Solving Stance Detection on Tweets as Multi-Domain and Multi-Task Text Classification

Stance detection on tweets aims at classifying the attitude of tweets towards given targets. Existing work leverage attention-based models to learn target-aware stance representations. While those methods achieve substantial success, most of them usually train a model for each target separately desp...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Limin Wang, Dexin Wang
Formato:	article
Lenguaje:	EN
Publicado:	IEEE 2021
Materias:	Artificial intelligence natural language processing sentiment analysis Electrical engineering. Electronics. Nuclear engineering TK1-9971
Acceso en línea:	https://doaj.org/article/738e07e8d23f47aaa75282cd5a4ed1c3
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:738e07e8d23f47aaa75282cd5a4ed1c3
record_format	dspace
spelling	oai:doaj.org-article:738e07e8d23f47aaa75282cd5a4ed1c32021-12-03T00:01:23ZSolving Stance Detection on Tweets as Multi-Domain and Multi-Task Text Classification2169-353610.1109/ACCESS.2021.3129468https://doaj.org/article/738e07e8d23f47aaa75282cd5a4ed1c32021-01-01T00:00:00Zhttps://ieeexplore.ieee.org/document/9622274/https://doaj.org/toc/2169-3536Stance detection on tweets aims at classifying the attitude of tweets towards given targets. Existing work leverage attention-based models to learn target-aware stance representations. While those methods achieve substantial success, most of them usually train a model for each target separately despite the scarcity of annotated data for each target. To alleviate limitation of annotated data, some methods turn to external linguistic resources, additional sentiment annotations or target-aware data augmentation techniques for better detection results. We argue that the sharedness of stance-related features across targets in the existing stance detection dataset is not fully exploited. However, directly training on mixed examples of all targets may confuse the model in learning best features for each target. To this end, we borrow the idea from transfer learning and multi-task learning, and formulate stance detection on tweets as a multi-domain multi-task classification problem. We apply the target adversarial learning to capture stance-related features shared by all targets and target descriptors for learning stance-informative features correlating to specific targets. Experimental results on the benchmark SemEval 2016 dataset demonstrate the effectiveness of our model, which outperforms BERT model by over 2% on macro average F1 and achieves superior performance than many recent methods utilizing external resources. We further provide detailed analyses to illustrate the superiority of fully utilizing features shared by different targets.Limin WangDexin WangIEEEarticleArtificial intelligencenatural language processingsentiment analysisElectrical engineering. Electronics. Nuclear engineeringTK1-9971ENIEEE Access, Vol 9, Pp 157780-157789 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Artificial intelligence natural language processing sentiment analysis Electrical engineering. Electronics. Nuclear engineering TK1-9971
spellingShingle	Artificial intelligence natural language processing sentiment analysis Electrical engineering. Electronics. Nuclear engineering TK1-9971 Limin Wang Dexin Wang Solving Stance Detection on Tweets as Multi-Domain and Multi-Task Text Classification
description	Stance detection on tweets aims at classifying the attitude of tweets towards given targets. Existing work leverage attention-based models to learn target-aware stance representations. While those methods achieve substantial success, most of them usually train a model for each target separately despite the scarcity of annotated data for each target. To alleviate limitation of annotated data, some methods turn to external linguistic resources, additional sentiment annotations or target-aware data augmentation techniques for better detection results. We argue that the sharedness of stance-related features across targets in the existing stance detection dataset is not fully exploited. However, directly training on mixed examples of all targets may confuse the model in learning best features for each target. To this end, we borrow the idea from transfer learning and multi-task learning, and formulate stance detection on tweets as a multi-domain multi-task classification problem. We apply the target adversarial learning to capture stance-related features shared by all targets and target descriptors for learning stance-informative features correlating to specific targets. Experimental results on the benchmark SemEval 2016 dataset demonstrate the effectiveness of our model, which outperforms BERT model by over 2% on macro average F1 and achieves superior performance than many recent methods utilizing external resources. We further provide detailed analyses to illustrate the superiority of fully utilizing features shared by different targets.
format	article
author	Limin Wang Dexin Wang
author_facet	Limin Wang Dexin Wang
author_sort	Limin Wang
title	Solving Stance Detection on Tweets as Multi-Domain and Multi-Task Text Classification
title_short	Solving Stance Detection on Tweets as Multi-Domain and Multi-Task Text Classification
title_full	Solving Stance Detection on Tweets as Multi-Domain and Multi-Task Text Classification
title_fullStr	Solving Stance Detection on Tweets as Multi-Domain and Multi-Task Text Classification
title_full_unstemmed	Solving Stance Detection on Tweets as Multi-Domain and Multi-Task Text Classification
title_sort	solving stance detection on tweets as multi-domain and multi-task text classification
publisher	IEEE
publishDate	2021
url	https://doaj.org/article/738e07e8d23f47aaa75282cd5a4ed1c3
work_keys_str_mv	AT liminwang solvingstancedetectionontweetsasmultidomainandmultitasktextclassification AT dexinwang solvingstancedetectionontweetsasmultidomainandmultitasktextclassification
_version_	1718373994418143232

Solving Stance Detection on Tweets as Multi-Domain and Multi-Task Text Classification

Ejemplares similares