TU Darmstadt / ULB / TUbiblio

Neural End-to-End Learning for Computational Argumentation Mining

Eger, Steffen and Daxenberger, Johannes and Gurevych, Iryna (2017):
Neural End-to-End Learning for Computational Argumentation Mining.
In: Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (ACL 2017), Association for Computational Linguistics, Vancouver, Canada, Volume 1: Long Papers, [Online-Edition: http://aclweb.org/anthology/P17-1002],
[Conference or Workshop Item]

Abstract

We investigate neural techniques for end-to-end computational argumentation mining (AM). We frame AM both as a token-based dependency parsing and as a token-based sequence tagging problem, including a multi-task learning setup. Contrary to models that operate on the argument component level, we find that framing AM as dependency parsing leads to subpar performance results. In contrast, less complex (local) tagging models based on BiLSTMs perform robustly across classification scenarios, being able to catch long-range dependencies inherent to the AM problem. Moreover, we find that jointly learning ‘natural’ subtasks, in a multi-task learning setup, improves performance.

Item Type: Conference or Workshop Item
Erschienen: 2017
Creators: Eger, Steffen and Daxenberger, Johannes and Gurevych, Iryna
Title: Neural End-to-End Learning for Computational Argumentation Mining
Language: English
Abstract:

We investigate neural techniques for end-to-end computational argumentation mining (AM). We frame AM both as a token-based dependency parsing and as a token-based sequence tagging problem, including a multi-task learning setup. Contrary to models that operate on the argument component level, we find that framing AM as dependency parsing leads to subpar performance results. In contrast, less complex (local) tagging models based on BiLSTMs perform robustly across classification scenarios, being able to catch long-range dependencies inherent to the AM problem. Moreover, we find that jointly learning ‘natural’ subtasks, in a multi-task learning setup, improves performance.

Title of Book: Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (ACL 2017)
Volume: Volume 1: Long Papers
Publisher: Association for Computational Linguistics
Uncontrolled Keywords: UKP_a_DLinNLP, UKP_a_ArMin, reviewed, UKP_p_ArgumenText
Divisions: 20 Department of Computer Science
20 Department of Computer Science > Ubiquitous Knowledge Processing
Event Location: Vancouver, Canada
Date Deposited: 31 Mar 2017 14:02
Official URL: http://aclweb.org/anthology/P17-1002
Identification Number: TUD-CS-2017-0070
Related URLs:
Projects: ArgumenText
Export:
Suche nach Titel in: TUfind oder in Google

Optionen (nur für Redakteure)

View Item View Item