Cross-Dialectal Transfer and Zero-Shot Learning for Armenian Varieties: A Comparative Analysis of RNNs, Transformers and LLMs - Inalco - Institut National des Langues et Civilisations Orientales
Communication Dans Un Congrès Année : 2024

Cross-Dialectal Transfer and Zero-Shot Learning for Armenian Varieties: A Comparative Analysis of RNNs, Transformers and LLMs

Résumé

This paper evaluates lemmatization, POS tagging, and morphological analysis for four Armenian varieties: Classical Armenian, Modern Eastern Armenian, Modern Western Armenian, and the under-documented Getashen dialect. It compares traditional RNN models, multilingual models like mDeBERTa, and large language models (ChatGPT) using supervised, transfer learning, and zero/few-shot learning approaches. The study finds that RNN models are particularly strong in POS-tagging, while large language models demonstrate high adaptability, especially in handling previously unseen dialect variations. The research highlights the value of cross-variational and in-context learning for enhancing NLP performance in low resource languages, offering crucial insights into model transferability and supporting the preservation of endangered dialects.
Fichier principal
Vignette du fichier
EMNLP_2024_Armenian_Dalih.pdf (2.13 Mo) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)
Licence

Dates et versions

hal-04722313 , version 1 (04-10-2024)

Licence

Identifiants

  • HAL Id : hal-04722313 , version 1

Citer

Chahan Vidal-Gorène, Nadi Tomeh, Victoria Khurshudyan. Cross-Dialectal Transfer and Zero-Shot Learning for Armenian Varieties: A Comparative Analysis of RNNs, Transformers and LLMs. 4th International Conference on Natural Language Processing for Digital Humanities, EMNLP 2024, Nov 2024, Miami, United States. ⟨hal-04722313⟩
0 Consultations
0 Téléchargements

Partager

More