A Comparative Study of Ensemble Approaches to Fact-Checking for the FEVER Shared Task

Oluwabamigbe O. Oni, Terence L.Van Zyl

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

3 Citations (Scopus)

Abstract

The surge of information globally motivates for automated rumour detection. We use Fact-checking to detect rumours of the type misinformation. The FEVER-shared task is the Fact-checking task used for this comparative study. The task is divided into Document Retrieval, Sentence Selection, and Claim Verification components. We standardise TF-IDF for Document Retrieval, before creating the pipelines of Sentence Selection and Claim Verification algorithms. We evaluate each unique combination on the FEVER score, then compare the four pipelines to the baseline and state of the art. Our results show that the 2-way classification task using the Siamese BiLSTM achieves better Evidence Retrieval F1 scores than the state of the art models, and that the pipeline combinations, rival the state of the art for the Shared Task. The novelty of this research lies in the standardised text processing on novel pipeline combinations, allowing for comparable results, as well as the evaluation of the Siamese BiLSTM.

Original languageEnglish
Title of host publication2020 IEEE Asia-Pacific Conference on Computer Science and Data Engineering, CSDE 2020
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781665419741
DOIs
Publication statusPublished - 16 Dec 2020
Event2020 IEEE Asia-Pacific Conference on Computer Science and Data Engineering, CSDE 2020 - Gold Coast, Australia
Duration: 16 Dec 202018 Dec 2020

Publication series

Name2020 IEEE Asia-Pacific Conference on Computer Science and Data Engineering, CSDE 2020

Conference

Conference2020 IEEE Asia-Pacific Conference on Computer Science and Data Engineering, CSDE 2020
Country/TerritoryAustralia
CityGold Coast
Period16/12/2018/12/20

Keywords

  • Claim Verification
  • Document Retrieval
  • Document selection
  • Evidence Retrieval
  • Fact-checking
  • FEVER

ASJC Scopus subject areas

  • Artificial Intelligence
  • Computer Networks and Communications
  • Computer Science Applications
  • Information Systems and Management
  • Health Informatics
  • Communication

Fingerprint

Dive into the research topics of 'A Comparative Study of Ensemble Approaches to Fact-Checking for the FEVER Shared Task'. Together they form a unique fingerprint.

Cite this