Logo do repositório
 
Publicação

BATS-PT: Assessing Portuguese Masked Language Models in Lexico-Semantic Analogy Solving and Relation Completion

dc.contributor.authorOliveira, Hugo Gonçalo
dc.contributor.authorRodrigues, Ricardo
dc.contributor.authorFerreira, Bruno
dc.contributor.authorSilvano, Purificação
dc.contributor.authorCarvalho, Sara
dc.date.accessioned2026-06-01T13:38:45Z
dc.date.available2026-06-01T13:38:45Z
dc.date.issued2024
dc.description.abstractThis paper presents BATS-PT, the manual translation of the lexicographic portion of the Bigger Analogy Test Set (BATS) to European Portuguese. BATS-PT covers ten types of lexicosemantic analogies and can be used for assessing word embeddings and language models. Following this, the dataset is showcased while assessing two pretrained language models for Portuguese, BERTimbau and Albertina, in two tasks: analogy solving and relation completion, both in zero- and few-shot mask-prediction approaches. Experiments reveal different performance across relations and, in both tasks, the best overall performance was achieved with BERTimbau, in a five-shot scenario. We further discuss the limitations of the reported experiments and directions towards future improvements in these tasks.eng
dc.identifier.urihttp://hdl.handle.net/10400.26/63452
dc.language.isoeng
dc.peerreviewedn/a
dc.relationC645008882-00000055
dc.rights.urihttp://creativecommons.org/licenses/by/4.0/
dc.titleBATS-PT: Assessing Portuguese Masked Language Models in Lexico-Semantic Analogy Solving and Relation Completioneng
dc.typeconference paper
dspace.entity.typePublication
oaire.citation.endPage217
oaire.citation.startPage207
oaire.citation.titleProceedings of the 16th International Conference on Computational Processing of Portuguese - Vol. 1
oaire.versionhttp://purl.org/coar/version/c_970fb48d4fbd8a85
person.familyNameRodrigues
person.givenNameRicardo
person.identifier.ciencia-idD31C-FB4A-FEAA
person.identifier.orcid0000-0002-6262-7920
relation.isAuthorOfPublicationc64ccf7c-eca2-43cf-a4a2-78e684499c00
relation.isAuthorOfPublication.latestForDiscoveryc64ccf7c-eca2-43cf-a4a2-78e684499c00

Ficheiros

Principais
A mostrar 1 - 1 de 1
A carregar...
Miniatura
Nome:
2024.propor-1.21.pdf
Tamanho:
209.46 KB
Formato:
Adobe Portable Document Format
Licença
A mostrar 1 - 1 de 1
Miniatura indisponível
Nome:
license.txt
Tamanho:
1.85 KB
Formato:
Item-specific license agreed upon to submission
Descrição: