Matches in SemOpenAlex for { <https://semopenalex.org/work/W3183824823> ?p ?o ?g. }
Showing items 1 to 82 of
82
with 100 items per page.
- W3183824823 endingPage "107313" @default.
- W3183824823 startingPage "107313" @default.
- W3183824823 abstract "Attention-based models have been widely used in image captioning. Nevertheless, most conventional deep attention models perform attention operations for each block/step independently, which neglects prior knowledge obtained by previous steps. In this paper, we propose a novel method — DYnamic Attention PRior (DY-APR), which combines both attention distribution prior and local linguistic context for caption generation. Like human beings, DY-APR can gradually shift its attention from a multitude of objects to the one of keen interest when coping with an image of a complex scene. DY-APR first captures rough information and then explicitly updates attention weights step by step. Besides, DY-APR fully leverages local linguistic context from the previous tokens, that is, capitalizes on local information when performing global attention — which we refer to as “local–global attention”. We show that the prior knowledge from previous steps provides meaningful semantic information, serving as guidance to build more accurate attention for the latter layers. Experiments on the MS-COCO dataset demonstrate the effectiveness of DY-APR, leading to CIDEr-D improvement by 2.32% with less than 0.2% additional FLOPs and parameters." @default.
- W3183824823 created "2021-08-02" @default.
- W3183824823 creator A5005905514 @default.
- W3183824823 creator A5009822141 @default.
- W3183824823 creator A5027937292 @default.
- W3183824823 creator A5059827848 @default.
- W3183824823 creator A5074960174 @default.
- W3183824823 date "2021-09-01" @default.
- W3183824823 modified "2023-09-23" @default.
- W3183824823 title "Reasoning like Humans: On Dynamic Attention Prior in Image Captioning" @default.
- W3183824823 cites W1895577753 @default.
- W3183824823 cites W1956340063 @default.
- W3183824823 cites W2113536531 @default.
- W3183824823 cites W2194775991 @default.
- W3183824823 cites W2277195237 @default.
- W3183824823 cites W2333091651 @default.
- W3183824823 cites W2575842049 @default.
- W3183824823 cites W2745461083 @default.
- W3183824823 cites W2795151422 @default.
- W3183824823 cites W2797733588 @default.
- W3183824823 cites W2885013662 @default.
- W3183824823 cites W2887585070 @default.
- W3183824823 cites W2890531016 @default.
- W3183824823 cites W2962843773 @default.
- W3183824823 cites W2963084599 @default.
- W3183824823 cites W2963101956 @default.
- W3183824823 cites W2966162142 @default.
- W3183824823 cites W2983141445 @default.
- W3183824823 cites W2984138079 @default.
- W3183824823 cites W2990818246 @default.
- W3183824823 cites W3000029975 @default.
- W3183824823 cites W3016489761 @default.
- W3183824823 cites W3034655362 @default.
- W3183824823 cites W3035160838 @default.
- W3183824823 cites W3035284526 @default.
- W3183824823 cites W3101313921 @default.
- W3183824823 cites W3103651098 @default.
- W3183824823 doi "https://doi.org/10.1016/j.knosys.2021.107313" @default.
- W3183824823 hasPublicationYear "2021" @default.
- W3183824823 type Work @default.
- W3183824823 sameAs 3183824823 @default.
- W3183824823 citedByCount "8" @default.
- W3183824823 countsByYear W31838248232021 @default.
- W3183824823 countsByYear W31838248232022 @default.
- W3183824823 crossrefType "journal-article" @default.
- W3183824823 hasAuthorship W3183824823A5005905514 @default.
- W3183824823 hasAuthorship W3183824823A5009822141 @default.
- W3183824823 hasAuthorship W3183824823A5027937292 @default.
- W3183824823 hasAuthorship W3183824823A5059827848 @default.
- W3183824823 hasAuthorship W3183824823A5074960174 @default.
- W3183824823 hasConcept C107457646 @default.
- W3183824823 hasConcept C115961682 @default.
- W3183824823 hasConcept C154945302 @default.
- W3183824823 hasConcept C157657479 @default.
- W3183824823 hasConcept C31972630 @default.
- W3183824823 hasConcept C41008148 @default.
- W3183824823 hasConceptScore W3183824823C107457646 @default.
- W3183824823 hasConceptScore W3183824823C115961682 @default.
- W3183824823 hasConceptScore W3183824823C154945302 @default.
- W3183824823 hasConceptScore W3183824823C157657479 @default.
- W3183824823 hasConceptScore W3183824823C31972630 @default.
- W3183824823 hasConceptScore W3183824823C41008148 @default.
- W3183824823 hasLocation W31838248231 @default.
- W3183824823 hasOpenAccess W3183824823 @default.
- W3183824823 hasPrimaryLocation W31838248231 @default.
- W3183824823 hasRelatedWork W2005185696 @default.
- W3183824823 hasRelatedWork W2092957489 @default.
- W3183824823 hasRelatedWork W2130228941 @default.
- W3183824823 hasRelatedWork W2161229648 @default.
- W3183824823 hasRelatedWork W2235753890 @default.
- W3183824823 hasRelatedWork W2366116130 @default.
- W3183824823 hasRelatedWork W2795359650 @default.
- W3183824823 hasRelatedWork W2923366293 @default.
- W3183824823 hasRelatedWork W2993674027 @default.
- W3183824823 hasRelatedWork W3008515501 @default.
- W3183824823 hasVolume "228" @default.
- W3183824823 isParatext "false" @default.
- W3183824823 isRetracted "false" @default.
- W3183824823 magId "3183824823" @default.
- W3183824823 workType "article" @default.