Skip to Main content Skip to Navigation
Theses

Identification et exploitation de contextes riches en connaissances pour l’aide à la traduction terminologique

Abstract : Computer-assisted translation and terminology management tools are often used to meet the needs in management of multilingual and monolingual writings. These tools facilitate the access to technical terms and expressions that are related to areas of specialty, and essential to any communication process. The understanding of technical terms can be potentiated by their “contextualization”. However, having access to a term or its translation is not enough, since it is also necessary to be able to use it properly and to understand its exact meaning. Thus, this contextualization is estabilished on two levels: in texts and in the terminology. In texts, the user must have access to information regarding the use of terms, namely linguistic knowledge-rich contexts. In the terminology, the user requires access to semantic or conceptual relationships between the terms to better understand its meaning, namely conceptual rich-knowlegde contexts. In the framework of this thesis, we proposed a strategy for extracting Knowledge-Rich Contexts (KRCs) to produce a new terminological dictionary. It is to provide, for each term and its possible translations, the KRCs in which it occurs. We continued our work in a bilingual phase part of specialized translation, under continuous revision. We propose a new generation of bilingual concordancers that take as input a term and its translation, and provides not parallel, but aligned Knowledge-Rich Contexts from specialized comparable corpora. The evaluation show that our concordancer can assist revisers despite the difficulty of the task.
Complete list of metadatas

Cited literature [160 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-01725324
Contributor : Firas Hmida <>
Submitted on : Wednesday, March 7, 2018 - 12:25:42 PM
Last modification on : Monday, October 19, 2020 - 11:08:12 AM
Long-term archiving on: : Friday, June 8, 2018 - 1:46:22 PM

File

phd-firashmida.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : tel-01725324, version 1

Citation

Firas Hmida. Identification et exploitation de contextes riches en connaissances pour l’aide à la traduction terminologique. Informatique et langage [cs.CL]. Université de Nantes, 2017. Français. ⟨tel-01725324⟩

Share

Metrics

Record views

242

Files downloads

503