Matches in SemOpenAlex for { <https://semopenalex.org/work/W4384080288> ?p ?o ?g. }
- W4384080288 endingPage "4087" @default.
- W4384080288 startingPage "4073" @default.
- W4384080288 abstract "Video-language pre-training has attracted considerable attention recently for its promising performance on various downstream tasks. Most existing methods utilize the modality-specific or modality-joint representation architectures for the cross-modality pre-training. Different from previous methods, this paper presents a novel architecture named Memory-augmented Inter-Modality Bridge (MemBridge), which uses the learnable intermediate modality representations as the bridge for the interaction between videos and language. Specifically, in the transformer-based cross-modality encoder, we introduce the learnable bridge tokens as the interaction approach, which means the video and language tokens can only perceive information from bridge tokens and themselves. Moreover, a memory bank is proposed to store abundant modality interaction information for adaptively generating bridge tokens according to different cases, enhancing the capacity and robustness of the inter-modality bridge. Through pre-training, MemBridge explicitly models the representations for more sufficient inter-modality interaction. Comprehensive experiments show that our approach achieves competitive performance with previous methods on various downstream tasks including video-text retrieval, video captioning, and video question answering on multiple datasets, demonstrating the effectiveness of the proposed method. The code has been available at https://github.com/jahhaoyang/MemBridge." @default.
- W4384080288 created "2023-07-13" @default.
- W4384080288 creator A5010880902 @default.
- W4384080288 creator A5021525470 @default.
- W4384080288 creator A5023582873 @default.
- W4384080288 creator A5042319605 @default.
- W4384080288 creator A5050256868 @default.
- W4384080288 creator A5067696234 @default.
- W4384080288 creator A5067977172 @default.
- W4384080288 creator A5085237331 @default.
- W4384080288 creator A5085719285 @default.
- W4384080288 date "2023-01-01" @default.
- W4384080288 modified "2023-10-18" @default.
- W4384080288 title "MemBridge: Video-Language Pre-training with Memory-Augmented Inter-Modality Bridge" @default.
- W4384080288 cites W1893116441 @default.
- W4384080288 cites W2425121537 @default.
- W4384080288 cites W2765716052 @default.
- W4384080288 cites W2885775891 @default.
- W4384080288 cites W2963017553 @default.
- W4384080288 cites W2984008963 @default.
- W4384080288 cites W3034730770 @default.
- W4384080288 cites W3035365026 @default.
- W4384080288 cites W3039060838 @default.
- W4384080288 cites W3043840704 @default.
- W4384080288 cites W3090602574 @default.
- W4384080288 cites W3105232955 @default.
- W4384080288 cites W3153005511 @default.
- W4384080288 cites W3168640669 @default.
- W4384080288 cites W3174441232 @default.
- W4384080288 cites W3176398504 @default.
- W4384080288 cites W3176481196 @default.
- W4384080288 cites W3176689360 @default.
- W4384080288 cites W3180463990 @default.
- W4384080288 cites W3197457832 @default.
- W4384080288 cites W3203711169 @default.
- W4384080288 cites W3204588463 @default.
- W4384080288 cites W3204670646 @default.
- W4384080288 cites W3217102353 @default.
- W4384080288 cites W4214692497 @default.
- W4384080288 cites W4221142658 @default.
- W4384080288 cites W4285606530 @default.
- W4384080288 cites W4312372711 @default.
- W4384080288 cites W4312784228 @default.
- W4384080288 cites W4313186260 @default.
- W4384080288 doi "https://doi.org/10.1109/tip.2023.3283916" @default.
- W4384080288 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/37436853" @default.
- W4384080288 hasPublicationYear "2023" @default.
- W4384080288 type Work @default.
- W4384080288 citedByCount "0" @default.
- W4384080288 crossrefType "journal-article" @default.
- W4384080288 hasAuthorship W4384080288A5010880902 @default.
- W4384080288 hasAuthorship W4384080288A5021525470 @default.
- W4384080288 hasAuthorship W4384080288A5023582873 @default.
- W4384080288 hasAuthorship W4384080288A5042319605 @default.
- W4384080288 hasAuthorship W4384080288A5050256868 @default.
- W4384080288 hasAuthorship W4384080288A5067696234 @default.
- W4384080288 hasAuthorship W4384080288A5067977172 @default.
- W4384080288 hasAuthorship W4384080288A5085237331 @default.
- W4384080288 hasAuthorship W4384080288A5085719285 @default.
- W4384080288 hasConcept C100776233 @default.
- W4384080288 hasConcept C104317684 @default.
- W4384080288 hasConcept C111919701 @default.
- W4384080288 hasConcept C115961682 @default.
- W4384080288 hasConcept C118505674 @default.
- W4384080288 hasConcept C121332964 @default.
- W4384080288 hasConcept C126322002 @default.
- W4384080288 hasConcept C154945302 @default.
- W4384080288 hasConcept C157657479 @default.
- W4384080288 hasConcept C165801399 @default.
- W4384080288 hasConcept C185592680 @default.
- W4384080288 hasConcept C204321447 @default.
- W4384080288 hasConcept C2780226545 @default.
- W4384080288 hasConcept C28490314 @default.
- W4384080288 hasConcept C41008148 @default.
- W4384080288 hasConcept C55493867 @default.
- W4384080288 hasConcept C62520636 @default.
- W4384080288 hasConcept C63479239 @default.
- W4384080288 hasConcept C66322947 @default.
- W4384080288 hasConcept C71924100 @default.
- W4384080288 hasConceptScore W4384080288C100776233 @default.
- W4384080288 hasConceptScore W4384080288C104317684 @default.
- W4384080288 hasConceptScore W4384080288C111919701 @default.
- W4384080288 hasConceptScore W4384080288C115961682 @default.
- W4384080288 hasConceptScore W4384080288C118505674 @default.
- W4384080288 hasConceptScore W4384080288C121332964 @default.
- W4384080288 hasConceptScore W4384080288C126322002 @default.
- W4384080288 hasConceptScore W4384080288C154945302 @default.
- W4384080288 hasConceptScore W4384080288C157657479 @default.
- W4384080288 hasConceptScore W4384080288C165801399 @default.
- W4384080288 hasConceptScore W4384080288C185592680 @default.
- W4384080288 hasConceptScore W4384080288C204321447 @default.
- W4384080288 hasConceptScore W4384080288C2780226545 @default.
- W4384080288 hasConceptScore W4384080288C28490314 @default.
- W4384080288 hasConceptScore W4384080288C41008148 @default.
- W4384080288 hasConceptScore W4384080288C55493867 @default.
- W4384080288 hasConceptScore W4384080288C62520636 @default.
- W4384080288 hasConceptScore W4384080288C63479239 @default.
- W4384080288 hasConceptScore W4384080288C66322947 @default.