Publication:
PROTEST-ER: retraining BERT for protest event extraction

Simple item page
dc.contributor.coauthorCaselli, Tommaso
dc.contributor.coauthorBasile, Angelo
dc.contributor.departmentDepartment of Sociology
dc.contributor.departmentDepartment of Computer Engineering
dc.contributor.kuauthorHürriyetoğlu, Ali
dc.contributor.kuauthorMutlu, Osman
dc.contributor.kuprofileTeaching Faculty
dc.contributor.kuprofileResearcher
dc.contributor.otherDepartment of Sociology
dc.contributor.otherDepartment of Computer Engineering
dc.contributor.schoolcollegeinstituteCollege of Social Sciences and Humanities
dc.contributor.schoolcollegeinstituteCollege of Engineering
dc.date.accessioned2024-11-09T13:23:33Z
dc.date.issued2021
dc.description.abstractWe analyze the effect of further pre-training BERT with different domain specific data as an unsupervised domain adaptation strategy for event extraction. Portability of event extraction models is particularly challenging, with large performance drops affecting data on the same text genres (e.g., news). We present PROTEST-ER, a retrained BERT model for protest event extraction. PROTEST-ER outperforms a corresponding generic BERT on out-of-domain data of 8.1 points. Our best performing models reach 51.91-46.39 F1 across both domains.
dc.description.fulltextYES
dc.description.indexedbyWoS
dc.description.indexedbyScopus
dc.description.openaccessYES
dc.description.publisherscopeInternational
dc.description.sponsoredbyTubitakEuEU
dc.description.sponsorshipEuropean Union (EU)
dc.description.sponsorshipHorizon 2020
dc.description.sponsorshipEuropean Research Council (ERC)
dc.description.versionPublisher version
dc.formatpdf
dc.identifier.doi10.18653/v1/2021.case-1.4
dc.identifier.embargoNO
dc.identifier.filenameinventorynoIR03287
dc.identifier.isbn978-1-954085-79-4
dc.identifier.linkhttps://doi.org/10.18653/v1/2021.case-1.4
dc.identifier.quartileN/A
dc.identifier.scopus2-s2.0-85119310309
dc.identifier.urihttps://hdl.handle.net/20.500.14288/3375
dc.identifier.wos694853100004
dc.keywordsAdaptation strategies
dc.keywordsDifferent domains
dc.keywordsDomain adaptation
dc.keywordsDomain specific
dc.keywordsEvents extractions
dc.keywordsExtraction modeling
dc.keywordsPerformance
dc.keywordsPre-training
dc.keywordsText genre
dc.languageEnglish
dc.publisherAssociation for Computational Linguistics (ACL)
dc.relation.grantno714868
dc.relation.urihttp://cdm21054.contentdm.oclc.org/cdm/ref/collection/IR/id/10071
dc.sourceProceedings of the 4th Workshop on Challenges and Applications of Automated Extraction of Socio-political Events from Text (CASE 2021)
dc.subjectComputer science
dc.subjectArtificial intelligence
dc.subjectComputer science, interdisciplinary applications
dc.subjectLinguistics
dc.titlePROTEST-ER: retraining BERT for protest event extraction
dc.typeConference proceeding
dspace.entity.typePublication
local.contributor.kuauthorHürriyetoğlu, Ali
local.contributor.kuauthorMutlu, Osman
relation.isOrgUnitOfPublication10f5be47-fab1-42a1-af66-1642ba4aff8e
relation.isOrgUnitOfPublication89352e43-bf09-4ef4-82f6-6f9d0174ebae
relation.isOrgUnitOfPublication.latestForDiscovery10f5be47-fab1-42a1-af66-1642ba4aff8e

Files

Original bundle

Now showing 1 - 1 of 1
Thumbnail Image
Name:
10071.pdf
Size:
420.57 KB
Format:
Adobe Portable Document Format
Download

Collections

Publications with Fulltext