Mixing and Matching Emotion Frameworks: Investigating Cross-Framework Transfer Learning for Dutch Emotion Detection

Emotion detection has become a growing field of study, especially seeing its broad application potential. Research usually focuses on emotion classification, but performance tends to be rather low, especially when dealing with more advanced emotion categories that are tailored to specific tasks and...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Luna De Bruyne, Orphée De Clercq, Véronique Hoste
Formato:	article
Lenguaje:	EN
Publicado:	MDPI AG 2021
Materias:	emotion detection multi-task learning transfer learning emotion frameworks Electronics TK7800-8360
Acceso en línea:	https://doaj.org/article/f479121d12454f8fa12800011ce928e5
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:f479121d12454f8fa12800011ce928e5
record_format	dspace
spelling	oai:doaj.org-article:f479121d12454f8fa12800011ce928e52021-11-11T15:39:06ZMixing and Matching Emotion Frameworks: Investigating Cross-Framework Transfer Learning for Dutch Emotion Detection10.3390/electronics102126432079-9292https://doaj.org/article/f479121d12454f8fa12800011ce928e52021-10-01T00:00:00Zhttps://www.mdpi.com/2079-9292/10/21/2643https://doaj.org/toc/2079-9292Emotion detection has become a growing field of study, especially seeing its broad application potential. Research usually focuses on emotion classification, but performance tends to be rather low, especially when dealing with more advanced emotion categories that are tailored to specific tasks and domains. Therefore, we propose the use of the dimensional emotion representations valence, arousal and dominance (VAD), in an emotion regression task. Firstly, we hypothesize that they can improve performance of the classification task, and secondly, they might be used as a pivot mechanism to map towards any given emotion framework, which allows tailoring emotion frameworks to specific applications. In this paper, we examine three cross-framework transfer methodologies: multi-task learning, in which VAD regression and classification are learned simultaneously; meta-learning, where VAD regression and emotion classification are learned separately and predictions are jointly used as input for a meta-learner; and a pivot mechanism, which converts the predictions of the VAD model to emotion classes. We show that dimensional representations can indeed boost performance for emotion classification, especially in the meta-learning setting (up to 7% macro F1-score compared to regular emotion classification). The pivot method was not able to compete with the base model, but further inspection suggests that it could be efficient, provided that the VAD regression model is further improved.Luna De BruyneOrphée De ClercqVéronique HosteMDPI AGarticleemotion detectionmulti-task learningtransfer learningemotion frameworksElectronicsTK7800-8360ENElectronics, Vol 10, Iss 2643, p 2643 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	emotion detection multi-task learning transfer learning emotion frameworks Electronics TK7800-8360
spellingShingle	emotion detection multi-task learning transfer learning emotion frameworks Electronics TK7800-8360 Luna De Bruyne Orphée De Clercq Véronique Hoste Mixing and Matching Emotion Frameworks: Investigating Cross-Framework Transfer Learning for Dutch Emotion Detection
description	Emotion detection has become a growing field of study, especially seeing its broad application potential. Research usually focuses on emotion classification, but performance tends to be rather low, especially when dealing with more advanced emotion categories that are tailored to specific tasks and domains. Therefore, we propose the use of the dimensional emotion representations valence, arousal and dominance (VAD), in an emotion regression task. Firstly, we hypothesize that they can improve performance of the classification task, and secondly, they might be used as a pivot mechanism to map towards any given emotion framework, which allows tailoring emotion frameworks to specific applications. In this paper, we examine three cross-framework transfer methodologies: multi-task learning, in which VAD regression and classification are learned simultaneously; meta-learning, where VAD regression and emotion classification are learned separately and predictions are jointly used as input for a meta-learner; and a pivot mechanism, which converts the predictions of the VAD model to emotion classes. We show that dimensional representations can indeed boost performance for emotion classification, especially in the meta-learning setting (up to 7% macro F1-score compared to regular emotion classification). The pivot method was not able to compete with the base model, but further inspection suggests that it could be efficient, provided that the VAD regression model is further improved.
format	article
author	Luna De Bruyne Orphée De Clercq Véronique Hoste
author_facet	Luna De Bruyne Orphée De Clercq Véronique Hoste
author_sort	Luna De Bruyne
title	Mixing and Matching Emotion Frameworks: Investigating Cross-Framework Transfer Learning for Dutch Emotion Detection
title_short	Mixing and Matching Emotion Frameworks: Investigating Cross-Framework Transfer Learning for Dutch Emotion Detection
title_full	Mixing and Matching Emotion Frameworks: Investigating Cross-Framework Transfer Learning for Dutch Emotion Detection
title_fullStr	Mixing and Matching Emotion Frameworks: Investigating Cross-Framework Transfer Learning for Dutch Emotion Detection
title_full_unstemmed	Mixing and Matching Emotion Frameworks: Investigating Cross-Framework Transfer Learning for Dutch Emotion Detection
title_sort	mixing and matching emotion frameworks: investigating cross-framework transfer learning for dutch emotion detection
publisher	MDPI AG
publishDate	2021
url	https://doaj.org/article/f479121d12454f8fa12800011ce928e5
work_keys_str_mv	AT lunadebruyne mixingandmatchingemotionframeworksinvestigatingcrossframeworktransferlearningfordutchemotiondetection AT orpheedeclercq mixingandmatchingemotionframeworksinvestigatingcrossframeworktransferlearningfordutchemotiondetection AT veroniquehoste mixingandmatchingemotionframeworksinvestigatingcrossframeworktransferlearningfordutchemotiondetection
_version_	1718434707212861440

Mixing and Matching Emotion Frameworks: Investigating Cross-Framework Transfer Learning for Dutch Emotion Detection

Ejemplares similares