Matches in SemOpenAlex for { <https://semopenalex.org/work/W3122325173> ?p ?o ?g. }
- W3122325173 abstract "This paper presents Prototypical Contrastive Learning (PCL), an unsupervised representation learning method that addresses the fundamental limitations of instance-wise contrastive learning. PCL not only learns low-level features for the task of instance discrimination, but more importantly, it implicitly encodes semantic structures of the data into the learned embedding space. Specifically, we introduce prototypes as latent variables to help find the maximum-likelihood estimation of the network parameters in an Expectation-Maximization framework. We iteratively perform E-step as finding the distribution of prototypes via clustering and M-step as optimizing the network via contrastive learning. We propose ProtoNCE loss, a generalized version of the InfoNCE loss for contrastive learning, which encourages representations to be closer to their assigned prototypes. PCL outperforms state-of-the-art instance-wise contrastive learning methods on multiple benchmarks with substantial improvement in low-resource transfer learning. Code and pretrained models are available at https://github.com/salesforce/PCL." @default.
- W3122325173 created "2021-02-01" @default.
- W3122325173 creator A5004557807 @default.
- W3122325173 creator A5010883708 @default.
- W3122325173 creator A5032046813 @default.
- W3122325173 creator A5074834854 @default.
- W3122325173 date "2020-05-11" @default.
- W3122325173 modified "2023-10-18" @default.
- W3122325173 title "Prototypical Contrastive Learning of Unsupervised Representations" @default.
- W3122325173 cites W1861492603 @default.
- W3122325173 cites W2031489346 @default.
- W3122325173 cites W2082290707 @default.
- W3122325173 cites W2108598243 @default.
- W3122325173 cites W2118585731 @default.
- W3122325173 cites W2134670479 @default.
- W3122325173 cites W2148349024 @default.
- W3122325173 cites W2152790380 @default.
- W3122325173 cites W2162833336 @default.
- W3122325173 cites W2187089797 @default.
- W3122325173 cites W2194775991 @default.
- W3122325173 cites W2321533354 @default.
- W3122325173 cites W2326925005 @default.
- W3122325173 cites W2533545350 @default.
- W3122325173 cites W2558661413 @default.
- W3122325173 cites W2560977758 @default.
- W3122325173 cites W2593864460 @default.
- W3122325173 cites W2601450892 @default.
- W3122325173 cites W2613718673 @default.
- W3122325173 cites W2779692282 @default.
- W3122325173 cites W2798991696 @default.
- W3122325173 cites W2842511635 @default.
- W3122325173 cites W2883725317 @default.
- W3122325173 cites W2887997457 @default.
- W3122325173 cites W2909671697 @default.
- W3122325173 cites W2917551568 @default.
- W3122325173 cites W2949517790 @default.
- W3122325173 cites W2950141105 @default.
- W3122325173 cites W2962742544 @default.
- W3122325173 cites W2962852342 @default.
- W3122325173 cites W2963420272 @default.
- W3122325173 cites W2964074409 @default.
- W3122325173 cites W2964159205 @default.
- W3122325173 cites W2970241862 @default.
- W3122325173 cites W2971155163 @default.
- W3122325173 cites W2979579363 @default.
- W3122325173 cites W2987741655 @default.
- W3122325173 cites W2990500698 @default.
- W3122325173 cites W2990873191 @default.
- W3122325173 cites W2995181141 @default.
- W3122325173 cites W2995480165 @default.
- W3122325173 cites W2998388430 @default.
- W3122325173 cites W3009561768 @default.
- W3122325173 cites W3034576826 @default.
- W3122325173 cites W3034781633 @default.
- W3122325173 cites W3034978746 @default.
- W3122325173 cites W3035524453 @default.
- W3122325173 cites W3095121901 @default.
- W3122325173 cites W3101821705 @default.
- W3122325173 cites W3110446398 @default.
- W3122325173 cites W343636949 @default.
- W3122325173 hasPublicationYear "2020" @default.
- W3122325173 type Work @default.
- W3122325173 sameAs 3122325173 @default.
- W3122325173 citedByCount "43" @default.
- W3122325173 countsByYear W31223251732020 @default.
- W3122325173 countsByYear W31223251732021 @default.
- W3122325173 countsByYear W31223251732022 @default.
- W3122325173 crossrefType "posted-content" @default.
- W3122325173 hasAuthorship W3122325173A5004557807 @default.
- W3122325173 hasAuthorship W3122325173A5010883708 @default.
- W3122325173 hasAuthorship W3122325173A5032046813 @default.
- W3122325173 hasAuthorship W3122325173A5074834854 @default.
- W3122325173 hasBestOaLocation W31223251731 @default.
- W3122325173 hasConcept C119857082 @default.
- W3122325173 hasConcept C150899416 @default.
- W3122325173 hasConcept C154945302 @default.
- W3122325173 hasConcept C162324750 @default.
- W3122325173 hasConcept C177264268 @default.
- W3122325173 hasConcept C17744445 @default.
- W3122325173 hasConcept C187736073 @default.
- W3122325173 hasConcept C199360897 @default.
- W3122325173 hasConcept C199539241 @default.
- W3122325173 hasConcept C204321447 @default.
- W3122325173 hasConcept C2776359362 @default.
- W3122325173 hasConcept C2776760102 @default.
- W3122325173 hasConcept C2780451532 @default.
- W3122325173 hasConcept C41008148 @default.
- W3122325173 hasConcept C41608201 @default.
- W3122325173 hasConcept C51167844 @default.
- W3122325173 hasConcept C59404180 @default.
- W3122325173 hasConcept C73555534 @default.
- W3122325173 hasConcept C8038995 @default.
- W3122325173 hasConcept C94625758 @default.
- W3122325173 hasConceptScore W3122325173C119857082 @default.
- W3122325173 hasConceptScore W3122325173C150899416 @default.
- W3122325173 hasConceptScore W3122325173C154945302 @default.
- W3122325173 hasConceptScore W3122325173C162324750 @default.
- W3122325173 hasConceptScore W3122325173C177264268 @default.
- W3122325173 hasConceptScore W3122325173C17744445 @default.
- W3122325173 hasConceptScore W3122325173C187736073 @default.