Powered by OpenAIRE graph
Found an issue? Give us feedback
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/ https://www.aclweb.o...arrow_drop_down
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/
https://www.aclweb.org/antholo...
Article
License: CC BY
Data sources: UnpayWall
https://doi.org/10.18653/v1/p1...
Article . 2019 . Peer-reviewed
Data sources: Crossref
https://dx.doi.org/10.60692/4e...
Other literature type . 2019
Data sources: Datacite
https://dx.doi.org/10.60692/p5...
Other literature type . 2019
Data sources: Datacite
DBLP
Conference object
Data sources: DBLP
versions View all 4 versions
addClaim

This Research product is the result of merged Research products in OpenAIRE.

You have already added 0 works in your ORCID record related to the merged Research product.

Employing the Correspondence of Relations and Connectives to Identify Implicit Discourse Relations via Label Embeddings

توظيف مراسلات العلاقات والروابط لتحديد علاقات الخطاب الضمنية عبر تضمينات الملصقات
Authors: Linh The Nguyen; Linh Van Ngo; Khoat Than; Thien Huu Nguyen;

Employing the Correspondence of Relations and Connectives to Identify Implicit Discourse Relations via Label Embeddings

Abstract

Il a été démontré que les connecteurs implicites peuvent être exploités pour améliorer les performances des modèles de reconnaissance implicite des relations de discours (IDRR). Une propriété importante des connecteurs implicites est qu'ils peuvent être cartographiés avec précision dans les relations de discours véhiculant leurs fonctions. Dans ce travail, nous explorons cette propriété dans un cadre d'apprentissage multi-tâches pour l'IDRR dans lequel les relations et les connecteurs sont simultanément prédits, et la cartographie est exploitée pour transférer des connaissances entre les deux tâches de prédiction via les intégrations de relations et de connecteurs. Nous proposons plusieurs techniques pour permettre un tel transfert de connaissances qui donnent les performances de pointe pour l'IDRR sur plusieurs paramètres de l'ensemble de données de référence (c'est-à-dire l'ensemble de données Penn Discourse Treebank).

Se ha demostrado que los conectivos implícitos se pueden explotar para mejorar el rendimiento de los modelos de reconocimiento implícito de relaciones discursivas (IDRR). Una propiedad importante de los conectivos implícitos es que se pueden mapear con precisión en las relaciones discursivas que transmiten sus funciones. En este trabajo, exploramos esta propiedad en un marco de aprendizaje multitarea para IDRR en el que las relaciones y los conectivos se predicen simultáneamente, y el mapeo se aprovecha para transferir conocimiento entre las dos tareas de predicción a través de las incrustaciones de relaciones y conectivos. Proponemos varias técnicas para permitir dicha transferencia de conocimiento que produzcan el rendimiento de vanguardia para IDRR en varios entornos del conjunto de datos de referencia (es decir, el conjunto de datos del Penn Discourse Treebank).

It has been shown that implicit connectives can be exploited to improve the performance of the models for implicit discourse relation recognition (IDRR). An important property of the implicit connectives is that they can be accurately mapped into the discourse relations conveying their functions. In this work, we explore this property in a multi-task learning framework for IDRR in which the relations and the connectives are simultaneously predicted, and the mapping is leveraged to transfer knowledge between the two prediction tasks via the embeddings of relations and connectives. We propose several techniques to enable such knowledge transfer that yield the state-of-the-art performance for IDRR on several settings of the benchmark dataset (i.e., the Penn Discourse Treebank dataset).

لقد تبين أنه يمكن استغلال الروابط الضمنية لتحسين أداء نماذج التعرف على العلاقة الضمنية للخطاب (IDRR). من الخصائص المهمة للروابط الضمنية أنه يمكن تعيينها بدقة في علاقات الخطاب التي تنقل وظائفها. في هذا العمل، نستكشف هذه الخاصية في إطار تعلم متعدد المهام لـ IDRR حيث يتم التنبؤ بالعلاقات والروابط في وقت واحد، ويتم الاستفادة من التخطيط لنقل المعرفة بين مهمتي التنبؤ عبر تضمين العلاقات والروابط. نقترح العديد من التقنيات لتمكين نقل المعرفة هذا الذي يؤدي إلى الأداء المتطور لـ IDRR في العديد من إعدادات مجموعة البيانات المعيارية (أي مجموعة بيانات Penn Discourse Treebank).

Keywords

Relation (database), Artificial intelligence, Economics, Annotation, Knowledge management, Knowledge transfer, Semantic Processing, Epistemology, Task (project management), Artificial Intelligence, Implicit knowledge, Data mining, Natural Language Processing, Spoken Dialogue Systems, Referring Expressions, Extension (predicate logic), Geography, Natural language processing, Word Representation, Statistical Machine Translation and Natural Language Processing, Computer science, Language Modeling, Programming language, FOS: Philosophy, ethics and religion, Management, Philosophy, Treebank, Computer Science, Physical Sciences, Dialogue Act Modeling for Spoken Language Systems, Property (philosophy), Benchmark (surveying), Geodesy

  • BIP!
    Impact byBIP!
    selected citations
    These citations are derived from selected sources.
    This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
    12
    popularity
    This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.
    Top 10%
    influence
    This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
    Average
    impulse
    This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.
    Average
Powered by OpenAIRE graph
Found an issue? Give us feedback
selected citations
These citations are derived from selected sources.
This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
BIP!Citations provided by BIP!
popularity
This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.
BIP!Popularity provided by BIP!
influence
This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
BIP!Influence provided by BIP!
impulse
This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.
BIP!Impulse provided by BIP!
12
Top 10%
Average
Average
hybrid