FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3124404582> ?p ?o ?g. }

• W3124404582 abstract "The limits of applicability of vision-and-language models are defined by the coverage of their training data. Tasks like vision question answering (VQA) often require commonsense and factual information beyond what can be learned from task-specific datasets. This paper investigates the injection of knowledge from general-purpose knowledge bases (KBs) into vision-and-language transformers. We use an auxiliary training objective that encourages the learned representations to align with graph embeddings of matching entities in a KB. We empirically study the relevance of various KBs to multiple tasks and benchmarks. The technique brings clear benefits to knowledge-demanding question answering tasks (OK-VQA, FVQA) by capturing semantic and relational knowledge absent from existing models. More surprisingly, the technique also benefits visual reasoning tasks (NLVR2, SNLI-VE). We perform probing experiments and show that the injection of additional knowledge regularizes the space of embeddings, which improves the representation of lexical and semantic similarities. The technique is model-agnostic and can expand the applicability of any vision-and-language transformer with minimal computational overhead." @default.
• W3124404582 created "2021-02-01" @default.
• W3124404582 creator A5015805464 @default.
• W3124404582 creator A5028024287 @default.
• W3124404582 creator A5067549788 @default.
• W3124404582 creator A5082520166 @default.
• W3124404582 date "2021-01-15" @default.
• W3124404582 modified "2023-09-27" @default.
• W3124404582 title "Reasoning over Vision and Language: Exploring the Benefits of Supplemental Knowledge" @default.
• W3124404582 cites W1861492603 @default.
• W3124404582 cites W1889081078 @default.
• W3124404582 cites W2080133951 @default.
• W3124404582 cites W2185175083 @default.
• W3124404582 cites W2231285690 @default.
• W3124404582 cites W2250539671 @default.
• W3124404582 cites W2251818205 @default.
• W3124404582 cites W2525778437 @default.
• W3124404582 cites W2555661914 @default.
• W3124404582 cites W2560730294 @default.
• W3124404582 cites W2626778328 @default.
• W3124404582 cites W2745461083 @default.
• W3124404582 cites W2767794011 @default.
• W3124404582 cites W2770653590 @default.
• W3124404582 cites W2790235966 @default.
• W3124404582 cites W2794325560 @default.
• W3124404582 cites W2804243936 @default.
• W3124404582 cites W2886641317 @default.
• W3124404582 cites W2890894339 @default.
• W3124404582 cites W2891394954 @default.
• W3124404582 cites W2896348597 @default.
• W3124404582 cites W2898695519 @default.
• W3124404582 cites W2899335602 @default.
• W3124404582 cites W2912371042 @default.
• W3124404582 cites W2928107702 @default.
• W3124404582 cites W2936262349 @default.
• W3124404582 cites W2940753425 @default.
• W3124404582 cites W2946345909 @default.
• W3124404582 cites W2947312908 @default.
• W3124404582 cites W2948036864 @default.
• W3124404582 cites W2950104027 @default.
• W3124404582 cites W2950577311 @default.
• W3124404582 cites W2952828476 @default.
• W3124404582 cites W2953084091 @default.
• W3124404582 cites W2962749469 @default.
• W3124404582 cites W2963341956 @default.
• W3124404582 cites W2963398599 @default.
• W3124404582 cites W2963717374 @default.
• W3124404582 cites W2963921132 @default.
• W3124404582 cites W2964138343 @default.
• W3124404582 cites W2964204621 @default.
• W3124404582 cites W2964207259 @default.
• W3124404582 cites W2964303913 @default.
• W3124404582 cites W2967593235 @default.
• W3124404582 cites W2968124245 @default.
• W3124404582 cites W2968289784 @default.
• W3124404582 cites W2968880719 @default.
• W3124404582 cites W2968908603 @default.
• W3124404582 cites W2969862959 @default.
• W3124404582 cites W2969876226 @default.
• W3124404582 cites W2970608575 @default.
• W3124404582 cites W2972969579 @default.
• W3124404582 cites W2974134141 @default.
• W3124404582 cites W2975501350 @default.
• W3124404582 cites W2980446769 @default.
• W3124404582 cites W2981851019 @default.
• W3124404582 cites W2983995706 @default.
• W3124404582 cites W2990752173 @default.
• W3124404582 cites W2997591391 @default.
• W3124404582 cites W2998374885 @default.
• W3124404582 cites W2998385486 @default.
• W3124404582 cites W3016672431 @default.
• W3124404582 cites W3020257313 @default.
• W3124404582 cites W3101703188 @default.
• W3124404582 cites W3106784008 @default.
• W3124404582 cites W3151929433 @default.
• W3124404582 cites W630532510 @default.
• W3124404582 hasPublicationYear "2021" @default.
• W3124404582 type Work @default.
• W3124404582 sameAs 3124404582 @default.
• W3124404582 citedByCount "1" @default.
• W3124404582 countsByYear W31244045822022 @default.
• W3124404582 crossrefType "posted-content" @default.
• W3124404582 hasAuthorship W3124404582A5015805464 @default.
• W3124404582 hasAuthorship W3124404582A5028024287 @default.
• W3124404582 hasAuthorship W3124404582A5067549788 @default.
• W3124404582 hasAuthorship W3124404582A5082520166 @default.
• W3124404582 hasConcept C119857082 @default.
• W3124404582 hasConcept C121332964 @default.
• W3124404582 hasConcept C154945302 @default.
• W3124404582 hasConcept C161301231 @default.
• W3124404582 hasConcept C162324750 @default.
• W3124404582 hasConcept C165801399 @default.
• W3124404582 hasConcept C187736073 @default.
• W3124404582 hasConcept C204321447 @default.
• W3124404582 hasConcept C2777508537 @default.
• W3124404582 hasConcept C2780451532 @default.
• W3124404582 hasConcept C2987255567 @default.
• W3124404582 hasConcept C30542707 @default.
• W3124404582 hasConcept C41008148 @default.
• W3124404582 hasConcept C44291984 @default.

• next