Innovation in phraseomatics : DiCoP project and DiCoP-Text corpus for the enrichment of Language Models and Automatic Translation - Archive ouverte HAL
Communication Dans Un Congrès Année : 2024

Innovation in phraseomatics : DiCoP project and DiCoP-Text corpus for the enrichment of Language Models and Automatic Translation

Résumé

This article examines advances in phraseomatics (L. Chen, 2023) and digital phraseography through the DiCoP project and its DiCoP-Text corpus, aimed at enriching linguistic models and machine translation. The project evaluates the frequency of use of phraseological units (PUs) and improves their translation in different contexts, drawing on recent research in phraseotranslation (Sułkowska, 2022) and natural language processing (NLP). It emphasizes French-Chinese and Chinese-French language pairs. We integrated 549 PUs from the novel The Three-Body Problem by Liu Cixin for our tests. Various processes, such as tokenization, identification, alignment, and annotation, were used to improve the translation of PUs. DiCoP-Text, a comprehensive database including newspaper articles, literary works, and textbooks, aims to enhance the performance of language models (LMs).
Fichier principal
Vignette du fichier
Euralex-XXI.pdf (4.4 Mo) Télécharger le fichier
Origine Fichiers éditeurs autorisés sur une archive ouverte

Dates et versions

hal-04664240 , version 1 (08-10-2024)

Identifiants

  • HAL Id : hal-04664240 , version 1

Citer

Lian Chen, Wenjun Sun, Flora Badin. Innovation in phraseomatics : DiCoP project and DiCoP-Text corpus for the enrichment of Language Models and Automatic Translation. EURALEX 2024 - 21st EURALEX International Congress Lexicography and Semantics, Kristina Štrkalj Despot; Ana Ostroški Anić; Ivana Brač, Oct 2024, Cavtat, Croatia. pp.227-234. ⟨hal-04664240⟩
66 Consultations
0 Téléchargements

Partager

More