Warrant generation through deep learning

Fatima Alkhawaldeh; Tommy Yuan; Dimitar Lubomirov Kazakov

doi:10.5121/csit.2021.112005

Warrant generation through deep learning

Fatima Alkhawaldeh, Tommy Yuan, Dimitar Lubomirov Kazakov

Computer Science

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

The warrant element of the Toulmin model is critical for fact-checking and assessing the strength of an argument. As implicit information, warrants justify the arguments and explain why the evidence supports the claim. Despite the critical role warrants play in facilitating argument comprehension, the fact that most works aim to select the best warrant from existing structured data and labelled data is scarce presents a fact-checking challenge, particularly when the evidence is insufficient, or the conclusion is not inferred or generated well based on the evidence. Additionally, deep learning methods for false information detection face a significant bottleneck due to their training requirement of a large amount of labelled data. Manually annotating data, on the other hand, is a time-consuming and laborious process. Thus, we examine the extent to which warrants can be retrieved or reconfigured using unstructured data obtained from their premises.

Original language	English
Title of host publication	Seventh International Conference on Natural Language Computing (NATL 2021), Proc. of
Editors	David C. Wyld, et al.
Publisher	AIRCC Publishing Corporation
Pages	53-75
Number of pages	23
DOIs	https://doi.org/10.5121/csit.2021.112005
Publication status	Published - 27 Nov 2021
Event	International Conference on Natural Language Computing - London, United Kingdom Duration: 27 Nov 2021 → 28 Feb 2022 Conference number: 7

Publication series

Name	Computer Science & Information Technology (CS & IT)
Publisher	AIRCC Publishing Corporation

Conference

Conference	International Conference on Natural Language Computing
Abbreviated title	NATL 2021
Country/Territory	United Kingdom
City	London
Period	27/11/21 → 28/02/22

Keywords

Toulmin model
warrant
Fact-checking
Deep learning

Access to Document

10.5121/csit.2021.112005Licence: CC BY

csit112005-WarrantGenerationFinal published version, 1.42 MBLicence: CC BY

Cite this

@inproceedings{8824d51205b74d309a45097aedbc747b,

title = "Warrant generation through deep learning",

abstract = "The warrant element of the Toulmin model is critical for fact-checking and assessing the strength of an argument. As implicit information, warrants justify the arguments and explain why the evidence supports the claim. Despite the critical role warrants play in facilitating argument comprehension, the fact that most works aim to select the best warrant from existing structured data and labelled data is scarce presents a fact-checking challenge, particularly when the evidence is insufficient, or the conclusion is not inferred or generated well based on the evidence. Additionally, deep learning methods for false information detection face a significant bottleneck due to their training requirement of a large amount of labelled data. Manually annotating data, on the other hand, is a time-consuming and laborious process. Thus, we examine the extent to which warrants can be retrieved or reconfigured using unstructured data obtained from their premises.",

keywords = "Toulmin model, warrant, Fact-checking, Deep learning",

author = "Fatima Alkhawaldeh and Tommy Yuan and Kazakov, {Dimitar Lubomirov}",

year = "2021",

month = nov,

day = "27",

doi = "10.5121/csit.2021.112005",

language = "English",

series = "Computer Science & Information Technology (CS & IT)",

publisher = "AIRCC Publishing Corporation",

pages = "53--75",

editor = "Wyld, {David C.} and et al.",

booktitle = "Seventh International Conference on Natural Language Computing (NATL 2021), Proc. of",

note = "International Conference on Natural Language Computing, NATL 2021 ; Conference date: 27-11-2021 Through 28-02-2022",

}

Alkhawaldeh, F, Yuan, T & Kazakov, DL 2021, Warrant generation through deep learning. in DC Wyld & E al. (eds), Seventh International Conference on Natural Language Computing (NATL 2021), Proc. of. Computer Science & Information Technology (CS & IT), AIRCC Publishing Corporation, pp. 53-75, International Conference on Natural Language Computing, London, United Kingdom, 27/11/21. https://doi.org/10.5121/csit.2021.112005

Warrant generation through deep learning. / Alkhawaldeh, Fatima; Yuan, Tommy; Kazakov, Dimitar Lubomirov.

Seventh International Conference on Natural Language Computing (NATL 2021), Proc. of. ed. / David C. Wyld; et al. AIRCC Publishing Corporation, 2021. p. 53-75 (Computer Science & Information Technology (CS & IT)).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Warrant generation through deep learning

AU - Alkhawaldeh, Fatima

AU - Yuan, Tommy

AU - Kazakov, Dimitar Lubomirov

N1 - Conference code: 7

PY - 2021/11/27

Y1 - 2021/11/27

N2 - The warrant element of the Toulmin model is critical for fact-checking and assessing the strength of an argument. As implicit information, warrants justify the arguments and explain why the evidence supports the claim. Despite the critical role warrants play in facilitating argument comprehension, the fact that most works aim to select the best warrant from existing structured data and labelled data is scarce presents a fact-checking challenge, particularly when the evidence is insufficient, or the conclusion is not inferred or generated well based on the evidence. Additionally, deep learning methods for false information detection face a significant bottleneck due to their training requirement of a large amount of labelled data. Manually annotating data, on the other hand, is a time-consuming and laborious process. Thus, we examine the extent to which warrants can be retrieved or reconfigured using unstructured data obtained from their premises.

AB - The warrant element of the Toulmin model is critical for fact-checking and assessing the strength of an argument. As implicit information, warrants justify the arguments and explain why the evidence supports the claim. Despite the critical role warrants play in facilitating argument comprehension, the fact that most works aim to select the best warrant from existing structured data and labelled data is scarce presents a fact-checking challenge, particularly when the evidence is insufficient, or the conclusion is not inferred or generated well based on the evidence. Additionally, deep learning methods for false information detection face a significant bottleneck due to their training requirement of a large amount of labelled data. Manually annotating data, on the other hand, is a time-consuming and laborious process. Thus, we examine the extent to which warrants can be retrieved or reconfigured using unstructured data obtained from their premises.

KW - Toulmin model

KW - warrant

KW - Fact-checking

KW - Deep learning

U2 - 10.5121/csit.2021.112005

DO - 10.5121/csit.2021.112005

M3 - Conference contribution

T3 - Computer Science & Information Technology (CS & IT)

SP - 53

EP - 75

BT - Seventh International Conference on Natural Language Computing (NATL 2021), Proc. of

A2 - Wyld, David C.

A2 - al., et

PB - AIRCC Publishing Corporation

T2 - International Conference on Natural Language Computing

Y2 - 27 November 2021 through 28 February 2022

ER -