Matches in SemOpenAlex for { <https://semopenalex.org/work/W4287803913> ?p ?o ?g. }
Showing items 1 to 57 of
57
with 100 items per page.
- W4287803913 abstract "Following a navigation instruction such as 'Walk down the stairs and stop at the brown sofa' requires embodied AI agents to ground scene elements referenced via language (e.g. 'stairs') to visual content in the environment (pixels corresponding to 'stairs'). We ask the following question -- can we leverage abundant 'disembodied' web-scraped vision-and-language corpora (e.g. Conceptual Captions) to learn visual groundings (what do 'stairs' look like?) that improve performance on a relatively data-starved embodied perception task (Vision-and-Language Navigation)? Specifically, we develop VLN-BERT, a visiolinguistic transformer-based model for scoring the compatibility between an instruction ('...stop at the brown sofa') and a sequence of panoramic RGB images captured by the agent. We demonstrate that pretraining VLN-BERT on image-text pairs from the web before fine-tuning on embodied path-instruction data significantly improves performance on VLN -- outperforming the prior state-of-the-art in the fully-observed setting by 4 absolute percentage points on success rate. Ablations of our pretraining curriculum show each stage to be impactful -- with their combination resulting in further positive synergistic effects." @default.
- W4287803913 created "2022-07-26" @default.
- W4287803913 creator A5004995502 @default.
- W4287803913 creator A5014035752 @default.
- W4287803913 creator A5050342343 @default.
- W4287803913 creator A5051259505 @default.
- W4287803913 creator A5062122421 @default.
- W4287803913 creator A5063008596 @default.
- W4287803913 date "2020-04-30" @default.
- W4287803913 modified "2023-10-16" @default.
- W4287803913 title "Improving Vision-and-Language Navigation with Image-Text Pairs from the Web" @default.
- W4287803913 doi "https://doi.org/10.48550/arxiv.2004.14973" @default.
- W4287803913 hasPublicationYear "2020" @default.
- W4287803913 type Work @default.
- W4287803913 citedByCount "0" @default.
- W4287803913 crossrefType "posted-content" @default.
- W4287803913 hasAuthorship W4287803913A5004995502 @default.
- W4287803913 hasAuthorship W4287803913A5014035752 @default.
- W4287803913 hasAuthorship W4287803913A5050342343 @default.
- W4287803913 hasAuthorship W4287803913A5051259505 @default.
- W4287803913 hasAuthorship W4287803913A5062122421 @default.
- W4287803913 hasAuthorship W4287803913A5063008596 @default.
- W4287803913 hasBestOaLocation W42878039131 @default.
- W4287803913 hasConcept C100609095 @default.
- W4287803913 hasConcept C107457646 @default.
- W4287803913 hasConcept C127413603 @default.
- W4287803913 hasConcept C147176958 @default.
- W4287803913 hasConcept C153083717 @default.
- W4287803913 hasConcept C154945302 @default.
- W4287803913 hasConcept C2777295749 @default.
- W4287803913 hasConcept C31972630 @default.
- W4287803913 hasConcept C41008148 @default.
- W4287803913 hasConceptScore W4287803913C100609095 @default.
- W4287803913 hasConceptScore W4287803913C107457646 @default.
- W4287803913 hasConceptScore W4287803913C127413603 @default.
- W4287803913 hasConceptScore W4287803913C147176958 @default.
- W4287803913 hasConceptScore W4287803913C153083717 @default.
- W4287803913 hasConceptScore W4287803913C154945302 @default.
- W4287803913 hasConceptScore W4287803913C2777295749 @default.
- W4287803913 hasConceptScore W4287803913C31972630 @default.
- W4287803913 hasConceptScore W4287803913C41008148 @default.
- W4287803913 hasLocation W42878039131 @default.
- W4287803913 hasOpenAccess W4287803913 @default.
- W4287803913 hasPrimaryLocation W42878039131 @default.
- W4287803913 hasRelatedWork W13447594 @default.
- W4287803913 hasRelatedWork W14011730 @default.
- W4287803913 hasRelatedWork W1559077 @default.
- W4287803913 hasRelatedWork W4170893 @default.
- W4287803913 hasRelatedWork W4255058 @default.
- W4287803913 hasRelatedWork W5578107 @default.
- W4287803913 hasRelatedWork W7978787 @default.
- W4287803913 hasRelatedWork W8147663 @default.
- W4287803913 hasRelatedWork W8344810 @default.
- W4287803913 hasRelatedWork W9728998 @default.
- W4287803913 isParatext "false" @default.
- W4287803913 isRetracted "false" @default.
- W4287803913 workType "article" @default.