Comparative Investigation of Deep Learning Components for End-to-end Implicit Discourse Relationship Parser

  • Dejian Li
  • , Man Lan*
  • , Yuanbin Wu
  • *Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

1 Scopus citations

Abstract

The neural components in deep learning framework are crucial for the performance of many natural language processing tasks. So far there is no systematic work to investigate the influence of neural components on the performance of implicit discourse relation recognition. To address it, in this work we compare many different components and build two implicit discourse parsers base on the sequence and structure of sentence respectively. Experimental results show due to different linguistic features, the neural components have different effects in English and Chinese. Besides, our models achieve state-of-the-art performance on CoNLL-2016 English and Chinese datasets.

Original languageEnglish
Title of host publicationChinese Computational Linguistics - 18th China National Conference, CCL 2019, Proceedings
EditorsMaosong Sun, Yang Liu, Zhiyuan Liu, Xuanjing Huang, Heng Ji
PublisherSpringer
Pages143-155
Number of pages13
ISBN (Print)9783030323806
DOIs
StatePublished - 2019
Event18th China National Conference on Computational Linguistics, CCL 2019 - Kunming, China
Duration: 18 Oct 201920 Oct 2019

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11856 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference18th China National Conference on Computational Linguistics, CCL 2019
Country/TerritoryChina
CityKunming
Period18/10/1920/10/19

Keywords

  • Deep learning
  • Implicit discourse relation classification
  • Neural network
  • Word embedding

Fingerprint

Dive into the research topics of 'Comparative Investigation of Deep Learning Components for End-to-end Implicit Discourse Relationship Parser'. Together they form a unique fingerprint.

Cite this