Matches in SemOpenAlex for { <https://semopenalex.org/work/W4320340076> ?p ?o ?g. }
Showing items 1 to 53 of
53
with 100 items per page.
- W4320340076 abstract "Spatial commonsense, the knowledge about spatial position and relationship between objects (like the relative size of a lion and a girl, and the position of a boy relative to a bicycle when cycling), is an important part of commonsense knowledge. Although pretrained language models (PLMs) succeed in many NLP tasks, they are shown to be ineffective in spatial commonsense reasoning. Starting from the observation that images are more likely to exhibit spatial commonsense than texts, we explore whether models with visual signals learn more spatial commonsense than text-based PLMs. We propose a spatial commonsense benchmark that focuses on the relative scales of objects, and the positional relationship between people and objects under different actions. We probe PLMs and models with visual signals, including vision-language pretrained models and image synthesis models, on this benchmark, and find that image synthesis models are more capable of learning accurate and consistent spatial knowledge than other models. The spatial knowledge from image synthesis models also helps in natural language understanding tasks that require spatial commonsense." @default.
- W4320340076 created "2023-02-13" @default.
- W4320340076 creator A5036495504 @default.
- W4320340076 creator A5037132097 @default.
- W4320340076 creator A5043693897 @default.
- W4320340076 creator A5076409244 @default.
- W4320340076 date "2022-03-15" @default.
- W4320340076 modified "2023-10-18" @default.
- W4320340076 title "Things not Written in Text: Exploring Spatial Commonsense from Visual Signals" @default.
- W4320340076 doi "https://doi.org/10.48550/arxiv.2203.08075" @default.
- W4320340076 hasPublicationYear "2022" @default.
- W4320340076 type Work @default.
- W4320340076 citedByCount "0" @default.
- W4320340076 crossrefType "posted-content" @default.
- W4320340076 hasAuthorship W4320340076A5036495504 @default.
- W4320340076 hasAuthorship W4320340076A5037132097 @default.
- W4320340076 hasAuthorship W4320340076A5043693897 @default.
- W4320340076 hasAuthorship W4320340076A5076409244 @default.
- W4320340076 hasBestOaLocation W43203400761 @default.
- W4320340076 hasConcept C154945302 @default.
- W4320340076 hasConcept C161301231 @default.
- W4320340076 hasConcept C185798385 @default.
- W4320340076 hasConcept C193221554 @default.
- W4320340076 hasConcept C204321447 @default.
- W4320340076 hasConcept C205649164 @default.
- W4320340076 hasConcept C30542707 @default.
- W4320340076 hasConcept C41008148 @default.
- W4320340076 hasConcept C58640448 @default.
- W4320340076 hasConceptScore W4320340076C154945302 @default.
- W4320340076 hasConceptScore W4320340076C161301231 @default.
- W4320340076 hasConceptScore W4320340076C185798385 @default.
- W4320340076 hasConceptScore W4320340076C193221554 @default.
- W4320340076 hasConceptScore W4320340076C204321447 @default.
- W4320340076 hasConceptScore W4320340076C205649164 @default.
- W4320340076 hasConceptScore W4320340076C30542707 @default.
- W4320340076 hasConceptScore W4320340076C41008148 @default.
- W4320340076 hasConceptScore W4320340076C58640448 @default.
- W4320340076 hasLocation W43203400761 @default.
- W4320340076 hasOpenAccess W4320340076 @default.
- W4320340076 hasPrimaryLocation W43203400761 @default.
- W4320340076 hasRelatedWork W2962777029 @default.
- W4320340076 hasRelatedWork W2962911543 @default.
- W4320340076 hasRelatedWork W2964532710 @default.
- W4320340076 hasRelatedWork W2971788065 @default.
- W4320340076 hasRelatedWork W3107474891 @default.
- W4320340076 hasRelatedWork W3115951983 @default.
- W4320340076 hasRelatedWork W3168455342 @default.
- W4320340076 hasRelatedWork W4221141571 @default.
- W4320340076 hasRelatedWork W4287777598 @default.
- W4320340076 hasRelatedWork W4289528260 @default.
- W4320340076 isParatext "false" @default.
- W4320340076 isRetracted "false" @default.
- W4320340076 workType "article" @default.