---
title: "Looking for Candidate Translational Equivalents in Specialized,"
description: "The paper looks at how to find French–English translation candidates without parallel text… 33% recall. 16 references, cited by 62 papers here."
authors: ["Yun-Chuang Chiao", "Pierre Zweigenbaum"]
year: 2002
venue: "COLING 2002: The 17th International Conference on Computational Linguistics: Project Notes"
source: "https://aclanthology.org/C02-2020/"
licence: "CC BY-NC-SA 3.0 (ACL Anthology, before 2016)"
licence_url: "https://creativecommons.org/licenses/by-nc-sa/3.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/chiao-2002-looking-candidate-translational-equivalents-specialized/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/chiao-2002-looking-candidate-translational-equivalents-specialized/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2002](https://md.research.elamur.ai/free/en/papers/2002/) / Looking for Candidate Translational Equivalents in Specialized, Comparable Corpora

# Looking for Candidate Translational Equivalents in Specialized, Comparable Corpora

**Authors:** Yun-Chuang Chiao and [Pierre Zweigenbaum](https://md.research.elamur.ai/free/en/authors/pierre-zweigenbaum/) · **Year:** 2002 · **Venue:** [COLING 2002: The 17th International Conference on Computational Linguistics: Project Notes](https://md.research.elamur.ai/free/en/venues/coling/) ([COLING 2002](https://md.research.elamur.ai/free/en/venues/coling/2002/)) · **Licence:** [CC BY-NC-SA 3.0 (ACL Anthology, before 2016)](https://creativecommons.org/licenses/by-nc-sa/3.0/)

**Links:** [Publisher's page](https://aclanthology.org/C02-2020/) · [PDF](https://aclanthology.org/C02-2020.pdf) · [OpenAlex](https://openalex.org/W2011248508)

**Citations:** 118 citations (OpenAlex, 2026-10-14) · 62 in our library, of the 141,840 papers whose bibliographies we read

## Summary
The paper looks at how to find French–English translation candidates without parallel text, in medicine, where corpora are small and specialized. Earlier work on comparable corpora relied on very large general-language collections. Here, the authors built two medical web corpora of roughly 600k words each. Pages were selected through MeSH-indexed catalogs: CISMeF for French and CliniWeb for English, both restricted to the 'Signs and Symptoms' (C23) subtree. They also assembled a bilingual medical lexicon from the Masson dictionary and UMLS resources. Each word gets a context vector from a small co-occurrence window. The vector is mapped into the other language through the lexicon's seed words, then compared with native target vectors. Three weightings and two similarity measures were tested. On 95 frequent test words, the correct translation came first for about 20% of words and appeared in the top 10 for about half. A reverse-translation filter raised top-candidate precision to 74% at 33% recall. The results suggest that modest domain corpora can help extend specialized lexicons, for example for cross-language retrieval.

## Key points
- Corpora: a French corpus of 602,484 words (41,295 unique), drawn from 2,338 CISMeF URLs plus the pages they link to. An English corpus of 608,320 words (32,919 unique), from 921 CliniWeb URLs.
- Lexicon: 18,437 entries combining the Dictionnaire Médical Masson with the French and English versions of MeSH, WHOART and ICPC from UMLS. After intersecting with both corpora, 4,963 cross-language pivot words remain.
- Method: the context window spans ±3 words, applied after stop-word removal and light stemming (the S-stemmer for English). The weightings are raw counts, tf:idf and log likelihood, and the similarity measures are Jaccard and cosine.
- Results: Jaccard with raw counts places about 20% of correct translations at rank 1. tf:idf combined with either measure puts more than 60% within the top 20. Log likelihood gave little benefit.
- Reverse filtering keeps 42 of the 95 source words and reaches 0.69 precision at 0.31 recall. Counting derived adjectives as correct (e.g., estomac/gastric) raises this to 0.74 precision at 0.33 recall.

## Interesting ideas
- **Corpus building via thesaurus-indexed catalogs**: the authors picked documents that both catalogs indexed under the same MeSH subtree, which gives topically matched comparable corpora cheaply. The same approach could work wherever bilingual catalogs share a controlled vocabulary.
- **Shared pivot space**: context dimensions are limited to target words that occur in the target corpus and are in the lexicon, and whose source equivalent occurs in the source corpus. This keeps transferred and native vectors directly comparable and filters out noise.
- **Context vectors as documents**: treating the concatenated windows around a word as a pseudo-document lets standard IR weightings such as tf:idf be reused for lexical similarity.
- **Round-trip consistency filter**: a candidate is kept only if translating it back ranks the original word in its top 10. This trades recall for much higher precision, which is a generic check for any bilingual lexicon induction method.
- **Controlled leave-one-out evaluation**: the authors test on known lexicon pairs and report the whole distribution of percentile ranks rather than a single accuracy figure. This shows how rankings behave at different cut-offs.

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

## Institutions

- [Assistance Publique – Hôpitaux de Paris](https://md.research.elamur.ai/free/en/hospitals/assistance-publique-hopitaux-de-paris/) — Hospital, Paris, France · 2 works · 180 citations

📚 Shelves (4): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 9,085 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 8,888 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 9,296 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 9,206 works

## Cite

Cite as: Yun-Chuang Chiao and Pierre Zweigenbaum. 2002. Looking for Candidate Translational Equivalents in Specialized, Comparable Corpora. COLING 2002: The 17th International Conference on Computational Linguistics: Project Notes.

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

```bibtex
@inproceedings{chiao2002looking,
  title = {{Looking for Candidate Translational Equivalents in Specialized, Comparable Corpora}},
  author = {Yun-Chuang Chiao and Pierre Zweigenbaum},
  booktitle = {COLING 2002: The 17th International Conference on Computational Linguistics: Project Notes},
  year = {2002},
  doi = {10.3115/1071884.1071904},
  url = {https://doi.org/10.3115/1071884.1071904}
}
```

- [← previous Floresta Sintá(c)tica: A treebank for Portuguese](https://md.research.elamur.ai/free/en/papers/afonso-2002-floresta-sinta-c-tica-treebank/)
- [next → Modeling English Past Tense Intuitions with Minimal Generalization](https://md.research.elamur.ai/free/en/papers/albright-2002-modeling-english-past-tense-intuitions/)
- [↑ back to papers of 2002](https://md.research.elamur.ai/free/en/papers/2002/)

## References

16 references, 1 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Jiang Chen and J-Y. Nie. 2000. Parallel web text mining for cross-language IR. In Proceedings of RIAO 2000: Content-Based Multimedia Information Access, volume 1, pages 62–78, Paris, France, April. C.I.D. [find](https://scholar.google.com/scholar?q=Jiang+Chen+and+J%2DY%2E+Nie%2E+2000%2E+Parallel+web+text+mining+for+cross%2Dlanguage+IR%2E+In+Proceedings+of+RIAO+2000%3A+Content%2DBased+Multimedia+Information+Access%2C+volume+1%2C+pages+62%E2%80%9378%2C+Paris%2C+France%2C+April%2E+C%2E)
2. The city-block metric is computed as the sum of the absolute differences of corresponding vectors positions. [find](https://scholar.google.com/scholar?q=The+city%2Dblock+metric+is+computed+as+the+sum+of+the+absolute+differences+of+corresponding+vectors+positions%2E)
3. Pascale Fung and Kathleen McKeown. 1997. Finding terminology translations from non-paralle corpora. In Proceedings of the 5th Annual Workshop on Very Large Copora, volume 1, pages 192–202, Hong Kong. [find](https://scholar.google.com/scholar?q=Pascale+Fung+and+Kathleen+McKeown%2E+1997%2E+Finding+terminology+translations+from+non%2Dparalle+corpora%2E+In+Proceedings+of+the+5th+Annual+Workshop+on+Very+Large+Copora%2C+volume+1%2C+pages+192%E2%80%93202%2C+Hong+Kong%2E)
4. Pascale Fung and L. Y. Yee. 1998. An IR approach for translating new words from non-parallel, comparable texts. In Proceedings of the 36 th ACL, pages 414–420, Montréal, August. [find](https://scholar.google.com/scholar?q=Pascale+Fung+and+L%2E+Y%2E+Yee%2E+1998%2E+An+IR+approach+for+translating+new+words+from+non%2Dparallel%2C+comparable+texts%2E+In+Proceedings+of+the+36+th+ACL%2C+pages+414%E2%80%93420%2C+Montr%C3%A9al%2C+August%2E)
5. Gregory Grefenstette. 1998a. Cross-Language Information Retrieval. Kluwer Academic Publishers, London. [find](https://scholar.google.com/scholar?q=Gregory+Grefenstette%2E+1998a%2E+Cross%2DLanguage+Information+Retrieval%2E+Kluwer+Academic+Publishers%2C+London%2E)
6. Gregory Grefenstette. 1998b. The problem of cross-language information retrieval. In Cross-Language Information Retrieval (Grefenstette, 1998a), pages 1–9. [find](https://scholar.google.com/scholar?q=Gregory+Grefenstette%2E+1998b%2E+The+problem+of+cross%2Dlanguage+information+retrieval%2E+In+Cross%2DLanguage+Information+Retrieval+%28Grefenstette%2C+1998a%29%2C+pages+1%E2%80%939%2E)
7. D. Harman. 1991. How effective is suffixing. Journal of the American Society for Information Science, 42:7–15. [find](https://scholar.google.com/scholar?q=D%2E+Harman%2E+1991%2E+How+effective+is+suffixing%2E+Journal+of+the+American+Society+for+Information+Science%2C+42%3A7%E2%80%9315%2E)
8. D. Hiemstra, F. de Jong, and W. Kraaij. 1997. A domain specific lexicon acquisition tool for cross-linguage information retrieval. In Proceedings of RIAO97, pages 217–232, Montreal, Canada. [find](https://scholar.google.com/scholar?q=D%2E+Hiemstra%2C+F%2E+de+Jong%2C+and+W%2E+Kraaij%2E+1997%2E+A+domain+specific+lexicon+acquisition+tool+for+cross%2Dlinguage+information+retrieval%2E+In+Proceedings+of+RIAO97%2C+pages+217%E2%80%93232%2C+Montreal%2C+Canada%2E)
9. J. Laffling. 1992. On constructiong a transfer dictionary for man and machine. Target, 4(1):17–31. [find](https://scholar.google.com/scholar?q=J%2E+Laffling%2E+1992%2E+On+constructiong+a+transfer+dictionary+for+man+and+machine%2E+Target%2C+4%281%29%3A17%E2%80%9331%2E)
10. M.L. Littman, S.T. Dumais, and T.K. Landauer. 1998. Automatic cross-language information retrieval using latent semantic indexing. In Grefenstette (Grefenstette, 1998a), chapter 5, pages 51– 62. [find](https://scholar.google.com/scholar?q=M%2EL%2E+Littman%2C+S%2ET%2E+Dumais%2C+and+T%2EK%2E+Landauer%2E+1998%2E+Automatic+cross%2Dlanguage+information+retrieval+using+latent+semantic+indexing%2E+In+Grefenstette+%28Grefenstette%2C+1998a%29%2C+chapter+5%2C+pages+51%E2%80%93+62%2E)
11. Robert M. Losee. 1998. Text Retrieval and Filtering: Analytic Models of Performance, volume 3 of Information Retrieval. Kluwer Academic Publishers, Dordrecht & Boston. [find](https://scholar.google.com/scholar?q=Robert+M%2E+Losee%2E+1998%2E+Text+Retrieval+and+Filtering%3A+Analytic+Models+of+Performance%2C+volume+3+of+Information+Retrieval%2E+Kluwer+Academic+Publishers%2C+Dordrecht+%26+Boston%2E)
12. E. Picchi and C. Peters. 1998. Cross-language information retrieval: A system for comparable corpus querying. In Grefenstette (Grefenstette, 1998a), chapter 7, pages 81–90. [find](https://scholar.google.com/scholar?q=E%2E+Picchi+and+C%2E+Peters%2E+1998%2E+Cross%2Dlanguage+information+retrieval%3A+A+system+for+comparable+corpus+querying%2E+In+Grefenstette+%28Grefenstette%2C+1998a%29%2C+chapter+7%2C+pages+81%E2%80%9390%2E)
13. Reinhard Rapp. 1999. [Automatic identification of word translations from unrelated English and German corpora](https://md.research.elamur.ai/free/en/papers/rapp-1999-automatic-identification-word-translations-unrelated/). In Proceedings of the 37 th ACL, College Park, Maryland, June.
14. H. Charles Romesburg. 1990. Cluster Analysis for Researchers. Krieger, Malabar, FL. [find](https://scholar.google.com/scholar?q=H%2E+Charles+Romesburg%2E+1990%2E+Cluster+Analysis+for+Researchers%2E+Krieger%2C+Malabar%2C+FL%2E)
15. Gerald Salton. 1970. Automatic processing of foreign language documents. Journal of the American Society for Information Science, 21(3):187– 194. [find](https://scholar.google.com/scholar?q=Gerald+Salton%2E+1970%2E+Automatic+processing+of+foreign+language+documents%2E+Journal+of+the+American+Society+for+Information+Science%2C+21%283%29%3A187%E2%80%93+194%2E)
16. Karen Sparck Jones. 1979. Experiments in relevance weighting of search terms. Information Processing and Management, 15:133–144. [find](https://scholar.google.com/scholar?q=Karen+Sparck+Jones%2E+1979%2E+Experiments+in+relevance+weighting+of+search+terms%2E+Information+Processing+and+Management%2C+15%3A133%E2%80%93144%2E)

*Updated 24 October 2026*