FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3174308973> ?p ?o ?g. }

• W3174308973 abstract "Contrastive self-supervised learning has largely narrowed the gap to supervised pre-training on ImageNet. However, its success highly relies on the object-centric priors of ImageNet, i.e., different augmented views of the same image correspond to the same object. Such a heavily curated constraint becomes immediately infeasible when pre-trained on more complex scene images with many objects. To overcome this limitation, we introduce Object-level Representation Learning (ORL), a new self-supervised learning framework towards scene images. Our key insight is to leverage image-level self-supervised pre-training as the prior to discover object-level semantic correspondence, thus realizing object-level representation learning from scene images. Extensive experiments on COCO show that ORL significantly improves the performance of self-supervised learning on scene images, even surpassing supervised ImageNet pre-training on several downstream tasks. Furthermore, ORL improves the downstream performance when more unlabeled scene images are available, demonstrating its great potential of harnessing unlabeled data in the wild. We hope our approach can motivate future research on more general-purpose unsupervised representation learning from scene data." @default.
• W3174308973 created "2021-07-05" @default.
• W3174308973 creator A5005626854 @default.
• W3174308973 creator A5006077923 @default.
• W3174308973 creator A5036271306 @default.
• W3174308973 creator A5040988693 @default.
• W3174308973 creator A5055183049 @default.
• W3174308973 date "2021-06-22" @default.
• W3174308973 modified "2023-09-27" @default.
• W3174308973 title "Unsupervised Object-Level Representation Learning from Scene Images" @default.
• W3174308973 cites W1665214252 @default.
• W3174308973 cites W1674866864 @default.
• W3174308973 cites W1836465849 @default.
• W3174308973 cites W1861492603 @default.
• W3174308973 cites W1903029394 @default.
• W3174308973 cites W1929856797 @default.
• W3174308973 cites W1955055330 @default.
• W3174308973 cites W2031489346 @default.
• W3174308973 cites W2088049833 @default.
• W3174308973 cites W2102605133 @default.
• W3174308973 cites W2108598243 @default.
• W3174308973 cites W2117539524 @default.
• W3174308973 cites W2118585731 @default.
• W3174308973 cites W2134670479 @default.
• W3174308973 cites W2138621090 @default.
• W3174308973 cites W2163605009 @default.
• W3174308973 cites W2194775991 @default.
• W3174308973 cites W2308529009 @default.
• W3174308973 cites W2321533354 @default.
• W3174308973 cites W2326925005 @default.
• W3174308973 cites W2412002662 @default.
• W3174308973 cites W2518108298 @default.
• W3174308973 cites W2558661413 @default.
• W3174308973 cites W2565639579 @default.
• W3174308973 cites W2613718673 @default.
• W3174308973 cites W2622263826 @default.
• W3174308973 cites W2625542360 @default.
• W3174308973 cites W2732026016 @default.
• W3174308973 cites W2747550417 @default.
• W3174308973 cites W2750549109 @default.
• W3174308973 cites W2798991696 @default.
• W3174308973 cites W2842511635 @default.
• W3174308973 cites W2887997457 @default.
• W3174308973 cites W2944828972 @default.
• W3174308973 cites W2946856970 @default.
• W3174308973 cites W2949517790 @default.
• W3174308973 cites W2962742544 @default.
• W3174308973 cites W2963016543 @default.
• W3174308973 cites W2963150697 @default.
• W3174308973 cites W2963420272 @default.
• W3174308973 cites W2963537932 @default.
• W3174308973 cites W2971155163 @default.
• W3174308973 cites W2979579363 @default.
• W3174308973 cites W2981952041 @default.
• W3174308973 cites W2987741655 @default.
• W3174308973 cites W2987936369 @default.
• W3174308973 cites W2990873191 @default.
• W3174308973 cites W2998388430 @default.
• W3174308973 cites W3009561768 @default.
• W3174308973 cites W3022061250 @default.
• W3174308973 cites W3034429256 @default.
• W3174308973 cites W3034781633 @default.
• W3174308973 cites W3034978746 @default.
• W3174308973 cites W3035524453 @default.
• W3174308973 cites W3080420240 @default.
• W3174308973 cites W3095121901 @default.
• W3174308973 cites W3101821705 @default.
• W3174308973 cites W3106428938 @default.
• W3174308973 cites W3106528393 @default.
• W3174308973 cites W3106539090 @default.
• W3174308973 cites W3107668149 @default.
• W3174308973 cites W3112137288 @default.
• W3174308973 cites W3128661784 @default.
• W3174308973 cites W3130807600 @default.
• W3174308973 cites W3133014250 @default.
• W3174308973 cites W3135715136 @default.
• W3174308973 cites W3135802977 @default.
• W3174308973 cites W3136027703 @default.
• W3174308973 cites W3137901979 @default.
• W3174308973 cites W3158714121 @default.
• W3174308973 cites W3172615411 @default.
• W3174308973 cites W343636949 @default.
• W3174308973 cites W3108262825 @default.
• W3174308973 hasPublicationYear "2021" @default.
• W3174308973 type Work @default.
• W3174308973 sameAs 3174308973 @default.
• W3174308973 citedByCount "2" @default.
• W3174308973 countsByYear W31743089732021 @default.
• W3174308973 crossrefType "posted-content" @default.
• W3174308973 hasAuthorship W3174308973A5005626854 @default.
• W3174308973 hasAuthorship W3174308973A5006077923 @default.
• W3174308973 hasAuthorship W3174308973A5036271306 @default.
• W3174308973 hasAuthorship W3174308973A5040988693 @default.
• W3174308973 hasAuthorship W3174308973A5055183049 @default.
• W3174308973 hasBestOaLocation W31743089731 @default.
• W3174308973 hasConcept C119857082 @default.
• W3174308973 hasConcept C136389625 @default.
• W3174308973 hasConcept C153083717 @default.
• W3174308973 hasConcept C153180895 @default.
• W3174308973 hasConcept C154945302 @default.

• next