Matches in SemOpenAlex for { <https://semopenalex.org/work/W3037932287> ?p ?o ?g. }
Showing items 1 to 89 of
89
with 100 items per page.
- W3037932287 abstract "Unsupervised image captioning with no annotations is an emerging challenge in computer vision, where the existing arts usually adopt GAN (Generative Adversarial Networks) models. In this paper, we propose a novel memory-based network rather than GAN, named Recurrent Relational Memory Network ($R^2M$). Unlike complicated and sensitive adversarial learning that non-ideally performs for long sentence generation, $R^2M$ implements a concepts-to-sentence memory translator through two-stage memory mechanisms: fusion and recurrent memories, correlating the relational reasoning between common visual concepts and the generated words for long periods. $R^2M$ encodes visual context through unsupervised training on images, while enabling the memory to learn from irrelevant textual corpus via supervised fashion. Our solution enjoys less learnable parameters and higher computational efficiency than GAN-based methods, which heavily bear parameter sensitivity. We experimentally validate the superiority of $R^2M$ than state-of-the-arts on all benchmark datasets." @default.
- W3037932287 created "2020-07-02" @default.
- W3037932287 creator A5003642180 @default.
- W3037932287 creator A5015177224 @default.
- W3037932287 creator A5036726873 @default.
- W3037932287 creator A5059530979 @default.
- W3037932287 date "2020-06-24" @default.
- W3037932287 modified "2023-10-17" @default.
- W3037932287 title "Recurrent Relational Memory Network for Unsupervised Image Captioning" @default.
- W3037932287 cites W1861492603 @default.
- W3037932287 cites W1933349210 @default.
- W3037932287 cites W2185175083 @default.
- W3037932287 cites W2557728737 @default.
- W3037932287 cites W2626778328 @default.
- W3037932287 cites W2886641317 @default.
- W3037932287 cites W2908747729 @default.
- W3037932287 cites W2924200781 @default.
- W3037932287 cites W2937305817 @default.
- W3037932287 cites W2955956881 @default.
- W3037932287 cites W2963062932 @default.
- W3037932287 cites W2963743213 @default.
- W3037932287 cites W2963877622 @default.
- W3037932287 cites W2964013027 @default.
- W3037932287 cites W2968549119 @default.
- W3037932287 cites W2970241862 @default.
- W3037932287 cites W2979304729 @default.
- W3037932287 cites W2983141445 @default.
- W3037932287 cites W3121480429 @default.
- W3037932287 doi "https://doi.org/10.48550/arxiv.2006.13611" @default.
- W3037932287 hasPublicationYear "2020" @default.
- W3037932287 type Work @default.
- W3037932287 sameAs 3037932287 @default.
- W3037932287 citedByCount "0" @default.
- W3037932287 crossrefType "posted-content" @default.
- W3037932287 hasAuthorship W3037932287A5003642180 @default.
- W3037932287 hasAuthorship W3037932287A5015177224 @default.
- W3037932287 hasAuthorship W3037932287A5036726873 @default.
- W3037932287 hasAuthorship W3037932287A5059530979 @default.
- W3037932287 hasBestOaLocation W30379322871 @default.
- W3037932287 hasConcept C115961682 @default.
- W3037932287 hasConcept C119857082 @default.
- W3037932287 hasConcept C13280743 @default.
- W3037932287 hasConcept C151730666 @default.
- W3037932287 hasConcept C154945302 @default.
- W3037932287 hasConcept C157657479 @default.
- W3037932287 hasConcept C185798385 @default.
- W3037932287 hasConcept C204321447 @default.
- W3037932287 hasConcept C205649164 @default.
- W3037932287 hasConcept C2777530160 @default.
- W3037932287 hasConcept C2779343474 @default.
- W3037932287 hasConcept C39890363 @default.
- W3037932287 hasConcept C41008148 @default.
- W3037932287 hasConcept C8038995 @default.
- W3037932287 hasConcept C86803240 @default.
- W3037932287 hasConcept C97931131 @default.
- W3037932287 hasConceptScore W3037932287C115961682 @default.
- W3037932287 hasConceptScore W3037932287C119857082 @default.
- W3037932287 hasConceptScore W3037932287C13280743 @default.
- W3037932287 hasConceptScore W3037932287C151730666 @default.
- W3037932287 hasConceptScore W3037932287C154945302 @default.
- W3037932287 hasConceptScore W3037932287C157657479 @default.
- W3037932287 hasConceptScore W3037932287C185798385 @default.
- W3037932287 hasConceptScore W3037932287C204321447 @default.
- W3037932287 hasConceptScore W3037932287C205649164 @default.
- W3037932287 hasConceptScore W3037932287C2777530160 @default.
- W3037932287 hasConceptScore W3037932287C2779343474 @default.
- W3037932287 hasConceptScore W3037932287C39890363 @default.
- W3037932287 hasConceptScore W3037932287C41008148 @default.
- W3037932287 hasConceptScore W3037932287C8038995 @default.
- W3037932287 hasConceptScore W3037932287C86803240 @default.
- W3037932287 hasConceptScore W3037932287C97931131 @default.
- W3037932287 hasLocation W30379322871 @default.
- W3037932287 hasLocation W30379322872 @default.
- W3037932287 hasOpenAccess W3037932287 @default.
- W3037932287 hasPrimaryLocation W30379322871 @default.
- W3037932287 hasRelatedWork W159132833 @default.
- W3037932287 hasRelatedWork W2541882558 @default.
- W3037932287 hasRelatedWork W2888227225 @default.
- W3037932287 hasRelatedWork W2903179935 @default.
- W3037932287 hasRelatedWork W2965707726 @default.
- W3037932287 hasRelatedWork W2971866238 @default.
- W3037932287 hasRelatedWork W3093267690 @default.
- W3037932287 hasRelatedWork W3196155444 @default.
- W3037932287 hasRelatedWork W3209574120 @default.
- W3037932287 hasRelatedWork W4307074315 @default.
- W3037932287 isParatext "false" @default.
- W3037932287 isRetracted "false" @default.
- W3037932287 magId "3037932287" @default.
- W3037932287 workType "article" @default.