Matches in SemOpenAlex for { <https://semopenalex.org/work/W4366343077> ?p ?o ?g. }
Showing items 1 to 63 of
63
with 100 items per page.
- W4366343077 abstract "Modern AI applications involving video, such as video-text alignment, video search, and video captioning, benefit from a fine-grained understanding of video semantics. Existing approaches for video understanding are either data-hungry and need low-level annotation, or are based on general embeddings that are uninterpretable and can miss important details. We propose LASER, a neuro-symbolic approach that learns semantic video representations by leveraging logic specifications that can capture rich spatial and temporal properties in video data. In particular, we formulate the problem in terms of alignment between raw videos and specifications. The alignment process efficiently trains low-level perception models to extract a fine-grained video representation that conforms to the desired high-level specification. Our pipeline can be trained end-to-end and can incorporate contrastive and semantic loss functions derived from specifications. We evaluate our method on two datasets with rich spatial and temporal specifications: 20BN-Something-Something and MUGEN. We demonstrate that our method not only learns fine-grained video semantics but also outperforms existing baselines on downstream tasks such as video retrieval." @default.
- W4366343077 created "2023-04-20" @default.
- W4366343077 creator A5027652148 @default.
- W4366343077 creator A5051078160 @default.
- W4366343077 creator A5065252217 @default.
- W4366343077 creator A5067605731 @default.
- W4366343077 creator A5075879790 @default.
- W4366343077 date "2023-04-15" @default.
- W4366343077 modified "2023-10-16" @default.
- W4366343077 title "LASER: Neuro-Symbolic Learning of Semantic Video Representations" @default.
- W4366343077 doi "https://doi.org/10.48550/arxiv.2304.07647" @default.
- W4366343077 hasPublicationYear "2023" @default.
- W4366343077 type Work @default.
- W4366343077 citedByCount "0" @default.
- W4366343077 crossrefType "posted-content" @default.
- W4366343077 hasAuthorship W4366343077A5027652148 @default.
- W4366343077 hasAuthorship W4366343077A5051078160 @default.
- W4366343077 hasAuthorship W4366343077A5065252217 @default.
- W4366343077 hasAuthorship W4366343077A5067605731 @default.
- W4366343077 hasAuthorship W4366343077A5075879790 @default.
- W4366343077 hasBestOaLocation W43663430771 @default.
- W4366343077 hasConcept C115961682 @default.
- W4366343077 hasConcept C154945302 @default.
- W4366343077 hasConcept C157657479 @default.
- W4366343077 hasConcept C17744445 @default.
- W4366343077 hasConcept C184337299 @default.
- W4366343077 hasConcept C199360897 @default.
- W4366343077 hasConcept C199539241 @default.
- W4366343077 hasConcept C204321447 @default.
- W4366343077 hasConcept C2776359362 @default.
- W4366343077 hasConcept C41008148 @default.
- W4366343077 hasConcept C43521106 @default.
- W4366343077 hasConcept C94625758 @default.
- W4366343077 hasConcept C98045186 @default.
- W4366343077 hasConceptScore W4366343077C115961682 @default.
- W4366343077 hasConceptScore W4366343077C154945302 @default.
- W4366343077 hasConceptScore W4366343077C157657479 @default.
- W4366343077 hasConceptScore W4366343077C17744445 @default.
- W4366343077 hasConceptScore W4366343077C184337299 @default.
- W4366343077 hasConceptScore W4366343077C199360897 @default.
- W4366343077 hasConceptScore W4366343077C199539241 @default.
- W4366343077 hasConceptScore W4366343077C204321447 @default.
- W4366343077 hasConceptScore W4366343077C2776359362 @default.
- W4366343077 hasConceptScore W4366343077C41008148 @default.
- W4366343077 hasConceptScore W4366343077C43521106 @default.
- W4366343077 hasConceptScore W4366343077C94625758 @default.
- W4366343077 hasConceptScore W4366343077C98045186 @default.
- W4366343077 hasLocation W43663430771 @default.
- W4366343077 hasOpenAccess W4366343077 @default.
- W4366343077 hasPrimaryLocation W43663430771 @default.
- W4366343077 hasRelatedWork W1541271503 @default.
- W4366343077 hasRelatedWork W2020540721 @default.
- W4366343077 hasRelatedWork W2406518842 @default.
- W4366343077 hasRelatedWork W2795359650 @default.
- W4366343077 hasRelatedWork W2809904748 @default.
- W4366343077 hasRelatedWork W2901467237 @default.
- W4366343077 hasRelatedWork W2911868273 @default.
- W4366343077 hasRelatedWork W3107474891 @default.
- W4366343077 hasRelatedWork W4307856881 @default.
- W4366343077 hasRelatedWork W4320016076 @default.
- W4366343077 isParatext "false" @default.
- W4366343077 isRetracted "false" @default.
- W4366343077 workType "article" @default.