BERT for Evidence Retrieval and Claim Verification

We investigate BERT in an evidence retrieval and claim verification pipeline for the task of evidence-based claim verification. To this end, we propose to use two BERT models, one for retrieving evidence sentences supporting or rejecting claims, and another for verifying claims based on the retrieve...

Descripción completa

Detalles Bibliográficos
Autores principales: Soleimani, Amir, Monz, Christof, Worring, Marcel
Formato: Online Artículo Texto
Lenguaje:English
Publicado: 2020
Materias:
Article
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7148011/
http://dx.doi.org/10.1007/978-3-030-45442-5_45
_version_ 1783520510970167296
author Soleimani, Amir
Monz, Christof
Worring, Marcel
author_facet Soleimani, Amir
Monz, Christof
Worring, Marcel
author_sort Soleimani, Amir
collection PubMed
description We investigate BERT in an evidence retrieval and claim verification pipeline for the task of evidence-based claim verification. To this end, we propose to use two BERT models, one for retrieving evidence sentences supporting or rejecting claims, and another for verifying claims based on the retrieved evidence sentences. To train the BERT retrieval system, we use pointwise and pairwise loss functions and examine the effect of hard negative mining. Our system achieves a new state of the art recall of 87.1 for retrieving evidence sentences out of the FEVER dataset 50K Wikipedia pages, and scores second in the leaderboard with the FEVER score of 69.7.
format Online
Article
Text
id pubmed-7148011
institution National Center for Biotechnology Information
language English
publishDate 2020
record_format MEDLINE/PubMed
spelling pubmed-71480112020-04-13 BERT for Evidence Retrieval and Claim Verification Soleimani, Amir Monz, Christof Worring, Marcel Advances in Information Retrieval Article We investigate BERT in an evidence retrieval and claim verification pipeline for the task of evidence-based claim verification. To this end, we propose to use two BERT models, one for retrieving evidence sentences supporting or rejecting claims, and another for verifying claims based on the retrieved evidence sentences. To train the BERT retrieval system, we use pointwise and pairwise loss functions and examine the effect of hard negative mining. Our system achieves a new state of the art recall of 87.1 for retrieving evidence sentences out of the FEVER dataset 50K Wikipedia pages, and scores second in the leaderboard with the FEVER score of 69.7. 2020-03-24 /pmc/articles/PMC7148011/ http://dx.doi.org/10.1007/978-3-030-45442-5_45 Text en © Springer Nature Switzerland AG 2020 This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic.
spellingShingle Article
Soleimani, Amir
Monz, Christof
Worring, Marcel
BERT for Evidence Retrieval and Claim Verification
title BERT for Evidence Retrieval and Claim Verification
title_full BERT for Evidence Retrieval and Claim Verification
title_fullStr BERT for Evidence Retrieval and Claim Verification
title_full_unstemmed BERT for Evidence Retrieval and Claim Verification
title_short BERT for Evidence Retrieval and Claim Verification
title_sort bert for evidence retrieval and claim verification
topic Article
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7148011/
http://dx.doi.org/10.1007/978-3-030-45442-5_45
work_keys_str_mv AT soleimaniamir bertforevidenceretrievalandclaimverification
AT monzchristof bertforevidenceretrievalandclaimverification
AT worringmarcel bertforevidenceretrievalandclaimverification

Ejemplares similares