An IR approach for translating new words from nonparallel, comparable texts. In Proceedings of the 36th Annual Meeting of the Association for Computational Linguistics and the 17th International Conference on Computational Linguistics, pp. 414-420. our card
  4. Kaji, Hiroyuki and Toshiko Aizono. (1996). Extracting word correspondences from bilingual corpora based on word co-occurrence information. In Proceedings of the 16th International Conference on Computational Linguistics, pp. 23-28. find
  5. Kaji, Hiroyuki and Yasutsugu Morimoto. (2002). Unsupervised word sense disambiguation using bilingual comparable corpora. In Proceedings of the 19th International Conference on Computational Linguistics, pp. 411- 417. find
  6. Kaji, Hiroyuki and Yasutsugu Morimoto. (2005). Unsupervised word-sense disambiguation using bilingual comparable corpora. IEICE Transactions on Information and Systems, E88-D(2): 289-301. find
  7. Nakagawa, Tetsuji and Kiyotaka Uchimoto. (2007). A hybrid approach to word segmentation and POS tagging, In Proceedings of the 45th Annual Meeting of the Association for Computational Linguistics, pp. 217-220. find
  8. Rapp, Reinhard. (1995). Identifying word translations in non-parallel texts. In Proceedings of the 33rd Annual Meeting of the Association for Computational Linguistics, pp. 320-322. our card
  9. Rapp, Reinhard. (1999). Automatic identification of word translations from unrelated English and German corpora. In Proceedings of the 37th Annual Meeting of the Association for Computational Linguistics, pp. 519-526. our card
  10. Sammer, Marcus and Stephen Soderland. (2007). Building a sense-distinguished multilingual lexicon from monolingual corpora and bilingual lexicons. In Proceedings of Machine Translation Summit XI, pp. 399-406. find
  11. Shirai, Satoshi and Kazuhide Yamamoto. (2001). Linking English words in two bilingual dictionaries to generate another language pair dictionary. In Proceedings of the 19th International Conference on Computer Processing of Oriental Languages, pp. 174-179. find
  12. Tanaka, Kumiko and Hideya Iwasaki. (1996). Extraction of lexical translations from non-aligned corpora. In Proceedings of the 16th International Conference on Computational Linguistics, pp. 580-585. find
  13. Tanaka, Kumiko and Kyoji Umemura. (1994). Construction of a bilingual dictionary intermediated by a third language. In Proceedings of the 15th International Conference on Computational Linguistics, pp. 297-303. our card
  14. Zhang, Yujie, Qing Ma, and Hitoshi Isahara. (2007). Building Japanese-Chinese translation dictionary based on EDR Japanese-English bilingual dictionary. In Proceedings of Machine Translation Summit XI, pp. 551-557. find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/kaji-2008-automatic-construction-japanese-chinese-dictionary/