Matches in SemOpenAlex for { <https://semopenalex.org/work/W4386076314> ?p ?o ?g. }
- W4386076314 abstract "We introduce LAVILA, a new approach to learning video-language representations by leveraging Large Language Models (LLMs). We repurpose pre-trained LLMs to be conditioned on visual input, and finetune them to create automatic video narrators. Our auto-generated narrations offer a number of advantages, including dense coverage of long videos, better temporal synchronization of the visual information and text, and much higher diversity of text. The video-language embedding learned contrastively with these narrations outperforms the previous state-of-the-art on multiple first-person and third-person video tasks, both in zero-shot and finetuned setups. Most notably, Lavilaobtains an absolute gain of 10.1% on EGTEA classification and 5.9% Epic-Kitchens-100 multi-instance retrieval benchmarks. Furthermore, LaVilatrained with only half the narrations from the Ego4D dataset outperforms models trained on the full set, and shows positive scaling behavior on increasing pre-training data and model size." @default.
- W4386076314 created "2023-08-23" @default.
- W4386076314 creator A5000623592 @default.
- W4386076314 creator A5006312307 @default.
- W4386076314 creator A5057535281 @default.
- W4386076314 creator A5088063692 @default.
- W4386076314 date "2023-06-01" @default.
- W4386076314 modified "2023-09-27" @default.
- W4386076314 title "Learning Video Representations from Large Language Models" @default.
- W4386076314 cites W1895577753 @default.
- W4386076314 cites W1927052826 @default.
- W4386076314 cites W1947481528 @default.
- W4386076314 cites W2064675550 @default.
- W4386076314 cites W2102765684 @default.
- W4386076314 cites W2108598243 @default.
- W4386076314 cites W2126579184 @default.
- W4386076314 cites W2194775991 @default.
- W4386076314 cites W2251658415 @default.
- W4386076314 cites W2799067027 @default.
- W4386076314 cites W2914699769 @default.
- W4386076314 cites W2952132648 @default.
- W4386076314 cites W2962784628 @default.
- W4386076314 cites W2963082988 @default.
- W4386076314 cites W2963524571 @default.
- W4386076314 cites W2964212550 @default.
- W4386076314 cites W2971296908 @default.
- W4386076314 cites W2972073579 @default.
- W4386076314 cites W2981851019 @default.
- W4386076314 cites W2984008963 @default.
- W4386076314 cites W3009803092 @default.
- W4386076314 cites W3035265375 @default.
- W4386076314 cites W3035635319 @default.
- W4386076314 cites W3098341425 @default.
- W4386076314 cites W3168640669 @default.
- W4386076314 cites W3172942063 @default.
- W4386076314 cites W3173621652 @default.
- W4386076314 cites W3174828871 @default.
- W4386076314 cites W3177419066 @default.
- W4386076314 cites W3182683290 @default.
- W4386076314 cites W3203574385 @default.
- W4386076314 cites W3203711169 @default.
- W4386076314 cites W3204588463 @default.
- W4386076314 cites W3205786327 @default.
- W4386076314 cites W3207758636 @default.
- W4386076314 cites W4214612132 @default.
- W4386076314 cites W4312271977 @default.
- W4386076314 cites W4312372834 @default.
- W4386076314 cites W4312658081 @default.
- W4386076314 cites W4312746376 @default.
- W4386076314 cites W4312772544 @default.
- W4386076314 doi "https://doi.org/10.1109/cvpr52729.2023.00637" @default.
- W4386076314 hasPublicationYear "2023" @default.
- W4386076314 type Work @default.
- W4386076314 citedByCount "0" @default.
- W4386076314 crossrefType "proceedings-article" @default.
- W4386076314 hasAuthorship W4386076314A5000623592 @default.
- W4386076314 hasAuthorship W4386076314A5006312307 @default.
- W4386076314 hasAuthorship W4386076314A5057535281 @default.
- W4386076314 hasAuthorship W4386076314A5088063692 @default.
- W4386076314 hasConcept C127162648 @default.
- W4386076314 hasConcept C137293760 @default.
- W4386076314 hasConcept C154945302 @default.
- W4386076314 hasConcept C177264268 @default.
- W4386076314 hasConcept C178790620 @default.
- W4386076314 hasConcept C185592680 @default.
- W4386076314 hasConcept C199360897 @default.
- W4386076314 hasConcept C204321447 @default.
- W4386076314 hasConcept C2778344882 @default.
- W4386076314 hasConcept C2778562939 @default.
- W4386076314 hasConcept C31258907 @default.
- W4386076314 hasConcept C41008148 @default.
- W4386076314 hasConcept C41608201 @default.
- W4386076314 hasConcept C49774154 @default.
- W4386076314 hasConceptScore W4386076314C127162648 @default.
- W4386076314 hasConceptScore W4386076314C137293760 @default.
- W4386076314 hasConceptScore W4386076314C154945302 @default.
- W4386076314 hasConceptScore W4386076314C177264268 @default.
- W4386076314 hasConceptScore W4386076314C178790620 @default.
- W4386076314 hasConceptScore W4386076314C185592680 @default.
- W4386076314 hasConceptScore W4386076314C199360897 @default.
- W4386076314 hasConceptScore W4386076314C204321447 @default.
- W4386076314 hasConceptScore W4386076314C2778344882 @default.
- W4386076314 hasConceptScore W4386076314C2778562939 @default.
- W4386076314 hasConceptScore W4386076314C31258907 @default.
- W4386076314 hasConceptScore W4386076314C41008148 @default.
- W4386076314 hasConceptScore W4386076314C41608201 @default.
- W4386076314 hasConceptScore W4386076314C49774154 @default.
- W4386076314 hasFunder F4320306076 @default.
- W4386076314 hasLocation W43860763141 @default.
- W4386076314 hasOpenAccess W4386076314 @default.
- W4386076314 hasPrimaryLocation W43860763141 @default.
- W4386076314 hasRelatedWork W1772447446 @default.
- W4386076314 hasRelatedWork W1989705153 @default.
- W4386076314 hasRelatedWork W2030250808 @default.
- W4386076314 hasRelatedWork W2355862304 @default.
- W4386076314 hasRelatedWork W2356108042 @default.
- W4386076314 hasRelatedWork W2358577407 @default.
- W4386076314 hasRelatedWork W2359001871 @default.
- W4386076314 hasRelatedWork W2376796979 @default.
- W4386076314 hasRelatedWork W2615259895 @default.