Matches in SemOpenAlex for { <https://semopenalex.org/work/W4220787965> ?p ?o ?g. }
Showing items 1 to 78 of
78
with 100 items per page.
- W4220787965 endingPage "e899" @default.
- W4220787965 startingPage "e899" @default.
- W4220787965 abstract "Pre-trained multilingual models have been extensively used in cross-lingual information processing tasks. Existing work focuses on improving the transferring performance of pre-trained multilingual models but ignores the linguistic properties that models preserve at encoding time—“language identity”. We investigated the capability of state-of-the-art pre-trained multilingual models (mBERT, XLM, XLM-R) to preserve language identity through language typology. We explored model differences and variations in terms of languages, typological features, and internal hidden layers. We found the order of ability in preserving language identity of whole model and each of its hidden layers is: mBERT > XLM-R > XLM. Furthermore, all three models capture morphological, lexical, word order and syntactic features well, but perform poorly on nominal and verbal features. Finally, our results show that the ability of XLM-R and XLM remains stable across layers, but the ability of mBERT fluctuates severely. Our findings summarize the ability of each pre-trained multilingual model and its hidden layer to store language identity and typological features. It provides insights for later researchers in processing cross-lingual information." @default.
- W4220787965 created "2022-04-03" @default.
- W4220787965 creator A5037149192 @default.
- W4220787965 creator A5060002817 @default.
- W4220787965 date "2022-03-15" @default.
- W4220787965 modified "2023-09-27" @default.
- W4220787965 title "Probing language identity encoded in pre-trained multilingual models: a typological view" @default.
- W4220787965 cites W1949981797 @default.
- W4220787965 cites W1975781079 @default.
- W4220787965 cites W2053856724 @default.
- W4220787965 cites W2056540093 @default.
- W4220787965 cites W2890225082 @default.
- W4220787965 cites W2914924671 @default.
- W4220787965 cites W2925618549 @default.
- W4220787965 cites W2948947170 @default.
- W4220787965 cites W2952638691 @default.
- W4220787965 cites W2963250244 @default.
- W4220787965 cites W2963341956 @default.
- W4220787965 cites W2963560568 @default.
- W4220787965 cites W2980708516 @default.
- W4220787965 cites W3035390927 @default.
- W4220787965 cites W3035459196 @default.
- W4220787965 cites W3048717973 @default.
- W4220787965 cites W3100110884 @default.
- W4220787965 cites W3103490574 @default.
- W4220787965 cites W4240298715 @default.
- W4220787965 cites W4256617900 @default.
- W4220787965 doi "https://doi.org/10.7717/peerj-cs.899" @default.
- W4220787965 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/35494801" @default.
- W4220787965 hasPublicationYear "2022" @default.
- W4220787965 type Work @default.
- W4220787965 citedByCount "0" @default.
- W4220787965 crossrefType "journal-article" @default.
- W4220787965 hasAuthorship W4220787965A5037149192 @default.
- W4220787965 hasAuthorship W4220787965A5060002817 @default.
- W4220787965 hasBestOaLocation W42207879651 @default.
- W4220787965 hasConcept C121332964 @default.
- W4220787965 hasConcept C125411270 @default.
- W4220787965 hasConcept C137293760 @default.
- W4220787965 hasConcept C138885662 @default.
- W4220787965 hasConcept C154945302 @default.
- W4220787965 hasConcept C204321447 @default.
- W4220787965 hasConcept C24890656 @default.
- W4220787965 hasConcept C2778355321 @default.
- W4220787965 hasConcept C41008148 @default.
- W4220787965 hasConcept C41895202 @default.
- W4220787965 hasConceptScore W4220787965C121332964 @default.
- W4220787965 hasConceptScore W4220787965C125411270 @default.
- W4220787965 hasConceptScore W4220787965C137293760 @default.
- W4220787965 hasConceptScore W4220787965C138885662 @default.
- W4220787965 hasConceptScore W4220787965C154945302 @default.
- W4220787965 hasConceptScore W4220787965C204321447 @default.
- W4220787965 hasConceptScore W4220787965C24890656 @default.
- W4220787965 hasConceptScore W4220787965C2778355321 @default.
- W4220787965 hasConceptScore W4220787965C41008148 @default.
- W4220787965 hasConceptScore W4220787965C41895202 @default.
- W4220787965 hasLocation W42207879651 @default.
- W4220787965 hasLocation W42207879652 @default.
- W4220787965 hasLocation W42207879653 @default.
- W4220787965 hasLocation W42207879654 @default.
- W4220787965 hasOpenAccess W4220787965 @default.
- W4220787965 hasPrimaryLocation W42207879651 @default.
- W4220787965 hasRelatedWork W1989705153 @default.
- W4220787965 hasRelatedWork W2250741237 @default.
- W4220787965 hasRelatedWork W2293457016 @default.
- W4220787965 hasRelatedWork W2359001871 @default.
- W4220787965 hasRelatedWork W2389073067 @default.
- W4220787965 hasRelatedWork W2496228846 @default.
- W4220787965 hasRelatedWork W250853007 @default.
- W4220787965 hasRelatedWork W2789919619 @default.
- W4220787965 hasRelatedWork W2896411932 @default.
- W4220787965 hasRelatedWork W3169305685 @default.
- W4220787965 hasVolume "8" @default.
- W4220787965 isParatext "false" @default.
- W4220787965 isRetracted "false" @default.
- W4220787965 workType "article" @default.