Matches in SemOpenAlex for { <https://semopenalex.org/work/W2989728968> ?p ?o ?g. }
Showing items 1 to 81 of
81
with 100 items per page.
- W2989728968 abstract "Current state-of-the-art models for video action recognition are mostly based on expensive 3D ConvNets. This results in a need for large GPU clusters to train and evaluate such architectures. To address this problem, we present a lightweight and memory-friendly architecture for action recognition that performs on par with or better than current architectures by using only a fraction of resources. The proposed architecture is based on a combination of a deep subnet operating on low-resolution frames with a compact subnet operating on high-resolution frames, allowing for high efficiency and accuracy at the same time. We demonstrate that our approach achieves a reduction by $3sim4$ times in FLOPs and $sim2$ times in memory usage compared to the baseline. This enables training deeper models with more input frames under the same computational budget. To further obviate the need for large-scale 3D convolutions, a temporal aggregation module is proposed to model temporal dependencies in a video at very small additional computational costs. Our models achieve strong performance on several action recognition benchmarks including Kinetics, Something-Something and Moments-in-time. The code and models are available at https://github.com/IBM/bLVNet-TAM." @default.
- W2989728968 created "2019-12-05" @default.
- W2989728968 creator A5011795407 @default.
- W2989728968 creator A5025006492 @default.
- W2989728968 creator A5052325705 @default.
- W2989728968 creator A5066501364 @default.
- W2989728968 creator A5080417608 @default.
- W2989728968 date "2019-12-02" @default.
- W2989728968 modified "2023-10-03" @default.
- W2989728968 title "More Is Less: Learning Efficient Video Representations by Big-Little Network and Depthwise Temporal Aggregation" @default.
- W2989728968 doi "https://doi.org/10.48550/arxiv.1912.00869" @default.
- W2989728968 hasPublicationYear "2019" @default.
- W2989728968 type Work @default.
- W2989728968 sameAs 2989728968 @default.
- W2989728968 citedByCount "34" @default.
- W2989728968 countsByYear W29897289682020 @default.
- W2989728968 countsByYear W29897289682021 @default.
- W2989728968 countsByYear W29897289682022 @default.
- W2989728968 countsByYear W29897289682023 @default.
- W2989728968 crossrefType "posted-content" @default.
- W2989728968 hasAuthorship W2989728968A5011795407 @default.
- W2989728968 hasAuthorship W2989728968A5025006492 @default.
- W2989728968 hasAuthorship W2989728968A5052325705 @default.
- W2989728968 hasAuthorship W2989728968A5066501364 @default.
- W2989728968 hasAuthorship W2989728968A5080417608 @default.
- W2989728968 hasBestOaLocation W29897289681 @default.
- W2989728968 hasConcept C108583219 @default.
- W2989728968 hasConcept C111335779 @default.
- W2989728968 hasConcept C113775141 @default.
- W2989728968 hasConcept C126042441 @default.
- W2989728968 hasConcept C153180895 @default.
- W2989728968 hasConcept C154945302 @default.
- W2989728968 hasConcept C171250308 @default.
- W2989728968 hasConcept C173608175 @default.
- W2989728968 hasConcept C177264268 @default.
- W2989728968 hasConcept C192562407 @default.
- W2989728968 hasConcept C199360897 @default.
- W2989728968 hasConcept C21099817 @default.
- W2989728968 hasConcept C2524010 @default.
- W2989728968 hasConcept C2776760102 @default.
- W2989728968 hasConcept C31258907 @default.
- W2989728968 hasConcept C33923547 @default.
- W2989728968 hasConcept C3826847 @default.
- W2989728968 hasConcept C41008148 @default.
- W2989728968 hasConcept C70388272 @default.
- W2989728968 hasConceptScore W2989728968C108583219 @default.
- W2989728968 hasConceptScore W2989728968C111335779 @default.
- W2989728968 hasConceptScore W2989728968C113775141 @default.
- W2989728968 hasConceptScore W2989728968C126042441 @default.
- W2989728968 hasConceptScore W2989728968C153180895 @default.
- W2989728968 hasConceptScore W2989728968C154945302 @default.
- W2989728968 hasConceptScore W2989728968C171250308 @default.
- W2989728968 hasConceptScore W2989728968C173608175 @default.
- W2989728968 hasConceptScore W2989728968C177264268 @default.
- W2989728968 hasConceptScore W2989728968C192562407 @default.
- W2989728968 hasConceptScore W2989728968C199360897 @default.
- W2989728968 hasConceptScore W2989728968C21099817 @default.
- W2989728968 hasConceptScore W2989728968C2524010 @default.
- W2989728968 hasConceptScore W2989728968C2776760102 @default.
- W2989728968 hasConceptScore W2989728968C31258907 @default.
- W2989728968 hasConceptScore W2989728968C33923547 @default.
- W2989728968 hasConceptScore W2989728968C3826847 @default.
- W2989728968 hasConceptScore W2989728968C41008148 @default.
- W2989728968 hasConceptScore W2989728968C70388272 @default.
- W2989728968 hasLocation W29897289681 @default.
- W2989728968 hasOpenAccess W2989728968 @default.
- W2989728968 hasPrimaryLocation W29897289681 @default.
- W2989728968 hasRelatedWork W2733060750 @default.
- W2989728968 hasRelatedWork W2738221750 @default.
- W2989728968 hasRelatedWork W2766146978 @default.
- W2989728968 hasRelatedWork W2773120646 @default.
- W2989728968 hasRelatedWork W3156786002 @default.
- W2989728968 hasRelatedWork W3208028783 @default.
- W2989728968 hasRelatedWork W4220775285 @default.
- W2989728968 hasRelatedWork W4245792239 @default.
- W2989728968 hasRelatedWork W4317987726 @default.
- W2989728968 hasRelatedWork W3108696707 @default.
- W2989728968 isParatext "false" @default.
- W2989728968 isRetracted "false" @default.
- W2989728968 magId "2989728968" @default.
- W2989728968 workType "article" @default.