Matches in SemOpenAlex for { <https://semopenalex.org/work/W3211550315> ?p ?o ?g. }
Showing items 1 to 97 of
97
with 100 items per page.
- W3211550315 endingPage "4366" @default.
- W3211550315 startingPage "4357" @default.
- W3211550315 abstract "Linguistic representations derived from text alone have been criticized for their lack of grounding, i.e., connecting words to their meanings in the physical world. Vision-and- Language (VL) models, trained jointly on text and image or video data, have been offered as a response to such criticisms. However, while VL pretraining has shown success on multimodal tasks such as visual question answering, it is not yet known how the internal linguistic representations themselves compare to their text-only counterparts. This paper compares the semantic representations learned via VL vs. text-only pretraining for two recent VL models using a suite of analyses (clustering, probing, and performance on a commonsense question answering task) in a language-only setting. We find that the multimodal models fail to significantly outperform the text-only variants, suggesting that future work is required if multimodal pretraining is to be pursued as a means of improving NLP in general." @default.
- W3211550315 created "2021-11-22" @default.
- W3211550315 creator A5002734812 @default.
- W3211550315 creator A5053850863 @default.
- W3211550315 creator A5060145891 @default.
- W3211550315 date "2021-11-01" @default.
- W3211550315 modified "2023-09-28" @default.
- W3211550315 title "Does Vision-and-Language Pretraining Improve Lexical Grounding?" @default.
- W3211550315 cites W1889081078 @default.
- W3211550315 cites W1948251820 @default.
- W3211550315 cites W2014009760 @default.
- W3211550315 cites W2138615112 @default.
- W3211550315 cites W2549835527 @default.
- W3211550315 cites W2896739098 @default.
- W3211550315 cites W2931316642 @default.
- W3211550315 cites W2946359678 @default.
- W3211550315 cites W2946609015 @default.
- W3211550315 cites W2953307569 @default.
- W3211550315 cites W2963159690 @default.
- W3211550315 cites W2963341956 @default.
- W3211550315 cites W2963871344 @default.
- W3211550315 cites W2964121744 @default.
- W3211550315 cites W3004346089 @default.
- W3211550315 cites W3011411500 @default.
- W3211550315 cites W3020712669 @default.
- W3211550315 cites W3023306062 @default.
- W3211550315 cites W3030163527 @default.
- W3211550315 cites W3034723486 @default.
- W3211550315 cites W3090048242 @default.
- W3211550315 cites W3099178230 @default.
- W3211550315 cites W3100475986 @default.
- W3211550315 cites W3101065397 @default.
- W3211550315 hasPublicationYear "2021" @default.
- W3211550315 type Work @default.
- W3211550315 sameAs 3211550315 @default.
- W3211550315 citedByCount "0" @default.
- W3211550315 crossrefType "proceedings-article" @default.
- W3211550315 hasAuthorship W3211550315A5002734812 @default.
- W3211550315 hasAuthorship W3211550315A5053850863 @default.
- W3211550315 hasAuthorship W3211550315A5060145891 @default.
- W3211550315 hasConcept C137293760 @default.
- W3211550315 hasConcept C138885662 @default.
- W3211550315 hasConcept C154945302 @default.
- W3211550315 hasConcept C162324750 @default.
- W3211550315 hasConcept C166957645 @default.
- W3211550315 hasConcept C187736073 @default.
- W3211550315 hasConcept C204321447 @default.
- W3211550315 hasConcept C2780451532 @default.
- W3211550315 hasConcept C41008148 @default.
- W3211550315 hasConcept C41895202 @default.
- W3211550315 hasConcept C44291984 @default.
- W3211550315 hasConcept C73555534 @default.
- W3211550315 hasConcept C79581498 @default.
- W3211550315 hasConcept C95457728 @default.
- W3211550315 hasConceptScore W3211550315C137293760 @default.
- W3211550315 hasConceptScore W3211550315C138885662 @default.
- W3211550315 hasConceptScore W3211550315C154945302 @default.
- W3211550315 hasConceptScore W3211550315C162324750 @default.
- W3211550315 hasConceptScore W3211550315C166957645 @default.
- W3211550315 hasConceptScore W3211550315C187736073 @default.
- W3211550315 hasConceptScore W3211550315C204321447 @default.
- W3211550315 hasConceptScore W3211550315C2780451532 @default.
- W3211550315 hasConceptScore W3211550315C41008148 @default.
- W3211550315 hasConceptScore W3211550315C41895202 @default.
- W3211550315 hasConceptScore W3211550315C44291984 @default.
- W3211550315 hasConceptScore W3211550315C73555534 @default.
- W3211550315 hasConceptScore W3211550315C79581498 @default.
- W3211550315 hasConceptScore W3211550315C95457728 @default.
- W3211550315 hasLocation W32115503151 @default.
- W3211550315 hasOpenAccess W3211550315 @default.
- W3211550315 hasPrimaryLocation W32115503151 @default.
- W3211550315 hasRelatedWork W2009732762 @default.
- W3211550315 hasRelatedWork W2082398129 @default.
- W3211550315 hasRelatedWork W2139647815 @default.
- W3211550315 hasRelatedWork W2268023435 @default.
- W3211550315 hasRelatedWork W2400109828 @default.
- W3211550315 hasRelatedWork W2415378728 @default.
- W3211550315 hasRelatedWork W2619713362 @default.
- W3211550315 hasRelatedWork W2756244695 @default.
- W3211550315 hasRelatedWork W2766435878 @default.
- W3211550315 hasRelatedWork W2767826640 @default.
- W3211550315 hasRelatedWork W2789283398 @default.
- W3211550315 hasRelatedWork W2792726131 @default.
- W3211550315 hasRelatedWork W2886501253 @default.
- W3211550315 hasRelatedWork W2897252456 @default.
- W3211550315 hasRelatedWork W2978714200 @default.
- W3211550315 hasRelatedWork W3092733346 @default.
- W3211550315 hasRelatedWork W3099037975 @default.
- W3211550315 hasRelatedWork W3120008787 @default.
- W3211550315 hasRelatedWork W3199694729 @default.
- W3211550315 hasRelatedWork W3200568167 @default.
- W3211550315 isParatext "false" @default.
- W3211550315 isRetracted "false" @default.
- W3211550315 magId "3211550315" @default.
- W3211550315 workType "article" @default.