Matches in SemOpenAlex for { <https://semopenalex.org/work/W3143187051> ?p ?o ?g. }
- W3143187051 abstract "Data augmentation is a ubiquitous technique for improving image classification when labeled data is scarce. Constraining the model predictions to be invariant to diverse data augmentations effectively injects the desired representational invariances to the model (e.g., invariance to photometric variations) and helps improve accuracy. Compared to image data, the appearance variations in videos are far more complex due to the additional temporal dimension. Yet, data augmentation methods for videos remain under-explored. This paper investigates various data augmentation strategies that capture different video invariances, including photometric, geometric, temporal, and actor/scene augmentations. When integrated with existing semi-supervised learning frameworks, we show that our data augmentation strategy leads to promising performance on the Kinetics-100/400, Mini-Something-v2, UCF-101, and HMDB-51 datasets in the low-label regime. We also validate our data augmentation strategy in the fully supervised setting and demonstrate improved performance." @default.
- W3143187051 created "2021-04-13" @default.
- W3143187051 creator A5013292656 @default.
- W3143187051 creator A5025631228 @default.
- W3143187051 creator A5053730462 @default.
- W3143187051 creator A5062638357 @default.
- W3143187051 date "2022-01-01" @default.
- W3143187051 modified "2023-09-26" @default.
- W3143187051 title "Learning Representational Invariances for Data-Efficient Action Recognition" @default.
- W3143187051 cites W1522734439 @default.
- W3143187051 cites W1927052826 @default.
- W3143187051 cites W2126579184 @default.
- W3143187051 cites W2129068307 @default.
- W3143187051 cites W2145494108 @default.
- W3143187051 cites W2156303437 @default.
- W3143187051 cites W2163605009 @default.
- W3143187051 cites W2194775991 @default.
- W3143187051 cites W2337252826 @default.
- W3143187051 cites W2342662179 @default.
- W3143187051 cites W24089286 @default.
- W3143187051 cites W2592691248 @default.
- W3143187051 cites W2619947201 @default.
- W3143187051 cites W2625366777 @default.
- W3143187051 cites W2746314669 @default.
- W3143187051 cites W2773514261 @default.
- W3143187051 cites W2794583223 @default.
- W3143187051 cites W2804047946 @default.
- W3143187051 cites W2895243423 @default.
- W3143187051 cites W2902617128 @default.
- W3143187051 cites W2902709482 @default.
- W3143187051 cites W2943305281 @default.
- W3143187051 cites W2946856970 @default.
- W3143187051 cites W2948242301 @default.
- W3143187051 cites W2955874753 @default.
- W3143187051 cites W2962934715 @default.
- W3143187051 cites W2963045696 @default.
- W3143187051 cites W2963091558 @default.
- W3143187051 cites W2963155035 @default.
- W3143187051 cites W2963349562 @default.
- W3143187051 cites W2963435192 @default.
- W3143187051 cites W2963446712 @default.
- W3143187051 cites W2963524571 @default.
- W3143187051 cites W2964159205 @default.
- W3143187051 cites W2978426779 @default.
- W3143187051 cites W2981952041 @default.
- W3143187051 cites W2982358316 @default.
- W3143187051 cites W2990152177 @default.
- W3143187051 cites W2990503944 @default.
- W3143187051 cites W2991133498 @default.
- W3143187051 cites W2992308087 @default.
- W3143187051 cites W2994268455 @default.
- W3143187051 cites W2996501936 @default.
- W3143187051 cites W3009561768 @default.
- W3143187051 cites W3010874390 @default.
- W3143187051 cites W3034572008 @default.
- W3143187051 cites W3034938110 @default.
- W3143187051 cites W3034978746 @default.
- W3143187051 cites W3035037113 @default.
- W3143187051 cites W3035104321 @default.
- W3143187051 cites W3035224233 @default.
- W3143187051 cites W3035413240 @default.
- W3143187051 cites W3035524453 @default.
- W3143187051 cites W3045441084 @default.
- W3143187051 cites W3047740097 @default.
- W3143187051 cites W3048918001 @default.
- W3143187051 cites W3102631365 @default.
- W3143187051 cites W3105422445 @default.
- W3143187051 cites W3106031848 @default.
- W3143187051 cites W3118508703 @default.
- W3143187051 cites W3120167236 @default.
- W3143187051 cites W3127260722 @default.
- W3143187051 cites W3145385912 @default.
- W3143187051 doi "https://doi.org/10.2139/ssrn.4035476" @default.
- W3143187051 hasPublicationYear "2022" @default.
- W3143187051 type Work @default.
- W3143187051 sameAs 3143187051 @default.
- W3143187051 citedByCount "2" @default.
- W3143187051 countsByYear W31431870512021 @default.
- W3143187051 crossrefType "journal-article" @default.
- W3143187051 hasAuthorship W3143187051A5013292656 @default.
- W3143187051 hasAuthorship W3143187051A5025631228 @default.
- W3143187051 hasAuthorship W3143187051A5053730462 @default.
- W3143187051 hasAuthorship W3143187051A5062638357 @default.
- W3143187051 hasBestOaLocation W31431870512 @default.
- W3143187051 hasConcept C119857082 @default.
- W3143187051 hasConcept C121332964 @default.
- W3143187051 hasConcept C154945302 @default.
- W3143187051 hasConcept C15744967 @default.
- W3143187051 hasConcept C188147891 @default.
- W3143187051 hasConcept C2777212361 @default.
- W3143187051 hasConcept C2780791683 @default.
- W3143187051 hasConcept C2987834672 @default.
- W3143187051 hasConcept C41008148 @default.
- W3143187051 hasConcept C62520636 @default.
- W3143187051 hasConceptScore W3143187051C119857082 @default.
- W3143187051 hasConceptScore W3143187051C121332964 @default.
- W3143187051 hasConceptScore W3143187051C154945302 @default.
- W3143187051 hasConceptScore W3143187051C15744967 @default.
- W3143187051 hasConceptScore W3143187051C188147891 @default.
- W3143187051 hasConceptScore W3143187051C2777212361 @default.