FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4304080240> ?p ?o ?g. }

• W4304080240 abstract "Video-and-language pre-training has shown promising results for learning generalizable representations. Most existing approaches usually model video and text in an implicit manner, without considering explicit structural representations of the multi-modal content. We denote such form of representations as structural knowledge, which express rich semantics of multiple granularities. There are related works that propose object-aware approaches to inject similar knowledge as inputs. However, the existing methods usually fail to effectively utilize such knowledge as regularizations to shape a superior cross-modal representation space. To this end, we propose a Cross-modaL knOwledge-enhanced Pre-training (CLOP) method with Knowledge Regularizations. There are two key designs of ours: 1) a simple yet effective Structural Knowledge Prediction (SKP) task to pull together the latent representations of similar videos; and 2) a novel Knowledge-guided sampling approach for Contrastive Learning (KCL) to push apart cross-modal hard negative samples. We evaluate our method on four text-video retrieval tasks and one multi-choice QA task. The experiments show clear improvements, outperforming prior works by a substantial margin. Besides, we provide ablations and insights of how our methods affect the latent representation space, demonstrating the value of incorporating knowledge regularizations into video-and-language pre-training." @default.
• W4304080240 created "2022-10-10" @default.
• W4304080240 creator A5000515409 @default.
• W4304080240 creator A5013438405 @default.
• W4304080240 creator A5028861603 @default.
• W4304080240 creator A5037949684 @default.
• W4304080240 creator A5043069455 @default.
• W4304080240 creator A5059003026 @default.
• W4304080240 creator A5078633131 @default.
• W4304080240 date "2022-10-10" @default.
• W4304080240 modified "2023-09-26" @default.
• W4304080240 title "CLOP: Video-and-Language Pre-Training with Knowledge Regularizations" @default.
• W4304080240 cites W1996430422 @default.
• W4304080240 cites W2123442489 @default.
• W4304080240 cites W2425121537 @default.
• W4304080240 cites W2745461083 @default.
• W4304080240 cites W2885775891 @default.
• W4304080240 cites W2963017553 @default.
• W4304080240 cites W2981851019 @default.
• W4304080240 cites W2984008963 @default.
• W4304080240 cites W2989322838 @default.
• W4304080240 cites W2990503944 @default.
• W4304080240 cites W2998356391 @default.
• W4304080240 cites W2998385486 @default.
• W4304080240 cites W2998702515 @default.
• W4304080240 cites W3035265375 @default.
• W4304080240 cites W3035356601 @default.
• W4304080240 cites W3035524453 @default.
• W4304080240 cites W3035635319 @default.
• W4304080240 cites W3043840704 @default.
• W4304080240 cites W3091588028 @default.
• W4304080240 cites W3096655658 @default.
• W4304080240 cites W3099206234 @default.
• W4304080240 cites W3109894131 @default.
• W4304080240 cites W3168640669 @default.
• W4304080240 cites W3175593095 @default.
• W4304080240 cites W3176750236 @default.
• W4304080240 cites W3204588463 @default.
• W4304080240 doi "https://doi.org/10.1145/3503161.3548346" @default.
• W4304080240 hasPublicationYear "2022" @default.
• W4304080240 type Work @default.
• W4304080240 citedByCount "0" @default.
• W4304080240 crossrefType "proceedings-article" @default.
• W4304080240 hasAuthorship W4304080240A5000515409 @default.
• W4304080240 hasAuthorship W4304080240A5013438405 @default.
• W4304080240 hasAuthorship W4304080240A5028861603 @default.
• W4304080240 hasAuthorship W4304080240A5037949684 @default.
• W4304080240 hasAuthorship W4304080240A5043069455 @default.
• W4304080240 hasAuthorship W4304080240A5059003026 @default.
• W4304080240 hasAuthorship W4304080240A5078633131 @default.
• W4304080240 hasBestOaLocation W43040802402 @default.
• W4304080240 hasConcept C111919701 @default.
• W4304080240 hasConcept C119857082 @default.
• W4304080240 hasConcept C154945302 @default.
• W4304080240 hasConcept C161301231 @default.
• W4304080240 hasConcept C162324750 @default.
• W4304080240 hasConcept C17744445 @default.
• W4304080240 hasConcept C184337299 @default.
• W4304080240 hasConcept C185592680 @default.
• W4304080240 hasConcept C187736073 @default.
• W4304080240 hasConcept C188027245 @default.
• W4304080240 hasConcept C199360897 @default.
• W4304080240 hasConcept C199539241 @default.
• W4304080240 hasConcept C204321447 @default.
• W4304080240 hasConcept C2776359362 @default.
• W4304080240 hasConcept C2778572836 @default.
• W4304080240 hasConcept C2780451532 @default.
• W4304080240 hasConcept C2781238097 @default.
• W4304080240 hasConcept C41008148 @default.
• W4304080240 hasConcept C59404180 @default.
• W4304080240 hasConcept C71139939 @default.
• W4304080240 hasConcept C774472 @default.
• W4304080240 hasConcept C94625758 @default.
• W4304080240 hasConceptScore W4304080240C111919701 @default.
• W4304080240 hasConceptScore W4304080240C119857082 @default.
• W4304080240 hasConceptScore W4304080240C154945302 @default.
• W4304080240 hasConceptScore W4304080240C161301231 @default.
• W4304080240 hasConceptScore W4304080240C162324750 @default.
• W4304080240 hasConceptScore W4304080240C17744445 @default.
• W4304080240 hasConceptScore W4304080240C184337299 @default.
• W4304080240 hasConceptScore W4304080240C185592680 @default.
• W4304080240 hasConceptScore W4304080240C187736073 @default.
• W4304080240 hasConceptScore W4304080240C188027245 @default.
• W4304080240 hasConceptScore W4304080240C199360897 @default.
• W4304080240 hasConceptScore W4304080240C199539241 @default.
• W4304080240 hasConceptScore W4304080240C204321447 @default.
• W4304080240 hasConceptScore W4304080240C2776359362 @default.
• W4304080240 hasConceptScore W4304080240C2778572836 @default.
• W4304080240 hasConceptScore W4304080240C2780451532 @default.
• W4304080240 hasConceptScore W4304080240C2781238097 @default.
• W4304080240 hasConceptScore W4304080240C41008148 @default.
• W4304080240 hasConceptScore W4304080240C59404180 @default.
• W4304080240 hasConceptScore W4304080240C71139939 @default.
• W4304080240 hasConceptScore W4304080240C774472 @default.
• W4304080240 hasConceptScore W4304080240C94625758 @default.
• W4304080240 hasLocation W43040802401 @default.
• W4304080240 hasLocation W43040802402 @default.
• W4304080240 hasLocation W43040802403 @default.
• W4304080240 hasOpenAccess W4304080240 @default.
• W4304080240 hasPrimaryLocation W43040802401 @default.

• next