TINA: Textual Inference with Negation Augmentation - Département Informatique et Réseaux Accéder directement au contenu
Communication Dans Un Congrès Année : 2022

TINA: Textual Inference with Negation Augmentation

Résumé

Transformer-based language models achieve state-of-the-art results on several natural language processing tasks. One of these is textual entailment, i.e., the task of determining whether a premise logically entails a hypothesis. However, the models perform poorly on this task when the examples contain negations. In this paper, we propose a new definition of textual entailment that captures also negation. This allows us to develop TINA (Textual Inference with Negation Augmentation), a principled technique for negated data augmentation that can be combined with the unlikelihood loss function. Our experiments with different transformer-based models show that our method can significantly improve the performance of the models on textual entailment datasets with negation-without sacrificing performance on datasets without negation.
Fichier principal
Vignette du fichier
TINA__Textual_Inference_with_Negation_Augmentation-2.pdf (299.02 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-03870605 , version 1 (24-11-2022)

Identifiants

  • HAL Id : hal-03870605 , version 1

Citer

Chadi Helwe, Simon Coumes, Chloé Clavel, Fabian Suchanek. TINA: Textual Inference with Negation Augmentation. The 2022 Conference on Empirical Methods in Natural Language Processing ( EMNLP 2022 ), Dec 2022, Abu Dhabi, United Arab Emirates. ⟨hal-03870605⟩
124 Consultations
86 Téléchargements

Partager

Gmail Facebook X LinkedIn More