Matches in SemOpenAlex for { <https://semopenalex.org/work/W4367171517> ?p ?o ?g. }
- W4367171517 abstract "Self-supervised learning models have been shown to learn rich visual representations without requiring human annotations. However, in many real-world scenarios, labels are partially available, motivating a recent line of work on semi-supervised methods inspired by self-supervised principles. In this paper, we propose a conceptually simple yet empirically powerful approach to turn clustering-based self-supervised methods such as SwAV or DINO into semi-supervised learners. More precisely, we introduce a multi-task framework merging a supervised objective using ground-truth labels and a self-supervised objective relying on clustering assignments with a single cross-entropy loss. This approach may be interpreted as imposing the cluster centroids to be class prototypes. Despite its simplicity, we provide empirical evidence that our approach is highly effective and achieves state-of-the-art performance on CI-FAR100 and ImageNet." @default.
- W4367171517 created "2023-04-28" @default.
- W4367171517 creator A5001459748 @default.
- W4367171517 creator A5049440980 @default.
- W4367171517 creator A5049825997 @default.
- W4367171517 creator A5062817741 @default.
- W4367171517 creator A5065059558 @default.
- W4367171517 creator A5066230537 @default.
- W4367171517 creator A5066621495 @default.
- W4367171517 date "2023-06-01" @default.
- W4367171517 modified "2023-09-27" @default.
- W4367171517 title "Semi-supervised learning made simple with self-supervised clustering" @default.
- W4367171517 cites W2097117768 @default.
- W4367171517 cites W2108598243 @default.
- W4367171517 cites W2194775991 @default.
- W4367171517 cites W2798991696 @default.
- W4367171517 cites W2808139377 @default.
- W4367171517 cites W2963080758 @default.
- W4367171517 cites W2992308087 @default.
- W4367171517 cites W2997131443 @default.
- W4367171517 cites W3034989734 @default.
- W4367171517 cites W3035524453 @default.
- W4367171517 cites W3091002423 @default.
- W4367171517 cites W3106743555 @default.
- W4367171517 cites W3109620645 @default.
- W4367171517 cites W3113757710 @default.
- W4367171517 cites W3114632476 @default.
- W4367171517 cites W3158278463 @default.
- W4367171517 cites W3158714121 @default.
- W4367171517 cites W3159481202 @default.
- W4367171517 cites W3171007011 @default.
- W4367171517 cites W3174699664 @default.
- W4367171517 cites W3180562345 @default.
- W4367171517 cites W3182493068 @default.
- W4367171517 cites W3185674647 @default.
- W4367171517 cites W3209458476 @default.
- W4367171517 cites W4213376695 @default.
- W4367171517 cites W4214687390 @default.
- W4367171517 cites W4221162144 @default.
- W4367171517 cites W4225777901 @default.
- W4367171517 cites W4312357822 @default.
- W4367171517 cites W4312421335 @default.
- W4367171517 cites W4312754496 @default.
- W4367171517 cites W4313156423 @default.
- W4367171517 doi "https://doi.org/10.1109/cvpr52729.2023.00311" @default.
- W4367171517 hasPublicationYear "2023" @default.
- W4367171517 type Work @default.
- W4367171517 citedByCount "0" @default.
- W4367171517 crossrefType "proceedings-article" @default.
- W4367171517 hasAuthorship W4367171517A5001459748 @default.
- W4367171517 hasAuthorship W4367171517A5049440980 @default.
- W4367171517 hasAuthorship W4367171517A5049825997 @default.
- W4367171517 hasAuthorship W4367171517A5062817741 @default.
- W4367171517 hasAuthorship W4367171517A5065059558 @default.
- W4367171517 hasAuthorship W4367171517A5066230537 @default.
- W4367171517 hasAuthorship W4367171517A5066621495 @default.
- W4367171517 hasBestOaLocation W43671715172 @default.
- W4367171517 hasConcept C106301342 @default.
- W4367171517 hasConcept C111472728 @default.
- W4367171517 hasConcept C119857082 @default.
- W4367171517 hasConcept C121332964 @default.
- W4367171517 hasConcept C136389625 @default.
- W4367171517 hasConcept C138885662 @default.
- W4367171517 hasConcept C146599234 @default.
- W4367171517 hasConcept C146849305 @default.
- W4367171517 hasConcept C154945302 @default.
- W4367171517 hasConcept C162324750 @default.
- W4367171517 hasConcept C187736073 @default.
- W4367171517 hasConcept C2776372474 @default.
- W4367171517 hasConcept C2780451532 @default.
- W4367171517 hasConcept C2780586882 @default.
- W4367171517 hasConcept C41008148 @default.
- W4367171517 hasConcept C50644808 @default.
- W4367171517 hasConcept C58973888 @default.
- W4367171517 hasConcept C62520636 @default.
- W4367171517 hasConcept C73555534 @default.
- W4367171517 hasConcept C8038995 @default.
- W4367171517 hasConceptScore W4367171517C106301342 @default.
- W4367171517 hasConceptScore W4367171517C111472728 @default.
- W4367171517 hasConceptScore W4367171517C119857082 @default.
- W4367171517 hasConceptScore W4367171517C121332964 @default.
- W4367171517 hasConceptScore W4367171517C136389625 @default.
- W4367171517 hasConceptScore W4367171517C138885662 @default.
- W4367171517 hasConceptScore W4367171517C146599234 @default.
- W4367171517 hasConceptScore W4367171517C146849305 @default.
- W4367171517 hasConceptScore W4367171517C154945302 @default.
- W4367171517 hasConceptScore W4367171517C162324750 @default.
- W4367171517 hasConceptScore W4367171517C187736073 @default.
- W4367171517 hasConceptScore W4367171517C2776372474 @default.
- W4367171517 hasConceptScore W4367171517C2780451532 @default.
- W4367171517 hasConceptScore W4367171517C2780586882 @default.
- W4367171517 hasConceptScore W4367171517C41008148 @default.
- W4367171517 hasConceptScore W4367171517C50644808 @default.
- W4367171517 hasConceptScore W4367171517C58973888 @default.
- W4367171517 hasConceptScore W4367171517C62520636 @default.
- W4367171517 hasConceptScore W4367171517C73555534 @default.
- W4367171517 hasConceptScore W4367171517C8038995 @default.
- W4367171517 hasLocation W43671715171 @default.
- W4367171517 hasLocation W43671715172 @default.
- W4367171517 hasLocation W43671715173 @default.