Näytä suppeat kuvailutiedot

Explainable Publication Year Prediction of Eighteenth Century Texts with the BERT Model

Repo Liina; Tiihonen Iiro; Ginter Filip; Babbar Rohit; Tolonen Mikko; Rastas Iiro; Qaraei Mohammedreza; Mäkelä Eetu; Ryan Yann

dc.contributor.authorRepo Liina
dc.contributor.authorTiihonen Iiro
dc.contributor.authorGinter Filip
dc.contributor.authorBabbar Rohit
dc.contributor.authorTolonen Mikko
dc.contributor.authorRastas Iiro
dc.contributor.authorQaraei Mohammedreza
dc.contributor.authorMäkelä Eetu
dc.contributor.authorRyan Yann
dc.date.accessioned2022-11-29T14:57:15Z
dc.date.available2022-11-29T14:57:15Z
dc.identifier.isbn978-1-955917-42-1
dc.identifier.urihttps://www.utupub.fi/handle/10024/173140
dc.description.abstractIn this paper, we describe a BERT model trained on the Eighteenth Century Collections Online (ECCO) dataset of digitized documents. The ECCO dataset poses unique modelling challenges due to the presence of Optical Character Recognition (OCR) artifacts. We establish the performance of the BERT model on a publication year prediction task against linear baseline models and human judgement, finding the BERT model to be superior to both and able to date the works, on average, with less than 7 years absolute error. We also explore how language change over time affects the model by analyzing the features the model uses for publication year predictions as given by the Integrated Gradients model explanation method.
dc.language.isoen
dc.titleExplainable Publication Year Prediction of Eighteenth Century Texts with the BERT Model
dc.identifier.urlhttps://aclanthology.org/2022.lchange-1.7.pdf
dc.identifier.urnURN:NBN:fi-fe2022110164053
dc.contributor.organizationfi=kieli- ja käännöstieteiden laitos, yhteiset|en=School of Languages and Translation Studies|
dc.contributor.organizationfi=data-analytiikka|en=Data-analytiikka|
dc.contributor.organization-code2602100
dc.contributor.organization-code2610301
dc.converis.publication-id176709131
dc.converis.urlhttps://research.utu.fi/converis/portal/Publication/176709131
dc.format.pagerange68
dc.format.pagerange77
dc.okm.affiliatedauthorGinter, Filip
dc.okm.affiliatedauthorRastas, Iiro
dc.okm.affiliatedauthorRepo, Liina
dc.okm.discipline113 Computer and information sciencesen_GB
dc.okm.discipline113 Tietojenkäsittely ja informaatiotieteetfi_FI
dc.okm.internationalcopublicationnot an international co-publication
dc.okm.internationalityInternational publication
dc.okm.typeConference proceedings article
dc.publisher.countryUnited Statesen_GB
dc.publisher.countryYhdysvallat (USA)fi_FI
dc.publisher.country-codeUS
dc.relation.conferenceWorkshop on Computational Approaches to Historical Language Change
dc.title.bookProceedings of the 3rd Workshop on Computational Approaches to Historical Language Change
dc.year.issued2022


Aineistoon kuuluvat tiedostot

Thumbnail

Aineisto kuuluu seuraaviin kokoelmiin

Näytä suppeat kuvailutiedot