Matches in SemOpenAlex for { <https://semopenalex.org/work/W4285192634> ?p ?o ?g. }
Showing items 1 to 70 of
70
with 100 items per page.
- W4285192634 abstract "Due to the limitations of the model structure and pre-training objectives, existing vision-and-language generation models cannot utilize pair-wise images and text through bi-directional generation. In this paper, we propose DU-VLG, a framework which unifies vision-and-language generation as sequence generation problems. DU-VLG is trained with novel dual pre-training tasks: multi-modal denoising autoencoder tasks and modality translation tasks. To bridge the gap between image understanding and generation, we further design a novel commitment loss. We compare pre-training objectives on image captioning and text-to-image generation datasets. Results show that DU-VLG yields better performance than variants trained with uni-directional generation objectives or the variant without the commitment loss. We also obtain higher scores compared to previous state-of-the-art systems on three vision-and-language generation tasks. In addition, human judges further confirm that our model generates real and relevant images as well as faithful and informative captions." @default.
- W4285192634 created "2022-07-14" @default.
- W4285192634 creator A5009352936 @default.
- W4285192634 creator A5036350345 @default.
- W4285192634 creator A5067689496 @default.
- W4285192634 creator A5068003633 @default.
- W4285192634 creator A5069371409 @default.
- W4285192634 date "2022-01-01" @default.
- W4285192634 modified "2023-09-25" @default.
- W4285192634 title "DU-VLG: Unifying Vision-and-Language Generation via Dual Sequence-to-Sequence Pre-training" @default.
- W4285192634 doi "https://doi.org/10.18653/v1/2022.findings-acl.201" @default.
- W4285192634 hasPublicationYear "2022" @default.
- W4285192634 type Work @default.
- W4285192634 citedByCount "0" @default.
- W4285192634 crossrefType "proceedings-article" @default.
- W4285192634 hasAuthorship W4285192634A5009352936 @default.
- W4285192634 hasAuthorship W4285192634A5036350345 @default.
- W4285192634 hasAuthorship W4285192634A5067689496 @default.
- W4285192634 hasAuthorship W4285192634A5068003633 @default.
- W4285192634 hasAuthorship W4285192634A5069371409 @default.
- W4285192634 hasBestOaLocation W42851926341 @default.
- W4285192634 hasConcept C101738243 @default.
- W4285192634 hasConcept C108583219 @default.
- W4285192634 hasConcept C115961682 @default.
- W4285192634 hasConcept C137293760 @default.
- W4285192634 hasConcept C138885662 @default.
- W4285192634 hasConcept C154945302 @default.
- W4285192634 hasConcept C157657479 @default.
- W4285192634 hasConcept C203005215 @default.
- W4285192634 hasConcept C204321447 @default.
- W4285192634 hasConcept C2778112365 @default.
- W4285192634 hasConcept C2780980858 @default.
- W4285192634 hasConcept C31972630 @default.
- W4285192634 hasConcept C41008148 @default.
- W4285192634 hasConcept C41895202 @default.
- W4285192634 hasConcept C54355233 @default.
- W4285192634 hasConcept C86803240 @default.
- W4285192634 hasConceptScore W4285192634C101738243 @default.
- W4285192634 hasConceptScore W4285192634C108583219 @default.
- W4285192634 hasConceptScore W4285192634C115961682 @default.
- W4285192634 hasConceptScore W4285192634C137293760 @default.
- W4285192634 hasConceptScore W4285192634C138885662 @default.
- W4285192634 hasConceptScore W4285192634C154945302 @default.
- W4285192634 hasConceptScore W4285192634C157657479 @default.
- W4285192634 hasConceptScore W4285192634C203005215 @default.
- W4285192634 hasConceptScore W4285192634C204321447 @default.
- W4285192634 hasConceptScore W4285192634C2778112365 @default.
- W4285192634 hasConceptScore W4285192634C2780980858 @default.
- W4285192634 hasConceptScore W4285192634C31972630 @default.
- W4285192634 hasConceptScore W4285192634C41008148 @default.
- W4285192634 hasConceptScore W4285192634C41895202 @default.
- W4285192634 hasConceptScore W4285192634C54355233 @default.
- W4285192634 hasConceptScore W4285192634C86803240 @default.
- W4285192634 hasLocation W42851926341 @default.
- W4285192634 hasLocation W42851926342 @default.
- W4285192634 hasOpenAccess W4285192634 @default.
- W4285192634 hasPrimaryLocation W42851926341 @default.
- W4285192634 hasRelatedWork W1512718085 @default.
- W4285192634 hasRelatedWork W2669956259 @default.
- W4285192634 hasRelatedWork W2939353110 @default.
- W4285192634 hasRelatedWork W3165097609 @default.
- W4285192634 hasRelatedWork W3165463024 @default.
- W4285192634 hasRelatedWork W4287178339 @default.
- W4285192634 hasRelatedWork W4288804510 @default.
- W4285192634 hasRelatedWork W4292874285 @default.
- W4285192634 hasRelatedWork W4327774331 @default.
- W4285192634 hasRelatedWork W2610387714 @default.
- W4285192634 isParatext "false" @default.
- W4285192634 isRetracted "false" @default.
- W4285192634 workType "article" @default.