PhrasIS: Phrase Inference and Similarity benchmark

dc.contributor.authorLópez Gazpio, Íñigo
dc.contributor.authorGaviria de la Puerta, José
dc.contributor.authorGarcía, P.
dc.contributor.authorSanjurjo González, Hugo
dc.contributor.authorSanz, B.
dc.contributor.authorZarranz, A.
dc.contributor.authorMaritxalar Anglada, Montse
dc.contributor.authorAgirre, E.
dc.date.accessioned2025-01-07T15:40:16Z
dc.date.available2025-01-07T15:40:16Z
dc.date.issued2024-12
dc.date.updated2025-01-07T15:40:16Z
dc.description.abstractWe present PhrasIS, a benchmark dataset composed of natural occurring Phrase pairs with Inference and Similarity annotations for the evaluation of semantic representations. The described dataset fills the gap between word and sentence-level datasets, allowing to evaluate compositional models at a finer granularity than sentences. Contrary to other datasets, the phrase pairs are extracted from naturally occurring text in image captions and news headlines. All the text fragments have been annotated by experts following a rigorous process also described in the manuscript achieving high inter annotator agreement. In this work we analyse the dataset, showing the relation between inference labels and similarity scores. With 10K phrase pairs split in development and test, the dataset is an excellent benchmark for testing meaning representation systems.en
dc.identifier.citationLopez-Gazpio, Gaviria, García, Sanjurjo-González, Sanz, Zarranz, Maritxalar, & Agirre. (2024). PhrasIS: Phrase Inference and Similarity benchmark. Logic Journal of the IGPL, 32(6), 1088-1101. https://doi.org/10.1093/JIGPAL/JZAE037
dc.identifier.doi10.1093/JIGPAL/JZAE037
dc.identifier.eissn1368-9894
dc.identifier.issn1367-0751
dc.identifier.urihttp://hdl.handle.net/20.500.14454/2197
dc.language.isoeng
dc.publisherOxford University Press
dc.rights© The Author(s) 2024
dc.titlePhrasIS: Phrase Inference and Similarity benchmarken
dc.typejournal article
dcterms.accessRightsmetadata only access
oaire.citation.endPage1101
oaire.citation.issue6
oaire.citation.startPage1088
oaire.citation.titleLogic Journal of the IGPL
oaire.citation.volume32
Archivos
Colecciones