How About Time? Probing a Multilingual Language Model for Temporal Relations

Tommaso Caselli, Irene Dini, Felice Dell'Orletta

OnderzoeksoutputAcademicpeer review

1 Citaat (Scopus)
52 Downloads (Pure)

Samenvatting

This paper presents a comprehensive set of probing experiments using a multilingual language model, XLM-R, for temporal relation classification between events in four languages. Results show an advantage of contextualized embeddings over static ones and a detrimen- tal role of sentence level embeddings. While obtaining competitive results against state-of-the-art systems, our probes indicate a lack of suitable encoded information to properly address this task.
pdf bib abs
Originele taal-2English
TitelProceedings of the 29th International Conference on Computational Linguistics
Plaats van productieGyeongju, Republic of Korea
UitgeverijInternational Committee on Computational Linguistics (ICCL)
Pagina's3197-3209
Aantal pagina's13
StatusPublished - 2022
Evenement29th International Conference on Computational Linguistics - Gyeongju, Korea, Republic of
Duur: 12-okt.-202217-okt.-2022

Conference

Conference29th International Conference on Computational Linguistics
Verkorte titelCOLING 2022
Land/RegioKorea, Republic of
StadGyeongju
Periode12/10/202217/10/2022

Vingerafdruk

Duik in de onderzoeksthema's van 'How About Time? Probing a Multilingual Language Model for Temporal Relations'. Samen vormen ze een unieke vingerafdruk.

Citeer dit