Matches in SemOpenAlex for { <https://semopenalex.org/work/W3097062010> ?p ?o ?g. }
- W3097062010 abstract "Diverse image captioning models aim to learn one-to-many mappings that are innate to cross-domain datasets, such as of images and texts. Current methods for this task are based on generative latent variable models, e.g. VAEs with structured latent spaces. Yet, the amount of multimodality captured by prior work is limited to that of the paired training data -- the true diversity of the underlying generative process is not fully captured. To address this limitation, we leverage the contextual descriptions in the dataset that explain similar contexts in different visual scenes. To this end, we introduce a novel factorization of the latent space, termed context-object split, to model diversity in contextual descriptions across images and texts within the dataset. Our framework not only enables diverse captioning through context-based pseudo supervision, but extends this to images with novel objects and without paired captions in the training data. We evaluate our COS-CVAE approach on the standard COCO dataset and on the held-out COCO dataset consisting of images with novel objects, showing significant gains in accuracy and diversity." @default.
- W3097062010 created "2020-11-09" @default.
- W3097062010 creator A5020568224 @default.
- W3097062010 creator A5056098625 @default.
- W3097062010 date "2020-11-02" @default.
- W3097062010 modified "2023-10-01" @default.
- W3097062010 title "Diverse Image Captioning with Context-Object Split Latent Spaces" @default.
- W3097062010 cites W1514535095 @default.
- W3097062010 cites W1861492603 @default.
- W3097062010 cites W1895577753 @default.
- W3097062010 cites W1895989618 @default.
- W3097062010 cites W1956340063 @default.
- W3097062010 cites W1959608418 @default.
- W3097062010 cites W1969616664 @default.
- W3097062010 cites W1976806664 @default.
- W3097062010 cites W2101105183 @default.
- W3097062010 cites W2133459682 @default.
- W3097062010 cites W2154652894 @default.
- W3097062010 cites W2173180041 @default.
- W3097062010 cites W2277195237 @default.
- W3097062010 cites W2334763311 @default.
- W3097062010 cites W2481240925 @default.
- W3097062010 cites W2506483933 @default.
- W3097062010 cites W2508429489 @default.
- W3097062010 cites W2575842049 @default.
- W3097062010 cites W2578466053 @default.
- W3097062010 cites W2604178507 @default.
- W3097062010 cites W2745461083 @default.
- W3097062010 cites W2756073160 @default.
- W3097062010 cites W2788277448 @default.
- W3097062010 cites W2794583223 @default.
- W3097062010 cites W2795151422 @default.
- W3097062010 cites W2950178297 @default.
- W3097062010 cites W2954841306 @default.
- W3097062010 cites W2955034273 @default.
- W3097062010 cites W2962706528 @default.
- W3097062010 cites W2962982762 @default.
- W3097062010 cites W2963049308 @default.
- W3097062010 cites W2963088515 @default.
- W3097062010 cites W2963175879 @default.
- W3097062010 cites W2963349562 @default.
- W3097062010 cites W2963499204 @default.
- W3097062010 cites W2963594498 @default.
- W3097062010 cites W2963686907 @default.
- W3097062010 cites W2963695091 @default.
- W3097062010 cites W2963758027 @default.
- W3097062010 cites W2963877622 @default.
- W3097062010 cites W2967834053 @default.
- W3097062010 cites W2970626077 @default.
- W3097062010 cites W2979747405 @default.
- W3097062010 cites W2982553922 @default.
- W3097062010 cites W2983141445 @default.
- W3097062010 cites W2986670728 @default.
- W3097062010 cites W2988793532 @default.
- W3097062010 cites W2995199006 @default.
- W3097062010 cites W2998116350 @default.
- W3097062010 cites W3034655362 @default.
- W3097062010 cites W3104279398 @default.
- W3097062010 cites W639708223 @default.
- W3097062010 hasPublicationYear "2020" @default.
- W3097062010 type Work @default.
- W3097062010 sameAs 3097062010 @default.
- W3097062010 citedByCount "2" @default.
- W3097062010 countsByYear W30970620102021 @default.
- W3097062010 crossrefType "posted-content" @default.
- W3097062010 hasAuthorship W3097062010A5020568224 @default.
- W3097062010 hasAuthorship W3097062010A5056098625 @default.
- W3097062010 hasConcept C115961682 @default.
- W3097062010 hasConcept C119857082 @default.
- W3097062010 hasConcept C153083717 @default.
- W3097062010 hasConcept C153180895 @default.
- W3097062010 hasConcept C154945302 @default.
- W3097062010 hasConcept C157657479 @default.
- W3097062010 hasConcept C166957645 @default.
- W3097062010 hasConcept C167966045 @default.
- W3097062010 hasConcept C204321447 @default.
- W3097062010 hasConcept C205649164 @default.
- W3097062010 hasConcept C2779343474 @default.
- W3097062010 hasConcept C2781238097 @default.
- W3097062010 hasConcept C39890363 @default.
- W3097062010 hasConcept C41008148 @default.
- W3097062010 hasConcept C51167844 @default.
- W3097062010 hasConceptScore W3097062010C115961682 @default.
- W3097062010 hasConceptScore W3097062010C119857082 @default.
- W3097062010 hasConceptScore W3097062010C153083717 @default.
- W3097062010 hasConceptScore W3097062010C153180895 @default.
- W3097062010 hasConceptScore W3097062010C154945302 @default.
- W3097062010 hasConceptScore W3097062010C157657479 @default.
- W3097062010 hasConceptScore W3097062010C166957645 @default.
- W3097062010 hasConceptScore W3097062010C167966045 @default.
- W3097062010 hasConceptScore W3097062010C204321447 @default.
- W3097062010 hasConceptScore W3097062010C205649164 @default.
- W3097062010 hasConceptScore W3097062010C2779343474 @default.
- W3097062010 hasConceptScore W3097062010C2781238097 @default.
- W3097062010 hasConceptScore W3097062010C39890363 @default.
- W3097062010 hasConceptScore W3097062010C41008148 @default.
- W3097062010 hasConceptScore W3097062010C51167844 @default.
- W3097062010 hasLocation W30970620101 @default.
- W3097062010 hasOpenAccess W3097062010 @default.
- W3097062010 hasPrimaryLocation W30970620101 @default.