Conference Papers Year : 2024

Zero-shot learning for multilingual discourse relation classification

Abstract

Classifying discourse relations is a hard task: discourse-annotated data is scarce, especially for languages other than English, and there exist different theoretical frameworks that affect textual spans to be linked and the label set used. Thus, work on transfer between languages is very limited, especially between frameworks, while it could improve our understanding of some theoretical aspects and enhance many applications. In this paper, we propose the first experiments on zero-shot learning for discourse relation classification and investigate several paths in the way source data can be combined, either based on languages, frameworks, or similarity measures. We demonstrate how difficult transfer is for the task at hand, and that the most impactful factor is label set divergence, where the notion of underlying framework possibly conceals crucial disagreements.
Fichier principal
Vignette du fichier
2024.lrec-main.1553.pdf (200.81 Ko) Télécharger le fichier
Origin Publisher files allowed on an open archive

Dates and versions

hal-04483805 , version 1 (29-02-2024)
hal-04483805 , version 2 (06-06-2024)

Licence

Identifiers

  • HAL Id : hal-04483805 , version 2

Cite

Eleni Metheniti, Philippe Muller, Chloé Braud, Margarita Hernández-Casas. Zero-shot learning for multilingual discourse relation classification. Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024), May 2024, Turin, Italy. pp.17858-17876. ⟨hal-04483805v2⟩
735 View
214 Download

Share

More