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

Iiro Rastas, Yann Ciarán Ryan, Iiro Tiihonen, Mohammadreza Mohammadnia Qaraei, Liina Repo, Rohit Babbar, Eetu Mäkelä, Mikko Tolonen, Filip Ginter

Tutkimustuotos: Artikkeli kirjassa/konferenssijulkaisussaConference article in proceedingsScientificvertaisarvioitu

35 Lataukset (Pure)

Abstrakti

In 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.
AlkuperäiskieliEnglanti
OtsikkoProceedings of the 3rd Workshop on Computational Approaches to Historical Language Change
KustantajaAssociation for Computational Linguistics
Sivut68-77
Sivumäärä10
ISBN (elektroninen)978-1-955917-42-1
DOI - pysyväislinkit
TilaJulkaistu - 2022
OKM-julkaisutyyppiA4 Artikkeli konferenssijulkaisussa
TapahtumaWorkshop on Computational Approaches to Historical Language Change - Dublin, Irlanti
Kesto: 26 toukok. 202227 toukok. 2022
Konferenssinumero: 3

Workshop

WorkshopWorkshop on Computational Approaches to Historical Language Change
Maa/AlueIrlanti
KaupunkiDublin
Ajanjakso26/05/202227/05/2022

Sormenjälki

Sukella tutkimusaiheisiin 'Explainable Publication Year Prediction of Eighteenth Century Texts with the BERT Model'. Ne muodostavat yhdessä ainutlaatuisen sormenjäljen.

Siteeraa tätä