Matches in SemOpenAlex for { <https://semopenalex.org/work/W4387197110> ?p ?o ?g. }
- W4387197110 endingPage "5549" @default.
- W4387197110 startingPage "5537" @default.
- W4387197110 abstract "Visual Question Answering (VQA) is fundamentally compositional in nature, and many questions are simply answered by decomposing them into modular sub-problems. The recent proposed Neural Module Network (NMN) employ this strategy to question answering, whereas heavily rest with off-the-shelf layout parser or additional expert policy regarding the network architecture design instead of learning from the data. These strategies result in the unsatisfactory adaptability to the semantically-complicated variance of the inputs, thereby hindering the representational capacity and generalizability of the model. To tackle this problem, we propose a Semantic-aware modUlar caPsulE Routing framework, termed as SUPER, to better capture the instance-specific vision-semantic characteristics and refine the discriminative representations for prediction. Particularly, five powerful specialized modules as well as dynamic routers are tailored in each layer of the SUPER network, and the compact routing spaces are constructed such that a variety of customizable routes can be sufficiently exploited and the vision-semantic representations can be explicitly calibrated. We comparatively justify the effectiveness and generalization ability of our proposed SUPER scheme over five benchmark datasets, as well as the parametric-efficient advantage. It is worth emphasizing that this work is not to pursue the state-of-the-art results in VQA. Instead, we expect that our model is responsible to provide a novel perspective towards architecture learning and representation calibration for VQA." @default.
- W4387197110 created "2023-09-30" @default.
- W4387197110 creator A5000872135 @default.
- W4387197110 creator A5038612499 @default.
- W4387197110 creator A5039731055 @default.
- W4387197110 creator A5048997484 @default.
- W4387197110 creator A5054031708 @default.
- W4387197110 date "2023-01-01" @default.
- W4387197110 modified "2023-10-16" @default.
- W4387197110 title "Semantic-aware Modular Capsule Routing for Visual Question Answering" @default.
- W4387197110 cites W1933349210 @default.
- W4387197110 cites W2157331557 @default.
- W4387197110 cites W2194775991 @default.
- W4387197110 cites W2277195237 @default.
- W4387197110 cites W2471094925 @default.
- W4387197110 cites W2560730294 @default.
- W4387197110 cites W2561529111 @default.
- W4387197110 cites W2561715562 @default.
- W4387197110 cites W2745461083 @default.
- W4387197110 cites W2883104598 @default.
- W4387197110 cites W2886641317 @default.
- W4387197110 cites W2891394954 @default.
- W4387197110 cites W2896902935 @default.
- W4387197110 cites W2905524945 @default.
- W4387197110 cites W2916723116 @default.
- W4387197110 cites W2947312908 @default.
- W4387197110 cites W2962749469 @default.
- W4387197110 cites W2962944050 @default.
- W4387197110 cites W2963150162 @default.
- W4387197110 cites W2963176022 @default.
- W4387197110 cites W2963224792 @default.
- W4387197110 cites W2963383024 @default.
- W4387197110 cites W2963717374 @default.
- W4387197110 cites W2963938081 @default.
- W4387197110 cites W2964067226 @default.
- W4387197110 cites W2964072591 @default.
- W4387197110 cites W2964118342 @default.
- W4387197110 cites W2964303913 @default.
- W4387197110 cites W2966683369 @default.
- W4387197110 cites W2969679616 @default.
- W4387197110 cites W2981578638 @default.
- W4387197110 cites W2983995706 @default.
- W4387197110 cites W2997545008 @default.
- W4387197110 cites W3004349648 @default.
- W4387197110 cites W3034787499 @default.
- W4387197110 cites W3035454069 @default.
- W4387197110 cites W3037773948 @default.
- W4387197110 cites W3092767330 @default.
- W4387197110 cites W3093200502 @default.
- W4387197110 cites W3095718427 @default.
- W4387197110 cites W3099884329 @default.
- W4387197110 cites W3175703099 @default.
- W4387197110 cites W3203022498 @default.
- W4387197110 cites W3203354307 @default.
- W4387197110 doi "https://doi.org/10.1109/tip.2023.3318949" @default.
- W4387197110 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/37773902" @default.
- W4387197110 hasPublicationYear "2023" @default.
- W4387197110 type Work @default.
- W4387197110 citedByCount "0" @default.
- W4387197110 crossrefType "journal-article" @default.
- W4387197110 hasAuthorship W4387197110A5000872135 @default.
- W4387197110 hasAuthorship W4387197110A5038612499 @default.
- W4387197110 hasAuthorship W4387197110A5039731055 @default.
- W4387197110 hasAuthorship W4387197110A5048997484 @default.
- W4387197110 hasAuthorship W4387197110A5054031708 @default.
- W4387197110 hasConcept C101468663 @default.
- W4387197110 hasConcept C119857082 @default.
- W4387197110 hasConcept C13280743 @default.
- W4387197110 hasConcept C136197465 @default.
- W4387197110 hasConcept C154945302 @default.
- W4387197110 hasConcept C185798385 @default.
- W4387197110 hasConcept C199360897 @default.
- W4387197110 hasConcept C205649164 @default.
- W4387197110 hasConcept C31258907 @default.
- W4387197110 hasConcept C41008148 @default.
- W4387197110 hasConcept C44291984 @default.
- W4387197110 hasConcept C74172769 @default.
- W4387197110 hasConcept C97931131 @default.
- W4387197110 hasConceptScore W4387197110C101468663 @default.
- W4387197110 hasConceptScore W4387197110C119857082 @default.
- W4387197110 hasConceptScore W4387197110C13280743 @default.
- W4387197110 hasConceptScore W4387197110C136197465 @default.
- W4387197110 hasConceptScore W4387197110C154945302 @default.
- W4387197110 hasConceptScore W4387197110C185798385 @default.
- W4387197110 hasConceptScore W4387197110C199360897 @default.
- W4387197110 hasConceptScore W4387197110C205649164 @default.
- W4387197110 hasConceptScore W4387197110C31258907 @default.
- W4387197110 hasConceptScore W4387197110C41008148 @default.
- W4387197110 hasConceptScore W4387197110C44291984 @default.
- W4387197110 hasConceptScore W4387197110C74172769 @default.
- W4387197110 hasConceptScore W4387197110C97931131 @default.
- W4387197110 hasFunder F4320321001 @default.
- W4387197110 hasFunder F4320324174 @default.
- W4387197110 hasLocation W43871971101 @default.
- W4387197110 hasLocation W43871971102 @default.
- W4387197110 hasOpenAccess W4387197110 @default.
- W4387197110 hasPrimaryLocation W43871971101 @default.
- W4387197110 hasRelatedWork W2110523656 @default.